CN114692655A - Translation system and text translation, download, quality check and editing method - Google Patents

Abstract

The embodiment of the invention provides a translation system and a text translation, download, quality check and editing method, wherein the translation system comprises a front-end device and a back-end server, the front-end device is provided with a translation interface, the front-end device transmits an original text to the back-end server after receiving the original text uploaded by a user through the translation interface, the back-end server translates the original text to obtain a translation of the original text, then the obtained translation is transmitted to the front-end device, and the front-end device displays the translation on the translation interface. Therefore, a user only needs to upload the original text on a translation interface provided by the front-end device, the front-end device interacts with the back-end server, the back-end server translates the original text, the translation of the original text is displayed on the front-end device, the user does not need to input the translation word by word and sentence by sentence, translation time is greatly reduced, and rapid translation of the text is achieved.

Description

Translation system and text translation, download, quality check and editing method

Technical Field

The invention relates to the technical field of machine translation, in particular to a translation system and a text translation, downloading, quality inspection and editing method.

Background

In daily work, life and learning, people often need to translate foreign characters into Chinese, or Chinese characters into foreign characters. Aiming at the translation requirements of people, a plurality of translation software and translation websites are available on the market.

When the translation software and the translation website are used for translation, a user needs to input words to be translated into a translation dialog box word by word and sentence by sentence, and then click a translation button, so that a translation of a word or a sentence can be obtained.

In practical applications, users usually need to translate the whole text, and if the translation is still performed in a word-by-word and sentence-by-sentence input manner, the translation process is inevitably time-consuming and enormous. Therefore, how to quickly translate the whole text becomes an urgent problem to be solved.

Disclosure of Invention

The embodiment of the invention aims to provide a translation system and a text translation, downloading, quality inspection and editing method so as to realize rapid translation of a text. The specific technical scheme is as follows:

in a first aspect, an embodiment of the present invention provides a translation system, where the translation system includes: the system comprises front-end equipment and a back-end server, wherein the front-end equipment is provided with a translation interface;

the front-end equipment is used for transmitting the original text to the back-end server after receiving the original text uploaded by the user through the translation interface;

the back-end server is used for obtaining the original text; translating the original text to obtain a translated text of the original text; and transmitting the translation to the front-end equipment so that the front-end equipment displays the translation on the translation interface.

In a second aspect, an embodiment of the present invention provides a text translation method, where the method includes:

acquiring an original text uploaded by a user through a translation interface provided by front-end equipment;

translating the original text to obtain a translated text of the original text;

and transmitting the translation to the front-end equipment so that the front-end equipment displays the translation on the translation interface.

In a third aspect, an embodiment of the present invention provides a text downloading method, where the method includes:

receiving a downloading instruction initiated by a user through a translation interface provided by front-end equipment;

acquiring a translation corresponding to the downloading instruction;

identifying a downloading mode selected by a user through a translation interface and contained in the downloading instruction, wherein the downloading mode comprises translation text downloading and comparison text downloading;

if the downloading mode selected by the user is the downloading of the translation text, generating the translation text according to the translation and the specified text format, and sending the translation text to the front-end equipment for the user to download;

if the downloading mode selected by the user is the comparison text downloading, adding the translated text of the part of content after each part of content of the original text of the translated text is divided based on a preset text content division strategy, and generating a comparison text according to a specified text format after the addition is finished; and sending the comparison text to the front-end equipment for downloading by the user.

In a fourth aspect, an embodiment of the present invention provides a quality inspection method, including:

receiving a quality inspection instruction initiated by a user through a translation interface provided by front-end equipment;

selecting a plurality of sentences from the translated text to be checked according to the sequence of translation scores of the sentences from low to high;

the front-end device is instructed to highlight the plurality of sentences to prompt the user to focus the examination of the plurality of sentences.

In a fifth aspect, an embodiment of the present invention provides a text translation method, where the method includes:

acquiring an original text and a user-defined term library uploaded by a user through a translation interface provided by front-end equipment;

identifying whether text content in a user-defined term library exists in the original text;

if the translation rule exists, translating the corresponding text content in the original text according to the translation rule in the user-defined term library, and translating other text contents in the original text according to the general translation rule to obtain the translated text of the original text.

In a sixth aspect, an embodiment of the present invention provides a text translation editing method, where the method includes:

acquiring an original text uploaded by a user through a translation interface provided by front-end equipment, and translating the original text to obtain a translation of the original text displayed on the translation interface;

and under the condition that the user selects a translation editing mode through the translation interface, setting the translation into an editing state for the user to edit the content in the translation.

In a seventh aspect, an embodiment of the present invention provides a text translation apparatus, where the apparatus includes:

the obtaining module is used for obtaining an original text uploaded by a user through a translation interface provided by front-end equipment;

the translation module is used for translating the original text to obtain a translated text of the original text;

and the transmission module is used for transmitting the translation to the front-end equipment so that the front-end equipment displays the translation on the translation interface.

In an eighth aspect, an embodiment of the present invention provides a text downloading apparatus, where the apparatus includes:

the receiving module is used for receiving a downloading instruction initiated by a user through a translation interface provided by the front-end equipment;

the acquisition module is used for acquiring a translation corresponding to the downloading instruction;

the recognition module is used for recognizing a downloading mode selected by a user through the translation interface, wherein the downloading mode comprises translation text downloading and comparison text downloading;

the processing module is used for generating a translation text according to the translation and a specified text format if the downloading mode selected by the user is the translation text downloading, and sending the translation text to the front-end equipment for the user to download; if the downloading mode selected by the user is the comparison text downloading, adding the translated text of the content after each part of content of the original text division of the translated text based on a preset text content division strategy, and generating a comparison text according to a specified text format after the addition is completed; and sending the comparison text to the front-end equipment for downloading by the user.

In a ninth aspect, an embodiment of the present invention provides a quality inspection apparatus, including:

the receiving module is used for receiving a quality inspection instruction initiated by a user through a translation interface provided by the front-end equipment;

the selection module is used for selecting a plurality of sentences from the translated text to be checked according to the sequence of translation scores of the sentences from low to high;

and the indicating module is used for indicating the front-end equipment to highlight the plurality of sentences so as to prompt the user to perform key inspection on the plurality of sentences.

In a tenth aspect, an embodiment of the present invention provides a text translation apparatus, where the apparatus includes:

the acquisition module is used for acquiring an original text and a custom term library uploaded by a user through a translation interface provided by front-end equipment;

the identification module is used for identifying whether the text content in the user-defined term library exists in the original text;

and the translation module is used for translating the corresponding text content in the original text according to the translation rule in the user-defined term library if the translation module exists, and translating other text contents in the original text according to the general translation rule to obtain the translated text of the original text.

In an eleventh aspect, an embodiment of the present invention provides a text translation editing apparatus, where the apparatus includes:

the acquisition module is used for acquiring an original text uploaded by a user through a translation interface provided by front-end equipment, and translating the original text to obtain a translated text of the original text displayed on the translation interface;

and the setting module is used for setting the translated text into an editing state under the condition that the user selects a translation editing mode through the translation interface, so that the user can edit the content of the translated text.

In a twelfth aspect, an embodiment of the present invention provides an electronic device, including a processor and a memory, where the memory is used for storing a computer program; a processor for implementing, when executing the computer program stored on the memory: the method provided by the second aspect of the embodiment of the present invention, or the method provided by the third aspect of the embodiment of the present invention, or the method provided by the fourth aspect of the embodiment of the present invention, or the method provided by the fifth aspect of the embodiment of the present invention, or the method provided by the sixth aspect of the embodiment of the present invention.

In a thirteenth aspect, an embodiment of the present invention provides a computer-readable storage medium, in which a computer program is stored, and when the computer program is executed by a processor, the computer program implements: a method provided by the second aspect of the embodiment of the present invention, or a method provided by the third aspect of the embodiment of the present invention, or a method provided by the fourth aspect of the embodiment of the present invention, or a method provided by the fifth aspect of the embodiment of the present invention, or a method provided by the sixth aspect of the embodiment of the present invention.

In a fourteenth aspect, an embodiment of the present invention further provides a computer program product including instructions, which when run on a computer, cause the computer to perform: the method provided by the second aspect of the embodiment of the present invention, or the method provided by the third aspect of the embodiment of the present invention, or the method provided by the fourth aspect of the embodiment of the present invention, or the method provided by the fifth aspect of the embodiment of the present invention, or the method provided by the sixth aspect of the embodiment of the present invention.

The embodiment of the invention has the following beneficial effects: in the scheme provided by the embodiment of the invention, the translation system comprises front-end equipment and a back-end server, wherein the front-end equipment is provided with a translation interface, the front-end equipment transmits an original text to the back-end server after receiving the original text uploaded by a user through the translation interface, the back-end server translates the original text to obtain a translation of the original text, then the obtained translation is transmitted to the front-end equipment, and the front-end equipment displays the translation on the translation interface. Therefore, a user only needs to upload the original text on a translation interface provided by the front-end device, the front-end device interacts with the back-end server, the back-end server translates the original text, the translation of the original text is displayed on the front-end device, the user does not need to input the translation word by word and sentence by sentence, translation time is greatly reduced, and rapid translation of the text is achieved.

Drawings

In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and it is obvious for those skilled in the art that other embodiments can be obtained by using the drawings without creative efforts.

FIG. 1 is a schematic diagram of a translation system according to an embodiment of the present invention;

FIG. 2 is a flowchart illustrating a text translation method according to an embodiment of the present invention;

FIG. 3 is a flowchart illustrating a text translation method according to another embodiment of the present invention;

FIG. 4 is a flowchart illustrating a text translation method according to yet another embodiment of the present invention;

FIG. 5 is a flowchart illustrating a text translation method according to another embodiment of the present invention;

FIG. 6 is a flowchart illustrating a text translation method according to another embodiment of the present invention;

FIG. 7 is a flowchart illustrating a text translation method according to another embodiment of the present invention;

FIG. 8 is a schematic diagram of an overall process flow of a translation system for text translation according to an embodiment of the present invention;

FIG. 9 is a diagram of a translation interface according to an embodiment of the present invention;

FIG. 10 is a diagram illustrating editing options for a translated document in a translation interface according to an embodiment of the present invention;

FIG. 11 is a diagram illustrating download options according to an embodiment of the present invention;

FIG. 12 is a flowchart illustrating a process of translating an original document according to an embodiment of the present invention;

FIG. 13 is a diagram of a translation system architecture according to an embodiment of the present invention;

FIG. 14 is a flowchart illustrating a text downloading method according to an embodiment of the present invention;

FIG. 15 is a flow chart illustrating a quality inspection method according to an embodiment of the present invention;

FIG. 16 is a flowchart illustrating another method for translating text according to an embodiment of the present invention;

FIG. 17 is a flowchart illustrating a text translation editing method according to an embodiment of the present invention;

fig. 18 is a schematic structural diagram of a text translation apparatus according to an embodiment of the present invention;

FIG. 19 is a diagram illustrating a structure of a text downloading device according to an embodiment of the present invention;

FIG. 20 is a schematic structural diagram of a quality inspection apparatus according to an embodiment of the present invention;

FIG. 21 is a schematic structural diagram of another text translation apparatus according to an embodiment of the present invention;

fig. 22 is a schematic structural diagram of a text translation editing apparatus according to an embodiment of the present invention;

fig. 23 is a schematic structural diagram of an electronic device according to an embodiment of the present invention.

Detailed Description

The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.

In order to realize rapid translation of a text, the embodiment of the invention provides a translation system and a text translation, downloading, quality inspection and editing method.

As shown in fig. 1, the translation system provided in the embodiment of the present invention includes a client at a front-end device 110 and a server at a back-end server 120, where the front-end device 110 may be a mobile phone, a personal computer, a portable computer, a tablet computer, and the like, the front-end device 110 is provided with a translation interface, and the translation interface may be a user interaction interface provided by a translation APP (application) installed on the front-end device, a web page interface of a translation website opened on the front-end device by a user, or a user interaction interface provided by application software installed at the front-end device, and for convenience of discussion, the translation interface is presented to the user through the clients in the above three presentation manners.

