The content of the invention
To overcome translation software translation quality poor, the slow defect of human translation translation speed, the invention discloses a kind of fast
Fast translating equipment and method.
A kind of rapid translation device of the present invention, including historical trace storehouse, input distribute module, first judge module,
Contrast module, first screening storehouse, mark module, recovery module and display device;
File to be translated inputs the input distribute module, and is distributed in units of whole sentence;
The judge module first judges each simple sentence to be translated that input distribute module distribution is obtained, and judges that this is to be translated
Whether simple sentence occurs first in file to be translated, is to be stored in screen storehouse first, otherwise sign module makes the sentence
Represent the mark do not translated;
The contrast module is connected with historical trace storehouse, in comparison library simple sentence with input distribute module distribution obtain it is each
Simple sentence to be translated, it is such as identical, it indicates that mark module makes the simple sentence to be translated the mark for representing not translate;
Display device shows the file to be translated made and do not translate mark all;
The recovery module is used for after the completion of translation, and translating for whole simple sentences in storehouse is screened by historical trace storehouse and first
Text returns to the correspondence position of file to be translated according to corresponding relation.
It is preferred that, the letter abbreviations in file to be translated are also identified as simple sentence by the input distribute module.
Specifically, not translating of making of the mark module is designated color displays different from normal display, and for that can not compile
The state of collecting.
File to be translated is assigned as simple sentence to be translated the invention also discloses a kind of rapid translation method, including in advance
Step, in addition to:
The step of I1 is compared simple sentence to be translated with the simple sentence in historical trace storehouse:Simple sentence and input distribution mould in comparison library
Each simple sentence to be translated that block distribution is obtained, it is such as identical, it indicates that the simple sentence to be translated is made expression not by mark module
The mark of translation;Make and do not translate the simple sentence of mark and do not enter back into I2 steps;
I2 screens step first:Judge whether the simple sentence to be translated occurs first in file to be translated, be then to be stored in head
Secondary screening storehouse, otherwise sign module this is made and represents the mark do not translated;
I3 recovering steps:After the completion of translation, the translation of whole simple sentences in screening storehouse is pressed by historical trace storehouse and first
The correspondence position of file to be translated is returned to according to corresponding relation.
It is preferred that, set one to match array D,, should if not having to store same simple sentence in D to each simple sentence
Simple sentence is stored in D, is otherwise made the simple sentence and is not translated mark.
Specifically, described in advance the step of file to be translated is assigned as into simple sentence to be translated including the word in file to be translated
Mother's abbreviation is identified as simple sentence.
Using rapid translation device and method of the present invention, treat translated document and carried out in advance using history file storehouse
Screening, carries out repeating filtering with reference to self-contrast, reduces translation word quantity, contrast ensures the translation of translation in units of sentence
Quality, by actual test, the present invention can reduce translation amount more than 30%.
Embodiment
Below in conjunction with the accompanying drawings, the embodiment to the present invention is described in further detail.
Rapid translation device of the present invention, including historical trace storehouse, first input distribute module, judge module, contrast
Module, first screening storehouse, mark module, recovery module and display device;
File to be translated inputs the input distribute module, and is distributed in units of whole sentence;
The judge module first judges each simple sentence to be translated that input distribute module distribution is obtained, and judges that this is to be translated
Whether simple sentence occurs first in file to be translated, is to be stored in screen storehouse first, otherwise sign module makes the sentence
Represent the mark do not translated;
The contrast module is connected with historical trace storehouse, in comparison library simple sentence with input distribute module distribution obtain it is each
Simple sentence to be translated, it is such as identical, it indicates that mark module makes the simple sentence to be translated the mark for representing not translate;
Display device shows the file to be translated made and do not translate mark all;
The recovery module is used for after the completion of translation, and translating for whole simple sentences in storehouse is screened by historical trace storehouse and first
Text returns to the correspondence position of file to be translated according to corresponding relation.
During using the present invention, file to be translated is inputted to input distribute module first, input distribute module is according to certain
File to be translated is divided into simple sentence by rule, and common processing mode is, with punctuation mark, to make such as fullstop, question mark, ellipsis
Simple sentence is marked off for simple sentence segmentation identifier.
The simple sentence that historical trace storehouse marks off file to be translated is compared with the simple sentence prestored in historical trace storehouse
Right, it is completely the same to compare principle, i.e., the order in tandem of each word and whole words is completely the same in simple sentence, has met
Complete consistent simple sentence, which is made, does not translate mark.
It is currently preferred individually to draw the letter abbreviations for possessing common art-recognized meanings as simple sentence during simple sentence is recognized
Branch away, without considering whether the letter abbreviations are separated by punctuation mark, such as WTO(World Trade Organization), USA(United States of America
United states)It is located at Deng, usual letter abbreviations in a sentence, then contrast abbreviation first, then contrast the sentence where the abbreviation.