The front-end device 110 may implement function operation and display in a mode of combining VUE (a set of frames for constructing a user interface) with Element UI (a user interface component), and the back-end server 120 may be composed of Springboot (a software development frame), Mysql (a relational database management system), Redis (remote dictionary service), Jodconvert (a file conversion tool), MongoDB (a distributed document storage database), and the like. The front-end device 110 and the back-end server 120 are provided with a translation Interface, such as an API (Application Programming Interface). In general, the data interaction may be performed through an HTTP (HyperText Transfer Protocol) interface and a translation interface.

The front-end device 110 transmits the original text to the back-end server 120 after receiving the original text uploaded by the user through the translation interface, where the original text uploaded by the user may be a document in a format, such as txt, doc/docx, xls/xlsx, ppt/pptx, pdf. The back-end server 120 translates the original text to obtain a translation of the original text after obtaining the original text, and then transmits the translation to the front-end device 110, so that the front-end device 110 displays the translation on the translation interface.

By applying the embodiment of the invention, the front-end equipment transmits the original text to the back-end server after receiving the original text uploaded by the user through the translation interface, the back-end server translates the original text to obtain the translation of the original text, then the obtained translation is transmitted to the front-end equipment, and the front-end equipment displays the translation on the translation interface. Therefore, a user only needs to upload the original text on a translation interface provided by the front-end device, the front-end device interacts with the back-end server, the back-end server translates the original text, the translated text of the original text is displayed on the front-end device, the user does not need to input the translated text word by word, and translation time is greatly shortened, so that the text can be quickly translated.

In a specific embodiment of the present invention, the translation system includes the following modules: the system comprises an automatic language identification module, a machine translation module, a document classification module, a post-translation editing module and a format conversion module. The language automatic translation module is used for identifying the language type of the original text; the machine translation module is used for translating each part of content of the original text; the document classification module is used for identifying the field type of the original text and generating language type identification; the post-translation editing module is used for providing a function of editing a translated text for a user; the format conversion module is used for identifying the format of the original text, converting the format of the original text into a format which can be identified by the machine translation module and generating the language type identifier.

As shown in fig. 2, the text translation method provided in the embodiment of the present invention is applied to a back-end server in the translation system, and includes the following steps. And the back-end server can be a server. The front-end equipment is provided with a client, and the client can be a web browser, an APP (application program) or a PC (personal computer) application program.

S201, obtaining an original text from a client.

Specifically, the method for the server to obtain the original text may be to obtain the original text uploaded by the user through a translation interface provided by the front-end device; the method can be used for acquiring the original text uploaded by a user through a translation interface provided by front-end equipment; or receiving an original text uploaded by a user through a translation interface from the front-end equipment; or receiving an original text uploaded by a user through a translation interface provided by front-end equipment; or the open API interface receives the original text sent by the client side from the front-end equipment.

Specifically, the original text uploaded by the user through the front-end device is encrypted and then sent to the back-end server, and after receiving the encrypted original text, the back-end server performs identity authentication to determine that the original text is a legal translation requirement and decrypts the original text. The identity authentication can be authentication of the user identity, authentication for judging whether the encrypted file is tampered or not, and authentication for judging the source of the encrypted file.

S202, translating the original text to obtain a translated text of the original text.

Specifically, the original text is respectively sent to a language automatic recognition module for language recognition, a document classification module for document classification, a format conversion module for format recognition and format conversion, and then the original text after format conversion, the document type identifier and the language type identifier are sent to a machine translation module for machine translation of the original text, so as to obtain a translation of the original text.

And S203, transmitting the translation to the client so that the translation is displayed on a translation interface at the front-end equipment.

Specifically, the back-end server transmits the translation to the client through the API interface, and the client displays the translation on the display of the front-end device through the translation interface.

By applying the embodiment of the invention, the back-end server translates the original text after obtaining the original text uploaded by the user through the translation interface provided by the front-end equipment to obtain the translation of the original text, then transmits the obtained translation to the front-end equipment, and the front-end equipment displays the translation on the translation interface. Therefore, a user only needs to upload the original text on a translation interface provided by the front-end device, the front-end device interacts with the back-end server, the back-end server translates the original text, the translation of the original text is displayed on the front-end device, the user does not need to input the translation word by word and sentence by sentence, translation time is greatly reduced, and rapid translation of the text is achieved. The invention realizes document translation by the machine translation technology of the machine translation module, and has high translation speed and high accuracy. The document translation technology is different from the traditional dictionary translation technology, the dictionary translation technology does not consider the relation between the whole document sentences and sentences, the translation is carried out word by word and sentence by sentence, the relevance of the context sentences cannot be solved, the document translation technology can solve the relation between the sentences, and the translation accuracy is improved.

The client side is provided with a translation interface, a selection box for uploading texts is arranged on the translation interface, a user can pop up a text to open a dialog box by clicking the selection box, and selects an original text to be translated in the dialog box, so that the original text is uploaded.

Specifically, after the user clicks an upload button or a mouse dragging action is completed, the client of the front-end device receives the upload instruction or the mouse dragging action, the client acquires an address path of the document, judges whether the instruction is a valid instruction or not and whether the instruction is an identifiable original text or not, acquires the original text under the address path, encrypts the original text, and sends the encrypted original text to the server. The server decrypts and converts the format of the document, and separates character parts from the original text in the format conversion process, such as characters in a table, characters in a picture, characters in a thought map, horizontal and vertical characters, formulas and the like, so as to obtain the Chinese character part in the original text. For format conversion, the client performs format conversion to obtain the original text with the format recognizable by the translation system, then encrypts the original text with the format recognizable and sends the encrypted original text to the server, and the server directly obtains the original text with the format recognizable.

Specifically, after a user clicks an upload button or drag and drop actions such as mouse drag, touch drag, action gesture drag and the like are completed, a client of the front-end device receives an upload instruction or the drag and drop action, the client acquires address path information of an original text, judges whether the address path information is a valid instruction or not, acquires format information of the original text, judges whether the text is a recognizable text or not, acquires the size and the number of characters contained in the original text, judges whether the translation can be performed or not, outputs prompt information to prompt the user to upload the original text again under the condition that any judgment condition is not met, acquires the original text under the address path under the condition that all judgment conditions are met, encrypts the original text, and transmits the encrypted original text to a back-end server. The back-end server decrypts the original text and separates each part of content, such as the text content of each paragraph, the text content of the table, the text content of the picture, the text content of the mind map, the formula, etc., from the original text.

The method for translating the original text by the back-end server after acquiring the original text may specifically be: and inputting each sentence in the original text into a pre-trained machine translation model, and outputting a translation of each sentence by the machine translation model. The machine translation model can be a deep learning model, the training mode of the deep learning model is based on a large number of corpus samples, the corpus samples are input into a first network model selected in advance, the output of the first network model is compared with the nominal translation of the corpus samples to obtain a loss value, if the loss value is larger than or equal to a preset threshold value, parameters of the first network model are adjusted, a corpus sample is input again, until the obtained loss value is smaller than the preset threshold value, the model training can be determined to be finished, and the current first network model is the machine translation model. The language material sample can be a sentence, a word, a paragraph or an article, the nominal translation of the language material sample refers to the standard translation of the set language material sample, and the first network model is any pre-selected deep learning model capable of realizing the translation function.

In practical application of the embodiment of the present invention, since the output is a translation of the whole text, semantic association needs to be performed during translation, that is, when a machine translation model is used for translation, the translated sentence needs to be taken as a reference input, and translation of the sentence needs to refer to the translated sentence, that is, context translation semantic association, so as to ensure that translations of the same words and sentences in the same text are consistent, and to maintain connectivity and continuity between sentences. The machine translation model is hereinafter referred to as a translation model.

In one implementation of the embodiments of the invention, the first network model may include a first encoder, a second encoder, and a decoder. The first encoder comprises an input port and an output port, wherein the input port is used for inputting a target object when the translation model is trained, and the input port is used for inputting the target object or a specified object when the translation model is translated into a sentence; the output port is used for outputting a first encoding vector obtained by encoding the target object or the specified object. The second encoder comprises two input ports and an output port, wherein when the translation model is trained, a first input port in the second encoder is used for inputting training sentences, and when the translation model is translated, a first input port in the second encoder is used for inputting sentences to be translated; a second input port in the second encoder is used for inputting the first coding vector; the output port is used for outputting the second coding vector. The decoder also comprises two input ports and an output port, wherein a first input port in the decoder is used for inputting the target statement, a second input port in the decoder is used for inputting the second coding vector, and the output port is used for outputting the coding vector.

The process of training the first network model of the structure mainly comprises the following steps: the method comprises the steps of firstly, receiving a training sample, wherein the training sample comprises a corpus sample, a nominal translation corresponding to the corpus sample and a target object corresponding to the corpus sample, and the target object comprises a target prefix and/or a target suffix; secondly, inputting the target object into a first encoder to carry out encoding processing to obtain a first encoding vector; inputting the corpus samples and the first coding vector into a second coder for coding to obtain a second coding vector; fourthly, inputting the nominal translation and the second coding vector into a decoder for decoding processing to obtain a decoding vector, and calculating a loss value according to the decoding vector; and fifthly, adjusting the model parameters of the first network model according to the loss value, continuing to train the first network model until a training stopping condition is reached, and determining the current first network model as a translation model.

The first encoder comprises a first embedded layer and x object coding layers, wherein x is a positive integer greater than or equal to 1, and specifically, the first embedded layer, the first object coding layer and the xth object coding layer are connected in sequence. Therefore, the specific process of inputting the target object into the first encoder to perform encoding processing to obtain the first encoding vector may be: inputting a target object into a first embedding layer for embedding processing to obtain a target object vector; inputting the target object vector into a first object coding layer for coding to obtain a first object coding vector; inputting the first object coding vector into a second object coding layer for coding to obtain a second object coding vector; until the x-1 coding vector is input into the x object coding layer to be coded to obtain a first coding vector. Specifically, the embedding processing refers to a process of representing an object by a low-dimensional vector, for example, representing a word or an object by a vector, and the embedded vector has the property that objects corresponding to vectors with close distances can have close meanings, and the characteristic that the embedding can encode the object by the low-dimensional vector and can retain the meanings of the objects is very suitable for deep learning.

The second encoder comprises a second embedded layer and y statement encoding layers, y is a positive integer greater than or equal to 1, and specifically, the second embedded layer, the first statement encoding layer and the y statement encoding layer are sequentially connected. Therefore, the specific process of inputting the corpus samples and the first encoding vector into the second encoder to perform encoding processing to obtain the second encoding vector may be as follows: inputting the corpus sample into a second embedding layer for embedding processing to obtain a training statement vector, and inputting the training statement vector and a first coding vector into a first statement coding layer for coding processing to obtain a first statement coding vector; inputting the first statement vector into a second statement coding layer for coding to obtain a second statement coding vector; until the y-1 coding vector is input into the y statement coding layer for coding processing to obtain a second coding vector.

The decoder comprises a third embedded layer and z decoding layers, wherein z is a positive integer greater than or equal to 1, and specifically, the third embedded layer, the first decoding layer and the z decoding layer are sequentially connected. Therefore, the specific process of inputting the nominal translation and the second encoding vector into the decoder for decoding processing to obtain the decoding vector may be: inputting the nominal translation into a third embedding layer for embedding processing to obtain a nominal translation vector, and inputting the nominal translation vector and the second coding vector into a first decoding layer for decoding processing to obtain a first decoding vector; inputting the first decoding vector into a second decoding layer for decoding to obtain a second decoding vector; until the z-1 decoding vector is input into the z decoding layer to be decoded to obtain a decoding vector.

The way of calculating the loss value according to the decoding vector may specifically be: and comparing the decoded vector with a preset standard vector to obtain a loss value of the decoded vector. When the loss value of the decoding vector is calculated, the loss value is calculated by comparing the standard vector with the obtained decoding vector, but the loss value is not calculated by directly comparing the decoding vector with the nominal translation vector. Therefore, overfitting when the decoding vector is directly compared with the nominal translation vector to calculate the loss value is avoided, the performance of the translation model in other sentence translations is optimized, and the translation effect is more natural and smooth.