By taking most common English to Chinese as an example, historical trace storehouse is according to accumulation in the past or discloses English-Chinese document and accumulated
The phrase that completely looks like of simple sentence or can express be data library that unit is stored, including one-to-one original English version and the Chinese
Language translation, it is well known that each word there may be multiple declarations of will in English, but in each specific sentence, should
The meaning of word generally immobilizes, and the meaning of simple sentence and phrase sanctified by usage is in different context of co-texts, meaning
Think statement also basically identical.
Using simple sentence file and divided simple sentence unit in file to be translated in contrast module contrast historical trace storehouse, press
Screened according to identical comparison principle, the identical simple sentence filtered out is made in file to be translated does not translate mark
Know.Make and do not translate the simple sentence of mark and no longer carry out follow-up first judging.
After the completion of the contrast of historical trace storehouse, continuation judges that remaining simple sentence that is, whether the simple sentence is at this first
Occur first in file to be translated, judge module itself is screened using file to be translated first, by the clause repeated
Among filtering, same piece article, due to being the description for the same subject that same author writes, there are quite a lot of simple sentence or phrase may
Repeatedly occur, judge module judges whether each simple sentence occurs first in file to be translated successively first, so long as not first
Occur, then make and do not translate mark, be to occur first, be stored in screening storehouse first.
The contrast of historical trace storehouse should be earlier than judgement contrast first, it is possible to reduce judge amount of calculation, for example certain word is in text
Middle first time occurs, if the word is also appeared in historical trace storehouse, and only needing to progress historical trace contrast can make
Mark is not translated.If being judged contrast first first, also need to progress historical trace and judge just to can obtain result, for
For one article, generally, the simple sentence quantity repeated always less than only there is simple sentence quantity once, and by
Simple sentence accumulation in historical trace storehouse is huge, appears in the simple sentence in historical trace storehouse and is often more than what is repeated in quantity
Simple sentence, therefore will should judge to postpone first.
The simple sentence for not translating mark is made in internal system, should be different from other lists in the dispaly state of display module
Sentence, such as color displays are different from normal display, to prevent translator from voluntarily translating or misoperation, can be with by the simple sentence do not translated
Being set as can not editing mode.Translator only needs to operate and translated to be needed to turn in the file to be translated that display module is shown
The simple sentence translated, and the simple sentence repeatedly occurred in file to be translated stored first in judge module.
After the completion of translation, what is obtained is the translation for including some vacancies, and vacancy is to make the simple sentence pair for not translating mark
Answer position, system simple sentence translation in judge module according to the translation stored in historical trace storehouse and first, according to corresponding relation
These translations for making the simple sentence for not translating mark are backfilling into the vacancy of translation, complete translation is obtained.
A kind of embodiment of file to be translated is handled in real time using loop nesting algorithm batch as Fig. 2 is provided,
System once reads in N files to be translated, and interpretation method of the present invention is taken to each piece file to be translated,
Full text is made pauses in reading unpunctuated ancient writings first, C simple sentence is obtained, to each simple sentence, judgement identification is carried out successively, judges that identification uses cycle accumulor
Mode, i.e., to J, it is first determined whether appearing in historical trace storehouse, do not translate after mark if so, then making, continue into
The judgement that row is J+1, if it is not, then continuing to determine whether occur first.
The embodiment of judgment step in the present embodiment is first:A customized matching array D is set up, just
The array is sky during the beginning, to each simple sentence, if occurring for the first time, is stored in matching array D, and then carrying out J+1 sentences
It is disconnected, when the simple sentence occurs for the second time, then the simple sentence is made and do not translate mark, then carry out J+1 judgements, thus match
Array D is finally stored in unduplicated whole simple sentences in file to be translated, the complete matching array D of translator's actual translations
Whole simple sentences, with reference to the history translation stored in historical trace storehouse, that is, complete whole translations of file to be translated.Using
The mode of matching array is set, and data process method is simple, program operation consumption resource is few.
Disclosed in this invention embodiment description method or can directly use hardware, computing device the step of algorithm
Software module, or the two combination implemented.Software module can be placed in random access memory(RAM), internal memory, read-only storage
Device(ROM), electrically programmable ROM, electrically erasable ROM, register, hard disk, moveable magnetic disc, CD-ROM or technology neck
In any other form of storage medium well known in domain.
Previously described each preferred embodiment for the present invention, if the preferred embodiment in each preferred embodiment
It is not substantially contradictory or premised on a certain preferred embodiment, each preferred embodiment can any stack combinations
Use, the design parameter in the embodiment and embodiment merely to clearly state inventor invention verification process, and
The scope of patent protection of the limitation present invention is not used to, scope of patent protection of the invention is still defined by its claims, all
It is the equivalent structure change made with the specification and accompanying drawing content of the present invention, similarly should be included in the protection model of the present invention
In enclosing.