When adjusting the model parameters of the first network model, the model parameters of the first network model need to be adjusted according to the loss value, and the specific implementation process is as follows: and the loss value is reversely propagated to sequentially update the decoding parameters of the decoder, the coding parameters of the second coder and the coding parameters of the first coder.

In the training process, a target value can be preset, and when the loss value is lower than the preset target value, the model reaches the translation standard or a certain translation level, namely the training stopping condition is reached.

Through the improved translation model, the first encoder special for encoding the target object is added in the translation model, the auxiliary translation model can better translate the result with the appointed prefix and/or suffix, the fluency of the translation result is ensured under the condition that the appointed prefix and/or suffix are not lost by the translation model, and the translation result is more natural.

After the translation model is obtained through the training in the above manner, text translation can be performed by using the translation model, and the specific translation process is as follows: extracting a sentence to be translated and a specified object from an original text, wherein the object comprises a prefix and/or a suffix; inputting an object into a first encoder of a translation model for encoding to obtain a first encoding vector; inputting the first coding vector and the statement to be translated into a second coder of the translation model for coding to obtain a second coding vector; inputting the second coding vector into a decoder of the translation model for decoding to obtain a translation of the sentence to be translated; and splicing the translations of the sentences to be translated to obtain the translations of the original text. The specific splicing mode is to arrange the translations of the sentences correspondingly according to the sequence of the sentences in the original text.

In an implementation manner of the embodiment of the present invention, the method for translating the original text in S202 may specifically be: extracting each part of content in the original text according to a preset text content extraction strategy, and generating a translation filling template; sequentially inputting the extracted contents of each part into a pre-trained translation model to obtain a translation of each part; and filling contents by using the translated texts of the contents of all the parts based on the translated text filling template to obtain the translated texts of the original texts. The extraction policy may be a preset text content extraction policy.

The text that a user takes for machine translation is generally in various typesetting styles (fonts, paragraphs, table styles, artistic words, header and footer, comments, revision marks and the like) that are customary for creators (and also users), and also various formats of formula, graph, table, thinking guide, flow chart, video, audio file and other types of elements exist in a document. In order to meet the habit of a user and improve the user experience, the generated translation should be typeset according to the typesetting mode of the original text, therefore, in the embodiment of the invention, according to a preset text content extraction strategy, for example, according to paragraph extraction, sentence extraction, table extraction and the like, each part of content in a document is extracted, a translation filling template is generated, the translation filling template appoints which paragraph, which sentence or which row and column in a table the translation of each part of content should be filled with, the extracted parts of content are sequentially input into a pre-trained translation model to obtain the translation of each part of content, and then based on the translation filling template, content filling is performed by using the translations of each part of content to obtain the translation of the original text. The translation filling template defines the same typesetting style as the original text, thereby ensuring that the typesetting of the translation is the same as that of the original text.

Specifically, the step of extracting each part of content in the original text according to a preset text content extraction policy and generating a translation filling template may specifically be: and extracting the contents of each paragraph in the original text according to the paragraph division, and generating a translation filling template taking the paragraph as a unit.

In an implementation manner of the embodiment of the present invention, before performing S202, the method may further perform: acquiring a specified language selected by a user through a translation interface; and identifying the original language of the original text, or acquiring the original language of the original text selected by the user through the translation interface. Accordingly, S202 may specifically be: the method comprises the steps of translating an original text by utilizing a pre-trained translation model to obtain a specified language translation of the original text, wherein the translation model is a model which is obtained by utilizing a corpus sample of an original language and is used for translating the original language into the specified language.

After the original text is obtained, the back-end server may perform language identification on the original text, and the specific identification manner may be: the method includes the steps of randomly extracting a plurality of words from an original text, identifying the languages of the words, and determining that the original text is of the same language if the words exceeding a preset threshold belong to the same language, wherein other methods for identifying the language of the text, such as a neural network, also belong to the protection scope of the invention. In some special cases, a text may contain words of multiple languages, which may cause errors in language identification, for example, a document composed of a mixture of chinese and english, where english is abundant in randomly extracted words, the backend server identifies that the document is english, but actually the whole document is chinese, in order to avoid identification errors, a selection item of the language of the original text may be provided on the translation interface, the backend server identifies the language of the original text, sends the language type to the client, may cause the front-end device to display the identified language on the selection item, if the user finds that the language identification is incorrect, the user may intervene in selection, specifically, may select a correct language in a selection box of the selection item such as a drag-down box or a pop-up selection box, and after the user adjusts to a new language type, the backend server records the change, and preferentially adopting the new language type for translation when the machine translates. As a specific embodiment, if a situation that a user actively adjusts a language type occurs, the back-end server records the current change in a language adjustment log, where the language adjustment log records an original automatically recognized language type, a language type specified by the user, a word extracted at random this time, and a random function called in recognition, and the language adjustment log can be used for research and development personnel to improve the performance of the language automatic recognition module. Of course, the back-end server may not provide the language identification function, and the user may select the original language of the original text on the language selection item of the translation interface after uploading the original text. Under the condition that the user does not select the language, after receiving the original text, the back-end server can start the automatic identification module to automatically identify the original text, and translate the text document after the identification is finished; or the back-end server extracts each part of content in the original text after receiving the original text, then starts the automatic identification module to randomly extract words from each part of content in the original text, and then identifies the language type. The back-end server sends the original text to the language automatic identification module and the format conversion module for processing respectively so as to improve the processing efficiency, and the server enters the machine translation module after obtaining the contents of each part in the original text and the language type automatically identified.

The translation interface can also provide a translation language selection item, and a user can select a specified language in a selection box of the selection item, such as a pull-down box or a pop-up selection box, according to the requirement of the user. Because different translation models are needed for translation of different languages, when translation is performed, the pre-trained translation model is used for translating the original text to obtain a specified language translation of the original text, wherein the translation model is a model obtained by training a corpus sample of the original language and used for translating the original language into the specified language.

In practical application, there are more than two languages in some texts, so that when performing translation, a user can specify one language as a main language, and when performing machine translation, select a translation model corresponding to the main language, and translate the text content of the main language in the original text into a translation of a target language, while the content of other languages can keep the original text unchanged.

Naturally, in another implementation manner, translation models corresponding to different languages may be selected according to the languages contained in the original text, then the text contents of different languages are respectively input into the corresponding translation models, target language translations of each part of contents are respectively obtained, and then the translations of each part of contents are spliced according to the contents of the original text, so as to obtain the translations of the original text.

In an implementation manner of the embodiment of the present invention, before performing S202, the method may further perform: and identifying the domain type to which the original text belongs, or acquiring the domain type selected by the user through the translation interface. Accordingly, S202 may specifically be: and translating the original text by using a translation model corresponding to the field type to obtain a translation of the original text, wherein the translation model is obtained by training a corpus sample of the field type.

After the original text is obtained, the back-end server can identify the field type of the original text, and the specific identification mode can be as follows: inputting the original text or each part of the content in the original text into a pre-trained text classification model, and identifying the field type of the original text by the text classification model. The text classification model is a deep learning model, the training mode of the text classification model is based on a large number of corpus samples, the corpus samples are input into a second network model, the output of the second network model is compared with the domain category labels of the corpus samples to obtain a loss value, if the loss value is greater than or equal to a preset threshold value, parameters of the second network model are adjusted, a corpus sample is input again, until the obtained loss value is less than the preset threshold value, the model training can be determined to be finished, and the current network model is the text classification model. The second network model is any pre-selected deep learning model capable of realizing the text classification function.

In an implementation manner of the embodiment of the present invention, the second Network model may be an Induction Network model, and the training process of the text classification model mainly includes the following steps: the method comprises the steps of firstly, acquiring preset domain category labels, keywords corresponding to the domain category labels and a plurality of corpus samples; expanding the keywords corresponding to the category labels of each field by using the pre-training word vector; coding all the keywords and the corpus samples into vectorization expression to obtain keyword vectors and corpus sample vectors; classifying the corpus sample vectors based on the keyword vectors, and labeling all corpus samples according to the domain types based on classification results; and performing few-sample training on a preset induction network model based on the labeled corpus samples to obtain a text classification model.

The way of encoding all the keywords and corpus samples into vectorization representation specifically may be: and coding all the keywords and the corpus samples by adopting a pre-trained bidirectional coder BERT to obtain keyword vectors and corpus sample vectors.

Based on the keyword vector, the method for classifying the material sample vector may specifically be: and classifying the vector of the language material sample by adopting a K-neighborhood algorithm based on the keyword vector. The K-nearest neighbor algorithm (KNN) is a non-parametric statistical method for classification and regression. The method adopts a vector space model for classification, the concept is the cases of the same category, the similarity between the cases is high, and the possible classification of the cases of unknown category can be evaluated by calculating the similarity between the cases of known category.

The learning with few samples (few-shot learning) means that after a machine learning model learns a large amount of data of a certain category, a small amount of samples are needed for fast learning of a new category. The process of the few-sample training may specifically be: randomly extracting a first preset number of field category labels from the field category labels, and randomly selecting a second preset number of corpus samples with the field category labels for each extracted field category label to serve as a support set of a text classification model; randomly selecting a third preset number of corpus samples from the residual corpus samples with the domain category labels according to the extracted domain category labels as prediction objects; inputting the corpus samples in the support set into a preset induction network model to obtain the field category of each corpus sample; calculating loss values between the field categories of the corpus samples obtained through the induction network model and the field categories of the corpus samples in the prediction object; if the loss value is larger than or equal to the preset threshold value, adjusting network parameters of the induction network model, returning to execute the step of randomly extracting a first preset number of field type labels from all field type labels, and randomly selecting a second preset number of corpus samples with the field type labels for each extracted field type label to serve as a support set of the text classification model; and determining the current induction network model as a text classification model until the loss value is smaller than a preset threshold value.

The sensing network model may specifically include: an encoder, a compression network, and a relationship network; the encoder is configured to encode the corpus samples into training sample vectors, and may be composed of a Bi-directional Long Short-Term Memory (Bi-directional Long Short-Term Memory) and an Attention model, where a compression Network (also referred to as a Capsule Network) is configured to convert the training sample vectors into class-level vectors, and a relationship Network (also referred to as a relationship Network) is configured to map a relationship between the class-level vectors and the domain categories.

When a new field type label is received, keywords corresponding to the new field type label can be set, and then the keywords corresponding to the new field type label are expanded by using the pre-training word vector; crawling the corpus sample by using the expanded keywords, and performing field category labeling on the crawled corpus sample; and taking the newly labeled corpus sample as a new sample, and performing less sample training on a preset induction network model to obtain a text classification model.

BERT stands for transform's bi-directional encoder. It is designed to pre-train unlabeled text to a deep bidirectional representation through the union of left and right contexts. Therefore, only one additional output layer is needed to fine-tune the pre-trained BERT model, thereby creating SOTA results for various NLP (Natural language processing) tasks. The Chinese bert pre-training model comprises simplified Chinese characters and traditional Chinese characters, 12 layers in total, 768 hidden units, 12 Attention heads and 110M parameters.

The pre-training word vector is a general term of a language model and a characterization learning technology in NLP. Conceptually, it refers to embedding a high-dimensional space with dimensions of the number of all words into a continuous vector space with much lower dimensions, each word or phrase being mapped as a vector on the real number domain.

Keyword extraction refers to an automated technique that identifies meaningful and representative segments or words. The keyword extraction technology used in the work is a TWE (local Word Embedding) topic model.

The text classification model training adopts an unsupervised classification mode to construct the labeled corpus samples, then model training is carried out by utilizing less sample learning, the unsupervised classification mode enables the corpus samples not to be labeled with the field types in advance, the workload of manual labeling is reduced, and less sample learning ensures that after a large amount of data of a certain category is learned, the corpus samples can be rapidly learned by only a small amount of samples for a new category, so that the efficiency of model training is improved.

The translation of words and sentences in different fields is different, such as "viruses (translatable into claims and statements)", "turnkey (translatable into turkeys and turkeys)". Therefore, different translation models correspond to different fields, and when translation is performed, the original text is translated by using the translation model corresponding to the field type to obtain a translation of the original text, wherein the translation model is obtained by training the corpus sample of the field type. The specific translation model training process is to input the corpus samples in the field into the first network model for training.

In another implementation manner of the embodiment of the present invention, the language of the original text may be identified first, a language translation model library is determined based on the language, the language translation model library has domain category translation models in multiple domains of the language type, the domain category of the original text is identified, and based on the domain category, a corresponding domain category translation model is selected from the language translation model library to translate the original text. Of course, the domain category of the original text may be recognized first, the translation model library is determined based on the domain category, the translation model library has translation models of different languages, then the language of the original text is recognized, and based on the language and the language of the translation, the corresponding translation model is selected from the translation model library to translate the original text. Specifically, the domain category may be information, novel, law, finance, medical treatment, treatise, e-commerce, military, politics, patent, IT, machinery, geography, traffic, chemical engineering, and construction, and in another implementation, the domain category may also be a vertical domain category, where one definition of a vertical domain refers to a small domain vertically and longitudinally subdivided in a large domain, for example, the domain category is a patent, and the vertical domain of the patent domain category is a patent in the machine translation technology field; one vertical field subdivided into medical field categories may be the health preserving field. Specifically, a translation model library can be formed by storing a novel field translation model from English to Chinese and a novel field translation model from Chinese to English at the back-end server, and by analogy, the translation models in different fields of different languages are stored.

In practical application, the text content in some texts belongs to more than two different field types, so that when performing translation, a user can specify one field type as a main field, and when performing machine translation, a translation model of a corresponding point of the field type is selected to translate the original text. Of course, in another implementation manner, translation models corresponding to different domain categories may be selected according to the domain categories to which different text contents in the original text belong, then the text contents of the different domain categories are respectively input into the corresponding translation models, translations of each part of content are respectively obtained, and then the translations of each part of content are spliced according to the content of the original text, so that the translation of the original text is obtained.

After the back-end server translates the translation of the original text, the translation is transmitted to the front-end device, and the front-end device displays the translation on the translation interface, for example, in a JAVA programming Language environment, the back-end server may specifically convert the translation into an HTML (HyperText Markup Language) format by using Jodconvert and transmit the converted translation to the front-end device for display. In practical application, after the back-end server translates the translation of the original text, the back-end server can encrypt the translation, and only a user who uploads the original text can view the translation. As a specific embodiment, certain authority can be set to ensure document confidentiality, that is, a background worker does not have the authority to view the translation according to different set authorities, or machine management operation can be performed on desensitized or fragmented translations. Wherein, Jodconvert refers to a document conversion tool in JAVA, and performs document format conversion by using OpenOffice (a source office suite) or LibreOffice (a derivative of OpenOffice).

In specific translation, different languages have different requirements on translation formats, for example, words of european languages are separated by spaces, and words of east asian languages are not separated by spaces, so that a space character needs to be added between words when european languages (such as english, french, and the like) are translated, and east asian languages (such as chinese, thai, and the like) can be directly translated.

Based on the embodiment shown in fig. 2, a text translation method provided by the embodiment of the present invention, as shown in fig. 3, includes the following steps.

S301, obtaining the original text uploaded by the user through the translation interface provided by the front-end equipment.

S302, identifying the attribute information of the original text, judging whether the original text is in a processable format type according to the attribute information, if so, executing S304, and if not, executing S303.

The attribute information of the original text includes information such as text format, format information, path, number of characters, and the like. The attribute information for identifying the original text may be a text format for identifying the original text. The text format can be txt, doc/docx, xls/xlsx, ppt/pptx, pdf, etc.

The server may pre-store a processable format type list, and the method for determining whether the text format is a processable format type may be to compare the extension name, determine whether the text format is a format in the format type list, if yes, execute step S304; if not, the step S303 is executed.

S303, converting the text format of the original text into a processable format.

Specifically, the format conversion module participates in format conversion operation, the server sends the original text to the format conversion module, and the format conversion module converts the original text into a processable format. For example, the original text input to the format conversion module is in a jpg format (a picture format), and the original text in the jpg format is converted into a processable PDF format by using a conversion algorithm for converting the picture format into the PDF format.

S304, translating the original text to obtain a translated text of the original text.

S305, transmitting the translation to the front-end equipment so that the front-end equipment displays the translation on the translation interface.

Generally, a back-end server can only translate texts in general formats such as WORD, PDF, etc., and in order to meet the text translation requirements of a user, if an original text uploaded by the user is not in a format processable by the back-end server, such as a picture format, a table, a thought chart, a flow chart, etc., the text format of the original text needs to be converted into a processable format, and then the original text needs to be translated.

S301, S304, and S305 are the same as the steps in the embodiment shown in fig. 2, and how to implement the steps is specifically shown in fig. 2.

Based on the embodiment shown in fig. 2, the text translation method provided by the embodiment of the present invention, as shown in fig. 4, includes the following steps.

S401, obtaining the original text uploaded by the user through the translation interface provided by the front-end equipment.

S402, translating the original text to obtain a translated text of the original text.

And S403, transmitting the translation to the front-end equipment so that the front-end equipment displays the translation on the translation interface.

S404, if the user selects the translation editing mode through the translation interface, setting the translation into an editing state for the user to edit the content of the translation.

In the specific embodiment of the present invention, after a user selects to translate a certain original text, the client jumps to a translation display page, where the translation display page is provided with at least two columns of display contents, one column is used to display the original text, and the other column is used to display a translation, that is, after receiving the translation transmitted by the back-end server, the front-end device may display the original text in the original display column and the translation in the translation display column; and a mode selection control is arranged on the translation display page, the mode selection control can provide whether the user performs comparison reading or translation editing, the default mode is the comparison reading mode, if the user selects the translation editing mode, the client sends a trigger instruction to the back-end server, the back-end server sets the translation into an editing state after receiving the trigger instruction, and in the editing state, the client can acquire the position change and clicking action of a mouse pointer of the user in real time, and the content of the translation can be edited and modified by the user.

When the two columns of display contents are displayed on the translation display page, in order to prompt the user of the corresponding relationship between the contents in the original text and the contents in the translation and facilitate the user to quickly locate the contents of the original text and the corresponding contents of the translation, the embodiment of the invention can provide automatic highlighting display, and the specific realization of the automatic highlighting display comprises the following steps: (1) the back-end server can count the translation frequency of each sentence, count how many times each sentence has been translated repeatedly, determine sentences of which the translation frequency exceeds a certain threshold, then send original text identifiers (which can be positions of the sentences in the original text, keywords in the sentences, and the like) and translated text identifiers (which can be positions of the sentences in the translated text, keywords of the sentences, and the like) of the sentences to the front-end equipment, and the front-end equipment performs highlight setting such as color base yellow, character color red, underline increase, wavy line increase, and the like on corresponding sentences in the original text and the translated text according to the original text identifiers and the translated text identifiers of the sentences; (2) the method comprises the following steps that a user selects a certain word in an original text or a translated text by placing a mouse, clicking the mouse or sliding an area, the front-end equipment identifies the word selected by the user and then sends a word identifier selected by the user to a rear-end server, the rear-end server determines the corresponding positions of the selected content in the original text and the translated text according to the word identifier selected by the user, then the determined position information is fed back to the front-end equipment, and the front-end equipment performs highlight setting such as bottom color yellow, character color red, underline or wavy line increase on the corresponding words in the original text and the translated text according to the position information; (3) the user selects a text in the original text or a certain sentence in the translated text by sliding the area, alternatively, the user may select a word or a word in a certain sentence by placing a mouse or clicking a mouse, then the front-end equipment identifies the sentence between two end symbols to which the character or word belongs through symbol identification, after identifying that a certain sentence is selected by the user, the sentence mark selected by the user is sent to a back-end server, the back-end server determines the corresponding positions of the selected content in the original text and the translated text according to the sentence mark selected by the user, and then feeding back the determined position information to the front-end equipment, and carrying out highlight setting such as base color yellow, character color red, underline or wavy line on corresponding sentences in the original text and the translated text by the front-end equipment according to the position information. (4) The client receives and stores the one-to-one corresponding relation between the translated text and the original text of each word in the original text sent by the back-end server, judges the word pointed by the mouse point position according to the position of the display interface under the condition that the client detects that a user places a mouse, clicks the mouse or slides the position of the area, finds the word or the sentence where the word is located, highlights the position where the word or the sentence is located, and highlights the translated text or the original text corresponding to the word or the sentence. That is, if the word or sentence is from the original text, finding the position of the translation according to the corresponding relationship between the original text and the translation, and highlighting the translation with the corresponding relationship; similarly, if the word or sentence is from the translated text, the original text corresponding to the word or sentence is highlighted at the same time.

Because the back-end server performs text translation based on the translation model, translated sentences may be translated wrongly or may not reach the language grammar desired by the user, and different users may have different translation habits, the invention also provides the function of interaction between the user and the machine translation, helps the user and the machine to know each other, and the translation interface of the front-end equipment also provides the function of editing translated text. Specifically, a translation editing option button can be provided on a translation page, and a user can start the translation editing module to enter a translation editing mode by clicking the translation editing option button, or the user can switch to the translation editing mode by double-clicking or single-clicking the translation on a translation interface. After the translation is set to be in the editing mode, the user can edit the content of the translation directly in the translation, or pop up an editing box to edit the content of the translation in the editing box, after the editing is finished and the user confirms, the editing box is closed, the translation is correspondingly modified into the edited translation, and particularly, when the user finishes modifying the translation somewhere, the similar words or sentences of the translation can be modified into the translation changed by the user all together.

In one implementation manner of the embodiment of the present invention, after performing S404, the method may further include: after identifying the content to be edited selected by the user from the translated text, determining the original content corresponding to the content to be edited, and sending a highlight instruction to the front-end equipment to enable the front-end equipment to highlight the content to be edited in the translated text and the original content in the original text, wherein the mode that the user selects the content to be edited from the translated text comprises the following steps: single click selection, double click selection, or sliding region selection.

The user can select the content to be edited from the translated text by means of single-click selection, double-click selection or sliding area selection, in order to make the user more intuitively observe which sentence is edited, the front-end device recognizes that after the user selects a certain sentence from the translation, the sentence mark selected by the user (which may be the position of the sentence in the translation, the keyword in the sentence, etc.) is sent to the back-end server, and the back-end server determines the position of the sentence in the original text according to the sentence mark selected by the user, then the determined position information is carried in a highlight instruction and fed back to the front-end equipment, after the front-end equipment receives the highlight instruction, and according to the position information carried in the text, carrying out high-brightness setting such as bottom color code yellow, character color red, underline or wavy line increasing on the corresponding sentence in the original text. In this way, the user can clearly see the contents to be edited in the translated text which the user wants to edit and the corresponding original text contents in the original text.

In one implementation manner of the embodiment of the present invention, after performing S404, the method may further include: and under the condition that the content to be edited selected by the user from the translated text is identified, sending an edit box setting instruction to the front-end equipment so as to enable the front-end equipment to display an edit box at the specified position of the translation interface or pop up the edit box, wherein the position and the size of the edit box can be adjusted, and the user can edit the content to be edited in the edit box.

After a user selects contents to be edited from the translated text, an edit box setting instruction can be sent to the front-end equipment, the front-end equipment is instructed to display an edit box or pop up the edit box at a specified position of a translation interface, the user can edit the contents to be edited in the edit box, under the condition that the edit box or the edit box is popped up, the front-end equipment can also set a character pattern of 'please edit' in a receiving and editing input area in the edit box so as to prompt the user to carry out editing operation, and after the user finishes editing, a confirmation button is clicked, so that the rear-end server can update the translated text based on the contents edited by the user. The method comprises the steps that when an edit box setting instruction is sent to front-end equipment by a back-end server, sentence marks selected by a user can be carried in the edit box setting instruction, the front-end equipment can extract sentences selected by the user from a translation according to the sentence marks, and then the sentences selected by the user are displayed in an edit box through display setting.

In one implementation manner of the embodiment of the present invention, after performing S404, the method may further include: after receiving the translation edited by the user, re-translating the sentence to which the editing content of the user belongs based on the editing content of the user, and transmitting the translation of the sentence to the front-end equipment, so that the front-end equipment recommends the translation of the sentence to the user for selection.

After a user edits certain words, the user usually only concerns the editing of certain words and neglects whether the whole sentence is smooth or not, in order to ensure the smooth of the edited translated text, the back-end server can re-translate the sentence to which the edited content of the user belongs based on the edited content of the user, and transmit the translated text of the translated sentence to the front-end equipment, the front-end equipment can display the translated text of the sentence, and after seeing the translated text, the user can select whether to replace the original translated text according to the self requirement.

S401-S403 are the same as the steps in the embodiment shown in FIG. 2, and how to implement the steps is shown in FIG. 2.

Based on the embodiment shown in fig. 2, a text translation method provided by the embodiment of the present invention, as shown in fig. 5, includes the following steps.

S501, obtaining the original text uploaded by the user through the translation interface provided by the front-end equipment.

S502, if the text content of the user-defined term base exists in the original text, translating the corresponding text content in the original text according to the translation rule in the user-defined term base, and translating other text content in the original text according to the general translation rule to obtain the translation of the original text.

In another specific embodiment, each part of the text content in the original text is translated according to the general translation rule, if the text content of the user-defined term base exists in the original text, the corresponding text content in the translated text is replaced according to the translation rule in the user-defined term base, and then the translated text meeting the user-defined rule is obtained.

And S503, transmitting the translation to the front-end equipment so that the front-end equipment displays the translation on the translation interface.

In practical application, a user has a special agreement for translation of some words, for example, an english translation of an application program is APP, but the user wants to translate the words into yingyongchengxu, the user can establish a custom term library according to the special requirements, various special translation rules are agreed in the custom term library, the user can upload the custom term library to a front-end device through a translation interface, the front-end device transmits the custom term library to a back-end server, the back-end server obtains an original text, compares the original text word by word with the custom term library to check whether text contents of the custom term library exist in the original text, if so, translates corresponding text contents in the original text according to the translation rules in the custom term library, and inputs other text contents into a translation model to obtain a translation of other text contents, and integrating the translated text translated based on the user-defined term library and the translated text translated by the translation model to obtain the translated text of the original text. Wherein the integration comprises splicing.

In the embodiment of the invention, the functions of historical translation task query and custom term library query can be provided on the translation interface, and a user can check which translation tasks have been done and which custom term libraries have been uploaded. Specifically, the front-end device records a corresponding translation task after receiving a certain translation transmitted by the back-end server, such as recording the text name of the original text in the personal center of the translation interface (obtained when the original text is received), the time to complete the translation (obtained when the translation transmitted by the back-end server is received), etc., and the back-end server can store the original text, the translated text, the contrast text and the like which finish the translation task, the back-end server may also send a link to the saved text corresponding to the translation task to the front-end device, and when the front-end device records the translation task, the text links of the original text, the translated text and the contrast text can be recorded at the same time, so that when a user inquires a translation task, the corresponding original text, translated text and contrast text can be directly downloaded by clicking the link; the front-end equipment can note self-defined term bank information after receiving a certain self-defined term bank uploaded by a user, for example, note the name of the self-defined term bank in the personal center of a translation interface, the uploading time of the self-defined term bank, and the like, and the back-end server can save the self-defined term bank after receiving the self-defined term bank uploaded by the front-end equipment, then the back-end server can also send the link of saving the self-defined term bank to the front-end equipment, the front-end equipment can record the link of the self-defined term bank simultaneously when recording the self-defined term bank information, and thus, when the user inquires the self-defined term bank, the user can directly download the self-defined term bank by clicking the link. In addition, corresponding search operation can be provided, that is, the personal center of the translation interface can provide search and add-delete controls of the translation task and the user-defined term library, the user can input the name of the translation task to be searched and the name of the user-defined term library in the search controls and click the search, the front-end equipment can search the corresponding translation task and the user-defined term library information according to the name input by the user, and then the searched translation task and the user-defined term library information are displayed on the translation interface.

S501 and S503 are the same as S201 and S203 in the embodiment shown in fig. 2, and how to implement the embodiments is shown in fig. 2.

Based on the embodiment shown in fig. 2, the text translation method provided by the embodiment of the present invention, as shown in fig. 6, includes the following steps.

S601, obtaining the original text uploaded by the user through the translation interface provided by the front-end equipment.

S602, translating the original text to obtain a translated text of the original text.

S603, transmitting the translation to the front-end device, so that the front-end device displays the translation on the translation interface.

S604, after the fact that the user selects quality inspection through the translation interface is recognized, selecting a plurality of sentences according to the sequence that translation scores of the sentences are from low to high, and instructing the front-end equipment to highlight the plurality of sentences so as to prompt the user to perform key inspection on the plurality of sentences.

The translation interface can also provide a quality check option, after the user clicks the quality check option, the front-end device can send a quality check instruction to the back-end server, after the back-end server receives the quality check instruction, a plurality of sentences can be selected from the translated text according to the sequence from low to high of the translation scores of the sentences, then sentence identifications (such as the positions of the sentences in the translated text, keywords in the sentences and the like) of the sentences are carried in the highlight instruction and fed back to the front-end device, after the front-end device receives the highlight instruction, the front-end device carries out highlight setting such as bottom color code yellow, character color red, underline or wavy line increase on the corresponding sentences in the translated text according to the carried sentence identifications, and therefore, the user can visually see which sentences have lower translation scores, and can carry out key check on the sentences.

In the embodiment of the present invention, the manner of performing translation scoring on the sentences in the translated text may specifically be: inputting each sentence in an original text uploaded by a user through a translation interface into a pre-trained translation model to obtain a plurality of translations of the sentence and the translation accuracy of each translation; and determining a translation score of the sentence based on the translation accuracy of each translation.

For each sentence in an original text, a translation model generates a plurality of translations, each translation has a translation accuracy, when the translation is determined, the translation with the highest accuracy can be selected, after the translation accuracy of each translation is obtained, the translation score of the sentence can be determined according to a certain scoring strategy based on the translation accuracy of each translation, for example, if the translation accuracy of each translation of the sentence is less than a preset threshold, a lower score can be set, and if the translation accuracy of each translation of the sentence exceeds the preset threshold, a higher score can be set.

Based on the embodiment shown in fig. 2, a text translation method provided by the embodiment of the present invention, as shown in fig. 7, includes the following steps.

S701, obtaining the original text uploaded by the user through a translation interface provided by the front-end equipment.

S702, translating the original text to obtain a translated text of the original text.

And S703, transmitting the translation to the front-end equipment so that the front-end equipment displays the translation on the translation interface.

And S704, after identifying that the user selects to download the text through the translation interface, identifying a download mode selected by the user through the translation interface, wherein the download mode comprises translation text download and comparison text download. If the download mode selected by the user is the downloading of the translation text, S705 is executed, and if the download mode selected by the user is the downloading of the comparison text, S706 is executed.

S705, generating a translation text according to the translation and a specified text format, and sending the translation text to the front-end equipment for downloading by the user.

S706, based on a preset text content division strategy, adding the translated text of each part of the original text of the translated text after each part of the original text is divided, and generating a contrast text according to a specified text format after the addition is finished; and sending the comparison text to the front-end equipment for downloading by the user.

The translation interface is further provided with a text downloading option, after the user clicks the text downloading option, a dialog box pops up to enable the user to select a downloading mode, the downloading mode comprises translation text downloading and comparison text downloading, and the default downloading mode is translation text downloading. When a user selects a translation text downloading mode, the front-end equipment identifies that the downloading mode selected by the user is translation text downloading, a translation text downloading instruction is sent to the back-end server, after the back-end server receives the instruction, a translation text can be generated according to the translation and a specified text format (such as WORD, PDF and the like), and then the translation text is sent to the front-end equipment, so that the user can download the translation text. When the user selects the comparison text downloading mode, the front-end device identifies that the downloading mode selected by the user is comparison text downloading, a comparison text downloading instruction is sent to the back-end server, after the back-end server receives the instruction, the back-end server adds the translated text of the part of the content after each part of the content of the original text of the translated text is divided based on a preset text content division strategy (for example, division according to paragraphs), after the addition is completed, the comparison text is generated according to a specified text format, and then the comparison text is sent to the front-end device, so that the user can realize the comparison text downloading.

In summary, the overall process of text translation by the translation system is shown in fig. 8, a user accesses the platform, uploads an original text on the translation interface shown in fig. 9, selects a field and a language, clicks on immediate translation, starts an internal translation process of the back-end server, and the back-end server first performs text type recognition, reads text contents, inputs a translation model sentence by sentence for translation, then integrates and stores a translated text, and performs original translation comparison display after completing translation. The user selects the comparison reading mode or the post-translation editing mode, as shown in fig. 10, the comparison reading mode is normally the default mode, the sentence to which the word belongs is subjected to whole sentence machine translation based on the word modified by the user in the post-translation editing mode, and the editing result is saved after the user confirms the modification. The user clicks the download translation button, selects a translation download mode (download of a translation text or download of an original translation comparison text) in a pop-up dialog box, finishes the translation work of the whole text after the download is finished as shown in fig. 11, and can manage historical translation tasks in the personal center.

The method for translating by inputting the translation model sentence by sentence specifically comprises the following steps: selecting a translation model meeting the domain type and the language according to the domain type and the language of the original text, then sequentially inputting each sentence in the original text into the selected translation model for translation to obtain a translation of the sentence, and for the sentences which are not in the selected language, the translation model can not translate the sentences, so that the original language is kept unchanged for the sentences in the translation.

The method for integrating the translated text specifically comprises the following steps: and typesetting the corresponding translated text according to the typesetting sequence of each sentence and each section in the original text, typesetting the translated text according to the rows and columns of the original table for tables and the like, and obtaining the translated text of the original text through typesetting and integrating.

After obtaining the translation, the back-end server may generate a translation text according to a specified text format, add the translation of the segment under each segment of the original text according to a segmentation mode in the original text, generate a comparison text according to the specified text format, and then store both the translation text and the comparison text in the local database of the front-end device, so that a subsequent user can quickly find a corresponding format text for the user to download after initiating a download instruction. The front-end device may be a plurality of front-end devices.

The presentation mode of the original translation contrast display can be as follows: (1) setting two columns of display frames in a translation interface, wherein one column of display frame displays an original text and the other column of display frame displays a translated text; (2) setting a column of display frame in the translation interface, and displaying the column of display frame according to a mode of one line of original text and one line of translated text; (3) and setting a column of display frame in the translation interface, and displaying the column of display frame according to a section of original text and a section of translated text.

The translation process of the back-end server to the original document is as shown in fig. 12, firstly, the text format needs to be identified, whether the text format is a processable text format is judged, if not, the text format is converted, if so, the paragraph of the original text is marked, the text is extracted according to the paragraph, a text skeleton is generated, the text skeleton is a template filled with the translation, the translation model is used for performing the paragraph-by-paragraph translation, a translation result is generated, the translation result is used for filling the text skeleton, the text skeleton is converted into HTML, the translation text is generated, and finally, the translation text is stored.

The architecture of the translation system is shown in fig. 13, and includes a front end and an external request (for receiving a translation request of a user, where the translation request includes an original text to be translated), authority management (for managing authority of an operable translation, that is, setting which persons can view and edit the translation), a configuration management center (for managing system configuration), a gateway, and a public application (for managing and setting the gateway, the public application, and the like.

In order to deal with the problem that a user often needs to download a translation after text translation is performed, and a download service is not provided in the current translation software and translation website, an embodiment of the present invention provides a text download method, as shown in fig. 14, which includes the following steps.

And S1401, receiving a downloading instruction initiated by a user through a translation interface provided by the front-end equipment.

S1402, obtaining a translation corresponding to the downloading instruction.

And S1403, identifying a downloading mode selected by the user through the translation interface and contained in the downloading instruction, wherein the downloading mode comprises translation text downloading and comparison text downloading. If the download mode selected by the user is the translation text download, then S1404 is performed, and if the download mode selected by the user is the comparison text download, then S1405 is performed.

And S1404, generating a translation text according to the translation and a specified text format, and sending the translation text to the front-end equipment for downloading by the user.

S1405, based on a preset text content division strategy, adding the translated text of the part of content after each part of content of the original text division of the translated text, and generating a contrast text according to a specified text format after the addition is finished; and sending the comparison text to the front-end equipment for downloading by the user.

The translation interface of the front-end equipment is provided with a text downloading option, and after clicking the text downloading option, a dialog box pops up to enable a user to select a downloading mode, wherein the downloading mode comprises translation text downloading and comparison text downloading. When a user selects a translation text downloading mode, the front-end equipment identifies that the downloading mode selected by the user is translation text downloading, a translation text downloading instruction is sent to the back-end server, after the back-end server receives the instruction, a translation text can be generated according to the translation and a specified text format (such as WORD, PDF and the like), and then the translation text is sent to the front-end equipment, so that the user can realize the translation text downloading. When the user selects the comparison text downloading mode, the front-end device identifies that the downloading mode selected by the user is comparison text downloading, a comparison text downloading instruction is sent to the back-end server, after the back-end server receives the instruction, the back-end server adds the translated text corresponding to each part of the original text of the translated text based on a preset text content division strategy (for example, division according to paragraphs), after the addition is completed, the comparison text is generated according to a specified text format, and then the comparison text is sent to the front-end device, so that the user can realize comparison text downloading.

After text translation is performed, users often pay attention to whether the translated text is translated accurately, and the full-text inspection of the users undoubtedly brings great workload to the users.

S1501, a quality check instruction is received. Wherein the quality check instruction is initiated by the user via a translation interface provided by the front-end device.

S1502, a plurality of sentences are selected from the translated text to be checked according to the sequence of translation scores of the sentences from low to high.

S1503, the front-end device is instructed to highlight the plurality of sentences to prompt the user to perform a focused check on the plurality of sentences.

The translation interface of the front-end device provides a quality check option, and after the user clicks the quality check option, the front-end equipment sends a quality inspection instruction to the back-end server, and after the back-end server receives the quality inspection instruction, a plurality of sentences are selected from the translation according to the sequence of the translation scores of the sentences from low to high, then, sentence identifications (such as the position of the sentence in the translation, keywords in the sentence, and the like) of the sentences are carried in the highlighting instructions and fed back to the front-end equipment, and after the front-end equipment receives the highlighting instructions, according to the sentence marks carried in the translation, the corresponding sentences in the translation are set with high brightness such as yellow bottom color code, red character color, underline or wavy line, in this way, the user can intuitively see which sentences in the translation have lower translation scores, and the sentences can be subjected to key inspection.

In an implementation manner of the embodiment of the present invention, a manner of performing translation scoring on sentences in a translated text may specifically be: inputting each sentence in the original text uploaded by the user through the translation interface into a pre-trained translation model to obtain a plurality of translations of the sentence and the translation accuracy of each translation; and determining the translation score of the sentence based on the translation accuracy of each translation.

For each sentence in an original text, a translation model generates a plurality of translations, each translation has a translation accuracy, when the translation is determined, the translation with the highest accuracy can be selected, after the translation accuracy of each translation is obtained, the translation score of the sentence can be determined based on the translation accuracy of each translation, for example, if the translation accuracy of each translation of the sentence is less than a preset threshold, a lower score can be set, and if the translation accuracy of each translation of the sentence exceeds the preset threshold, a higher score can be set.

Specifically, the method may further include: checking an original text of the translated text, and identifying error content in the original text; and instructing the front-end equipment to prompt the error content in the original text. The wrong content comprises the careless content of misspelled words, misgrammatical sentences, abnormal front and back fonts and the like.

The embodiment of the invention can also provide a quality inspection function of the original text, and by inspecting the original text, wrong sentences with wrong word spelling, semantic logic and the like in the original text are identified, and then sentence identifications (such as the positions of the sentences in the translated text, keywords in the sentences and the like) of the wrong sentences are carried in the highlight instruction and fed back to the front-end equipment, and after receiving the highlight instruction, the front-end equipment carries out highlight setting such as bottom color code yellow, character color red, underline or wavy line increase on the corresponding wrong sentences in the original text according to the carried sentence identifications.

Specifically, in the step of instructing the front-end device to prompt the error content in the original text, besides highlighting, a pop-up prompt may be provided to the user, or a right-click menu may be provided, and a plurality of recommended modification suggestions are displayed in the right-click menu.

In practical applications, a user has a special agreement on translation of some words, for example, an english translation of an application program is APP, but the user wants to translate the words into yingyongchengxu, which cannot meet the user requirement in the current translation method, and in order to meet the problem, the embodiment of the present invention provides a text translation method, as shown in fig. 16, which includes the following steps.

S1601, obtaining a text and a custom term library uploaded by a user through a translation interface provided by the front-end equipment.

S1602, identifying whether the text content in the custom term library exists in the original text.

S1603, if the translation rule exists, translating corresponding text contents in the original text according to the translation rule in the user-defined term library, and translating other text contents in the original text according to the general translation rule to obtain a translated text of the original text.

In the embodiment of the invention, a user can establish a self-defined term library according to special requirements, various special translation rules are appointed in the self-defined term library, the user can upload the self-defined term library to the front-end equipment through a translation interface, the front-end equipment transmits the self-defined term library to the back-end server, after the back-end server obtains the original text, firstly, comparing the words and words of the original text with the self-defined term library to check whether the text content of the self-defined term library exists in the original text, if so, according to the translation rule in the self-defined term library, translating corresponding text contents in the original text, inputting other text contents into the translation model to obtain translations of other text contents, and then splicing the translated text obtained by translation based on the user-defined term library and the translated text obtained by the translation model to obtain the translated text of the original text.

In practical applications, since machine-translated sentences are more mechanical, and different users may have different translation habits or have different translation requirements, the embodiment of the present invention provides a text translation editing method for better adapting to the user habits, as shown in fig. 17, which includes the following steps.

S1701, obtain an original text uploaded by a user through a translation interface provided by the front-end device, and translate the original text to obtain a translated version of the original text displayed on the translation interface.

S1702, under the condition that the user selects the translation editing mode through the translation interface, setting the translation into an editing state for the user to edit the content in the translation.

The translation editing function is provided on the translation interface of the front-end device, a translation editing option can be specifically provided on a page of a translation, and a user can perform a translation editing mode by clicking the translation editing option, or the user can switch to the translation editing mode by double clicking a word or the like. After the translation is set to be in the editing state, the user can edit the content of the translation directly in the translation, or pop up an editing box to edit the content of the translation in the editing box, after the user finishes editing, the translation is modified into the edited translation correspondingly under the condition that a confirmation button is clicked or the editing box is closed, and particularly when the translation is modified at a certain position, similar words or sentences of the translation are modified into the user-modified translation together throughout.

In an implementation manner of the embodiment of the present invention, after executing S1702, the method may further include: after identifying the content to be edited selected by the user from the translated text, determining the original text content corresponding to the content to be edited, and sending a highlight instruction to the front-end device, so that the front-end device highlights the content to be edited in the translated text and the original text content in the original text, wherein the mode for selecting the content to be edited by the user from the translated text comprises the following steps: single click selection, double click selection, or sliding region selection.

The user can select the content to be edited from the translated text by means of single-click selection, double-click selection or sliding area selection, in order to make it more intuitive for the user to observe that a sentence is edited, the front-end device recognizes that after the user selects a certain sentence from the translation, the sentence mark selected by the user (which may be the position of the sentence in the translation, the keyword in the sentence, etc.) is sent to the back-end server, and the back-end server determines the position of the sentence in the original text according to the sentence mark selected by the user, then the determined position information is carried in a highlight instruction and fed back to the front-end equipment, after the front-end equipment receives the highlight instruction, and according to the position information carried in the text, carrying out high-brightness setting such as bottom color code yellow, character color red, underline or wavy line increasing on the corresponding sentence in the original text. In this way, the user can clearly see the content to be edited in the translated text and the corresponding original text content in the original text which the user wants to edit.

In an implementation manner of the embodiment of the present invention, after executing S1702, the method may further include: after the content to be edited selected by the user from the translated text is identified, an edit box setting instruction is sent to the front-end equipment, so that the front-end equipment displays an edit box at a specified position of the translation interface or pops up the edit box, and the user can edit the content to be edited in the edit box. Wherein the position and size of the edit box can be adjusted, specifically, the display or pop-up position of the edit box has an initial default value, and preferably, the position and size of the edit box are set to be fixed relative to the display interface of the front-end device, so that when the user drags the scroll bar up and down or left and right in the editing process, the edit box can always be in a relatively fixed position and maintain a fixed size. After the fact that the user selects the editing frame and drag operation is carried out is recognized, the front-end device responds to the selection and drag operation of the user to adjust the position of the editing frame to the corresponding position, and the mode of selecting the editing frame optionally comprises clicking a non-editing area of the editing frame. After recognizing that the user selects the edit box and has a zooming operation, the front-end device adjusts the size of the edit box to a corresponding size in response to the selection and the zooming operation of the user, and the zooming operation can be selected by a zooming gesture, a zooming-in or zooming-out button is clicked, and the like.

After a user selects contents to be edited from the translated text, an edit box setting instruction can be sent to the front-end equipment, the front-end equipment is instructed to display an edit box or pop up the edit box at a specified position of a translation interface, the user can edit the contents to be edited in the edit box, when the edit box is popped up or displayed, the front-end equipment can further set characters of 'please edit' displayed in the edit box so as to prompt the user to carry out editing operation, and after the user finishes editing, a confirmation button is clicked, and the rear-end server can update the translated text based on the contents edited by the user. When the back-end server sends an edit box setting instruction to the front-end device, the edit box setting instruction can also carry a sentence identifier selected by the user, the front-end device can extract a sentence selected by the user from the translation according to the sentence identifier, and then the sentence selected by the user is displayed in the edit box through display setting.

In an implementation manner of the embodiment of the present invention, an input prompt text is displayed in an edit box, and the input prompt text is used for instructing a user to update the content to be edited according to the input prompt text;

the method may further comprise: and acquiring the text content input in the edit box by the user aiming at the input prompt text, and updating the translation.

In an implementation manner of the embodiment of the present invention, after executing S1702, the method may further include: after receiving the translation edited by the user, re-translating the sentence to which the editing content of the user belongs based on the editing content of the user, and transmitting the translation of the sentence to the front-end equipment, so that the front-end equipment recommends the translation of the sentence to the user for selection.

After a user edits certain words, the user usually only concerns the editing of certain words and neglects whether the whole sentence is smooth or not, in order to ensure the smooth of the edited translated text, the back-end server can re-translate the sentence to which the edited content of the user belongs based on the edited content of the user, and transmit the translated text of the translated sentence to the front-end equipment, the front-end equipment can display the translated text of the sentence, and after seeing the translated text, the user can select whether to replace the original translated text according to the self requirement.

Based on the foregoing method embodiment, an embodiment of the present invention provides a text translation apparatus, as shown in fig. 18, the apparatus may include:

an obtaining module 1810, configured to obtain an original text uploaded by a user through a translation interface provided by a front-end device;

the translation module 1820 is configured to translate the original text to obtain a translation of the original text;

the transmitting module 1830 is configured to transmit the translation to the front-end device, so that the front-end device displays the translation on the translation interface.

Optionally, the translation module 1820 may be specifically configured to: extracting each part of content in the original text according to a preset text content extraction strategy, and generating a translation filling template; sequentially inputting the extracted contents of each part into a pre-trained translation model to obtain a translation of each part; and based on the translation filling template, filling contents by using the translations of the contents of all the parts to obtain the translation of the original text.

Optionally, the translation module 1820 may be specifically configured to: and extracting the contents of each paragraph in the original text according to the paragraph division, and generating a translation filling template taking the paragraph as a unit.

Optionally, the apparatus may further include:

the identification module is used for identifying the attribute information of the original text and judging whether the original text is in a processable format type according to the attribute information;

translation module 1820 may be specifically configured to: if the recognition result of the recognition module is positive, translating the original text to obtain a translated text of the original text;

and the conversion module is used for converting the text format of the original text into a processable format if the identification result of the identification module is negative.

Optionally, the obtaining module 1810 may further be configured to: acquiring a specified language selected by a user through a translation interface; identifying the original language of the original text, or acquiring the original language of the original text selected by a user through a translation interface;

translation module 1820 may be specifically configured to: the method comprises the steps of translating an original text by utilizing a pre-trained translation model to obtain a specified language translation of the original text, wherein the translation model is a model which is obtained by utilizing a corpus sample of an original language and is used for translating the original language into the specified language.

Optionally, the obtaining module 1810 may further be configured to: identifying the domain type to which the original text belongs, or acquiring the domain type selected by a user through a translation interface;

translation module 1820 may be specifically configured to: and translating the original text by using a translation model corresponding to the field type to obtain a translation of the original text, wherein the translation model is obtained by training a corpus sample of the field type.

Optionally, the apparatus may further include:

and the setting module is used for setting the translation into an editing state if the user selects a translation editing mode through the translation interface, so that the user can edit the content of the translation.

Optionally, the setting module may be further configured to: after identifying the content to be edited selected by the user from the translated text, determining the original text content corresponding to the content to be edited, and sending a highlight instruction to the front-end device, so that the front-end device highlights the content to be edited in the translated text and the original text content in the original text, wherein the mode for selecting the content to be edited by the user from the translated text comprises the following steps: single click selection, double click selection, or sliding region selection.

Optionally, the setting module may be further configured to: after the content to be edited selected by the user from the translated text is identified, an edit box setting instruction is sent to the front-end equipment, so that the front-end equipment displays an edit box at a specified position of the translation interface or pops up the edit box, wherein the position and the size of the edit box can be adjusted, and the user can edit the content to be edited in the edit box.

Optionally, the translation module 1820 may be further configured to: after receiving the translation edited by the user, re-translating the sentence to which the editing content of the user belongs based on the editing content of the user, and transmitting the translation of the sentence to the front-end equipment, so that the front-end equipment recommends the translation of the sentence to the user for selection.

Optionally, the translation module 1820 may be specifically configured to: if the text content of the user-defined term base exists in the original text, translating the corresponding text content in the original text according to the translation rule in the user-defined term base, and translating other text content in the original text according to the general translation rule to obtain the translated text of the original text.

Optionally, the apparatus may further include:

and the indicating module is used for selecting a plurality of sentences according to the sequence of the translation scores of the sentences from low to high after the user selects the quality check through the translation interface, and indicating the front-end equipment to highlight the plurality of sentences so as to prompt the user to perform key check on the plurality of sentences.

Optionally, the apparatus further comprises:

the downloading module is used for identifying a downloading mode selected by the user through the translation interface after identifying that the user selects to download the text through the translation interface, wherein the downloading mode comprises translation text downloading and comparison text downloading; if the downloading mode selected by the user is downloading of the translation text, generating the translation text according to the translation and the specified text format, and sending the translation text to the front-end equipment for the user to download; if the downloading mode selected by the user is the comparison text downloading, adding the translated text of the part of content after each part of content of the original text of the translated text is divided based on a preset text content division strategy, and generating a comparison text according to a specified text format after the addition is finished; and sending the comparison text to the front-end equipment for downloading by the user.

By applying the embodiment of the invention, the back-end server translates the original text after obtaining the original text uploaded by the user through the translation interface provided by the front-end equipment to obtain the translation of the original text, then transmits the obtained translation to the front-end equipment, and the front-end equipment displays the translation on the translation interface. Therefore, a user only needs to upload the original text on a translation interface provided by the front-end device, the front-end device interacts with the back-end server, the back-end server translates the original text, the translation of the original text is displayed on the front-end device, the user does not need to input the translation word by word and sentence by sentence, translation time is greatly reduced, and rapid translation of the text is achieved.

An embodiment of the present invention further provides a text downloading device, as shown in fig. 19, where the device may include:

a receiving module 1910, configured to receive a download instruction initiated by a user through a translation interface provided by a front-end device;

an obtaining module 1920, configured to obtain a translation corresponding to the download instruction;

the identifying module 1930 is configured to identify a downloading mode selected by a user through a translation interface and included in the downloading instruction, where the downloading mode includes downloading of a translation text and downloading of a comparison text;

the processing module 1940 is configured to, if the download mode selected by the user is downloading of the translation text, generate the translation text according to the translation and a specified text format, and send the translation text to the front-end device for downloading by the user; if the downloading mode selected by the user is the comparison text downloading, adding the translated text of the part of content after each part of content of the original text of the translated text is divided based on a preset text content division strategy, and generating a comparison text according to a specified text format after the addition is finished; and sending the comparison text to the front-end equipment for downloading by the user.

By applying the embodiment of the invention, the translation downloading service is provided, and the downloading of the translation text or the downloading of the contrast text can be provided for the user according to the requirements of the user.

An embodiment of the present invention further provides a quality inspection apparatus, as shown in fig. 20, the apparatus may include:

a receiving module 2010, configured to receive a quality inspection instruction initiated by a user through a translation interface provided by a front-end device;

the selection module 2020 is used for selecting a plurality of sentences from the translated text to be checked according to the sequence of translation scores of the sentences from low to high;

an indicating module 2030, configured to instruct the front-end device to highlight the multiple sentences so as to prompt the user to perform a focused check on the multiple sentences.

Optionally, the apparatus may further include:

the translation module is used for inputting each sentence in the original text uploaded by the user through the translation interface into a pre-trained translation model to obtain a plurality of translations of the sentence and the translation accuracy of each translation;

and the scoring module is used for determining the translation score of the sentence based on the translation accuracy of each translation.

Optionally, the apparatus may further include:

the recognition module is used for checking the original text of the translated text and recognizing the error content in the original text;

the instruction module 2030 may further be configured to: and instructing the front-end equipment to prompt the error content in the original text.

By applying the embodiment of the invention, the translation interface of the front-end equipment provides a quality check option, and after a user clicks the quality check option, the back-end server can determine a plurality of sentences with the lowest translation scores and instruct the front-end equipment to highlight the plurality of sentences, so that the user can visually see that the translation scores of the sentences are lower, and can perform key check on the sentences. The workload brought by full-text inspection is reduced.

An embodiment of the present invention further provides a text translation apparatus, as shown in fig. 21, the apparatus may include:

the obtaining module 2110 is used for obtaining an original text and a custom term library uploaded by a user through a translation interface provided by front-end equipment;

the identification module 2120 is configured to identify whether text content in a custom term library exists in the original text;

the translation module 2130 is configured to, if the recognition module 2120 recognizes that the text exists, translate corresponding text contents in the original text according to a translation rule in the user-defined term library, and translate other text contents in the original text according to a general translation rule to obtain a translated text of the original text.

By applying the embodiment of the invention, a user can establish a self-defined term library according to special requirements, various special translation rules are appointed in the self-defined term library, if the text content of the user-defined term library exists in the original text, the corresponding text content in the original text is translated according to the translation rules in the self-defined term library, and other text contents in the original text are translated according to the general translation rules to obtain the translated text of the original text. Can meet the special requirements of users.

An embodiment of the present invention further provides a text translation editing apparatus, as shown in fig. 22, the apparatus may include:

the obtaining module 2210 is configured to obtain an original text uploaded by a user through a translation interface provided by a front-end device, and translate the original text to obtain a translation of the original text displayed on the translation interface;

the setting module 2220 is configured to, in a case that it is recognized that the user selects the translation editing mode through the translation interface, set the translation into an editing state, so that the user can edit the content in the translation.

Optionally, the apparatus may further include:

the first sending module is used for determining the original text content corresponding to the content to be edited after identifying the content to be edited selected by the user from the translated text, and sending a highlighting instruction to the front-end device so that the front-end device highlights the content to be edited in the translated text and the original text content in the original text, wherein the mode for selecting the content to be edited by the user from the translated text comprises the following steps: single click selection, double click selection, or sliding region selection.

Optionally, the apparatus may further include:

and the second sending module is used for sending an edit box setting instruction to the front-end equipment after identifying the content to be edited selected by the user from the translated text, so that the front-end equipment displays an edit box at the specified position of the translation interface or pops up the edit box, wherein the position and the size of the edit box can be adjusted, and the user can edit the content to be edited in the edit box.

Optionally, an input prompt text is displayed in the edit box, and the input prompt text is used for instructing a user to update the content to be edited according to the input prompt text;

the acquisition module 2210 may also be used to: and acquiring the text content input in the edit box by the user aiming at the input prompt text, and updating the translation.

Optionally, the apparatus may further include:

and the translation module is used for re-translating the sentence to which the user editing content belongs based on the user editing content after receiving the translation edited by the user, and transmitting the translation of the sentence to the front-end equipment so that the front-end equipment recommends the translation of the sentence to the user for selection.

An embodiment of the present invention provides an electronic device, as shown in fig. 23, including a processor 2301 and a memory 2302, where the memory 2302 is used for storing computer programs; the processor 2301, when executing the computer program stored in the memory, implements: any of the above methods.

The Memory may include a RAM (Random Access Memory) or an NVM (Non-volatile Memory), such as at least one disk Memory. Optionally, the memory may also be at least one memory device located remotely from the processor.

The Processor may be a general-purpose Processor including a CPU, an NP (Network Processor), and the like; but also a DSP (Digital Signal Processing), an ASIC (Application Specific Integrated Circuit), an FPGA (Field Programmable Gate Array) or other Programmable logic device, discrete Gate or transistor logic device, discrete hardware component.

In addition, an embodiment of the present invention provides a computer-readable storage medium, in which a computer program is stored, and when the computer program is executed by a processor, the computer program implements: any of the methods described above.

Embodiments of the present invention also provide a computer program product comprising instructions which, when run on a computer, cause the computer to perform: any of the above methods.

In the above embodiments, the implementation may be wholly or partially realized by software, hardware, firmware, or any combination thereof. When implemented in software, may be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions. When loaded and executed on a computer, cause the processes or functions described in accordance with the embodiments of the invention to occur, in whole or in part. The computer may be a general purpose computer, a special purpose computer, a network of computers, or other programmable device. The computer instructions may be stored on a computer readable storage medium or transmitted from one computer readable storage medium to another, for example, from one website, computer, server, or data center to another website, computer, server, or data center via wire (e.g., coaxial cable, fiber, DSL (Digital Subscriber Line)) or wireless (e.g., infrared, wireless, microwave, etc.). The computer-readable storage medium can be any available medium that can be accessed by a computer or a data storage device, such as a server, a data center, etc., that incorporates one or more of the available media. The usable medium may be a magnetic medium (e.g., a floppy Disk, a hard Disk, a magnetic tape), an optical medium (e.g., a DVD (Digital Versatile Disk)), or a semiconductor medium (e.g., a SSD (Solid State Disk)), etc.

For the embodiments of the apparatus, the electronic device, the computer-readable storage medium and the computer program product, since the contents of the related methods are substantially similar to the foregoing method embodiments, the description is relatively simple, and the related points can be referred to the partial description of the method embodiments.

It should be noted that, in this document, relational terms such as first and second, and the like are used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Also, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising an … …" does not exclude the presence of other identical elements in a process, method, article, or apparatus that comprises the element.

All the embodiments in the present specification are described in a related manner, and the same and similar parts among the embodiments may be referred to each other, and each embodiment focuses on the differences from the other embodiments. In particular, for the apparatus, the electronic device, the computer-readable storage medium, and the computer program product embodiments, since they are substantially similar to the method embodiments, the description is relatively simple, and for the relevant points, reference may be made to the partial description of the method embodiments.

The above description is only for the preferred embodiment of the present invention, and is not intended to limit the scope of the present invention. Any modification, equivalent replacement, or improvement made within the spirit and principle of the present invention shall fall within the protection scope of the present invention.

Claims (31)

1. A translation system, the translation system comprising: the system comprises front-end equipment and a back-end server, wherein the front-end equipment is provided with a translation interface;

the front-end equipment is used for transmitting the original text to the back-end server after receiving the original text uploaded by the user through the translation interface;

the back-end server is used for obtaining the original text; translating the original text to obtain a translated text of the original text; and transmitting the translation to the front-end equipment so that the front-end equipment displays the translation on the translation interface.

2. A method of text translation, the method comprising:

acquiring a text of a primary text uploaded by a user through a translation interface provided by front-end equipment;

translating the original text to obtain a translated text of the original text;

and transmitting the translation to the front-end equipment so that the front-end equipment displays the translation on the translation interface.

3. The method of claim 2, wherein the step of translating the textual text to obtain a translation of the textual text comprises:

extracting contents of all parts in the original text according to a preset text content extraction strategy, and generating a translation filling template;

inputting the extracted contents of each part into a pre-trained translation model in sequence to obtain a translation of each content of each part;

and based on the translation filling template, filling contents by using the translations of the contents of all the parts to obtain the translation of the original text.

4. The method according to claim 3, wherein the step of extracting the contents of the parts in the original text and generating the translation filling template according to a preset text content extraction strategy comprises:

and according to paragraph division, extracting the contents of each paragraph in the original text, and generating a translation filling template taking the paragraph as a unit.

5. The method according to any one of claims 2-4, wherein after the step of obtaining the original text uploaded by the user via a translation interface provided by a front-end device, the method further comprises:

identifying attribute information of the original text, and judging whether the original text is in a processable format type according to the attribute information;

if so, executing the step of translating the original text to obtain a translation of the original text;

if not, the text format of the original text is converted into a processable format.

6. The method according to any one of claims 2-4, wherein prior to the step of translating the textual text to obtain a translation of the textual text, the method further comprises:

acquiring a specified language selected by a user through the translation interface;

identifying the original language of the original text, or acquiring the original language of the original text selected by a user through the translation interface;

the step of translating the original text to obtain a translation of the original text comprises:

and translating the original text by utilizing a pre-trained translation model to obtain a translation of the specified language of the original text, wherein the translation model is a model which is obtained by utilizing a corpus sample of the original language to train the original language into the specified language.

7. The method according to any one of claims 2-4, wherein prior to the step of translating the textual text to obtain a translation of the textual text, the method further comprises:

identifying the domain type to which the original text belongs, or acquiring the domain type selected by a user through the translation interface;

the step of translating the original text to obtain a translated text of the original text comprises the following steps:

and translating the original text by using a translation model corresponding to the field type to obtain a translation of the original text, wherein the translation model is obtained by training a corpus sample of the field type.

8. The method of claim 2, wherein after the step of transmitting the translation to the front-end device, the method further comprises:

and if the user selects a translation editing mode through the translation interface is identified, setting the translation into an editing state for the user to edit the content of the translation.

9. The method of claim 8, wherein after the step of setting the translation to an edit state, the method further comprises:

after identifying the content to be edited selected by the user from the translated text, determining the original content corresponding to the content to be edited, and sending a highlight instruction to the front-end device, so that the front-end device highlights the content to be edited in the translated text and the original content in the original text, wherein the manner for the user to select the content to be edited from the translated text comprises the following steps: single click selection, double click selection, or sliding region selection.

10. The method of claim 8 or 9, wherein after the step of setting the translation to an edit state, the method further comprises:

after the content to be edited selected by the user from the translation is identified, sending an edit box setting instruction to the front-end equipment so as to enable the front-end equipment to display an edit box at a specified position of the translation interface or pop up the edit box, wherein the position and the size of the edit box are adjustable, and the user can edit the content to be edited in the edit box.

11. The method of claim 8, wherein after the step of setting the translation to an edit state, the method further comprises:

after receiving the translation edited by the user, re-translating the sentence to which the user editing content belongs based on the user editing content, and transmitting the translation of the sentence to the front-end equipment, so that the front-end equipment recommends the translation of the sentence to the user for selection.

12. The method of claim 2, wherein the step of translating the textual text to obtain a translation of the textual text comprises:

if the text content of the user-defined term base exists in the original text, translating the corresponding text content in the original text according to the translation rule in the user-defined term base, and translating other text content in the original text according to a general translation rule to obtain a translation of the original text.

13. The method of claim 2, wherein after the step of transmitting the translation to the front-end device, the method further comprises:

after the fact that the user selects quality inspection through the translation interface is recognized, a plurality of sentences are selected according to the sequence of translation scores of the sentences from low to high, and the front-end equipment is indicated to highlight the sentences so as to prompt the user to perform key inspection on the sentences.

14. The method of claim 2, wherein after the step of transmitting the translation to the front-end device, the method further comprises:

after identifying that a user selects to download texts through the translation interface, identifying a download mode selected by the user through the translation interface, wherein the download mode comprises translation text download and comparison text download;

if the downloading mode selected by the user is the downloading of the translation text, generating the translation text according to the translation and a specified text format, and sending the translation text to the front-end equipment for the user to download;

if the downloading mode selected by the user is the comparison text downloading, adding the translation of the part of the content after the content of each part of the original text division of the translation based on a preset text content division strategy, and generating a comparison text according to a specified text format after the addition is finished; and sending the comparison text to the front-end equipment for downloading by a user.

15. A method for text download, the method comprising:

receiving a downloading instruction initiated by a user through a translation interface provided by front-end equipment;

obtaining a translation corresponding to the downloading instruction;

identifying a downloading mode selected by a user through the translation interface and contained in the downloading instruction, wherein the downloading mode comprises translation text downloading and comparison text downloading;

if the downloading mode selected by the user is the downloading of the translation text, generating the translation text according to the translation and a specified text format, and sending the translation text to the front-end equipment for the user to download;

if the downloading mode selected by the user is the comparison text downloading, adding the translation of the part of the content after the content of each part of the original text division of the translation based on a preset text content division strategy, and generating a comparison text according to a specified text format after the addition is finished; and sending the comparison text to the front-end equipment for downloading by a user.

16. A quality inspection method, characterized in that the method comprises:

receiving a quality inspection instruction initiated by a user through a translation interface provided by front-end equipment;

selecting a plurality of sentences from the translated text to be checked according to the sequence of translation scores of the sentences from low to high;

and instructing the front-end equipment to highlight the plurality of sentences so as to prompt a user to perform key inspection on the plurality of sentences.

17. The method of claim 16, wherein the manner of scoring the translation of the sentence in the translation comprises:

inputting each sentence in the original text uploaded by the user through the translation interface into a pre-trained translation model to obtain a plurality of translations of the sentence and the translation accuracy of each translation;

and determining the translation score of the sentence based on the translation accuracy of each translation.

18. The method of claim 16, further comprising:

checking an original text of the translated text, and identifying error content in the original text;

and instructing the front-end equipment to prompt the error content in the original text.

19. A method of text translation, the method comprising:

acquiring an original text and a user-defined term library uploaded by a user through a translation interface provided by front-end equipment;

identifying whether the text content in the user-defined term library exists in the original text;

if the translation rule exists, translating the corresponding text content in the original text according to the translation rule in the user-defined term library, and translating other text contents in the original text according to a general translation rule to obtain a translation of the original text.

20. A method for text translation editing, the method comprising:

acquiring an original text uploaded by a user through a translation interface provided by front-end equipment, and translating the original text to obtain a translation of the original text displayed on the translation interface;

and under the condition that the user selects a translation editing mode through the translation interface, setting the translation into an editing state for the user to edit the content of the translation.

21. The method of claim 20, wherein after the step of setting the translation to an edit state, the method further comprises:

after identifying the content to be edited selected by the user from the translated text, determining the original content corresponding to the content to be edited, and sending a highlight instruction to the front-end device, so that the front-end device highlights the content to be edited in the translated text and the original content in the original text, wherein the manner for the user to select the content to be edited from the translated text comprises the following steps: single click selection, double click selection, or sliding region selection.

22. The method of claim 20 or 21, wherein after the step of setting the translation to an edit state, the method further comprises:

after the content to be edited selected by the user from the translation is identified, sending an edit box setting instruction to the front-end equipment so as to enable the front-end equipment to display an edit box at a specified position of the translation interface or pop up the edit box, wherein the position and the size of the edit box are adjustable, and the user can edit the content to be edited in the edit box.

23. The method according to claim 22, wherein an input prompt text is displayed in the edit box, and the input prompt text is used for instructing a user to update the content to be edited according to the input prompt text;

the method further comprises the following steps:

and acquiring text content input in the edit box by the user aiming at the input prompt text, and updating the translation.

24. The method of claim 20, wherein after the step of setting the translation to an edit state, the method further comprises:

after receiving the translation edited by the user, re-translating the sentence to which the user editing content belongs based on the user editing content, and transmitting the translation of the sentence to the front-end equipment, so that the front-end equipment recommends the translation of the sentence to the user for selection.

25. A text translation apparatus, characterized in that the apparatus comprises:

the obtaining module is used for obtaining an original text uploaded by a user through a translation interface provided by front-end equipment;

the translation module is used for translating the original text to obtain a translated text of the original text;

and the transmission module is used for transmitting the translation to the front-end equipment so that the front-end equipment displays the translation on the translation interface.

26. A text download apparatus, the apparatus comprising:

the receiving module is used for receiving a downloading instruction initiated by a user through a translation interface provided by the front-end equipment;

the acquisition module is used for acquiring a translation corresponding to the downloading instruction;

the recognition module is used for recognizing a downloading mode selected by a user through the translation interface and contained in the downloading instruction, wherein the downloading mode comprises translation text downloading and comparison text downloading;

the processing module is used for generating a translation text according to a specified text format and sending the translation text to the front-end equipment for downloading by a user if the downloading mode selected by the user is the translation text downloading; if the downloading mode selected by the user is the comparison text downloading, adding the translation of the part of the content after the content of each part of the original text division of the translation based on a preset text content division strategy, and generating a comparison text according to a specified text format after the addition is finished; and sending the comparison text to the front-end equipment for downloading by a user.

27. A quality inspection apparatus, characterized in that the apparatus comprises:

the receiving module is used for receiving a quality inspection instruction initiated by a user through a translation interface provided by the front-end equipment;

the selection module is used for selecting a plurality of sentences from the translated text to be checked according to the sequence of translation scores of the sentences from low to high;

and the indicating module is used for indicating the front-end equipment to highlight the sentences so as to prompt a user to perform key inspection on the sentences.

28. A text translation apparatus, characterized in that the apparatus comprises:

the acquisition module is used for acquiring an original text and a custom term library uploaded by a user through a translation interface provided by front-end equipment;

the identification module is used for identifying whether the text content in the user-defined term library exists in the original text;

and if the translation module exists, translating corresponding text contents in the original text according to the translation rules in the user-defined term library, and translating other text contents in the original text according to a general translation rule to obtain a translation of the original text.

29. A text translation editing apparatus, characterized in that the apparatus comprises:

the acquisition module is used for acquiring an original text uploaded by a user through a translation interface provided by front-end equipment, and translating the original text to obtain a translation of the original text displayed on the translation interface;

and the setting module is used for setting the translation into an editing state under the condition that the user selects a translation editing mode through the translation interface, so that the user can edit the content of the translation.

30. An electronic device comprising a processor and a memory, wherein the memory is configured to store a computer program; the processor, when executing the computer program stored in the memory, implements: the method of any one of claims 1-14, or the method of claim 15, or the method of any one of claims 16-18, or the method of claim 19, or the method of any one of claims 20-24.

31. A computer-readable storage medium, having a computer program stored therein, which when executed by a processor, implements: the method of any one of claims 1-14, or the method of claim 15, or the method of any one of claims 16-18, or the method of claim 19, or the method of any one of claims 20-24.

Images