JP2776069B2 - Document inspection device - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a sentence inspection apparatus for inspecting a sentence, and more particularly to a sentence inspection apparatus for inspecting a drawing reference code used in a Japanese sentence.

[0002]

2. Description of the Related Art Improvements in document creation devices such as word processors have made it easier for ordinary users to create documents efficiently using a document creation device. In recent years, a function has been realized in which not only Japanese input by kana-kanji conversion or the like but also the input text is checked by a computer.
For example, B-1
-2 "Practical application of Japanese text creation support tool 'Elaboration' on a personal computer" introduces many sentence inspection functions such as "Search for words that should be paraphrased" and "KWIC list creation". ing. By utilizing these functions and detecting and correcting errors in Japanese sentences, the quality of sentences can be improved.

[0003]

A specification (hereinafter referred to as a patent specification) prepared when applying for a patent application or a utility model registration of an invention or device is accompanied by a drawing. The above-described sentence inspection function is also effective when creating the sentence of this patent specification. However, the above-described sentence checking function is intended for processing a general Japanese sentence, and the conventional sentence checking function is not effective for checking expressions and terms specific to drawings in patent documents. In the following, an example will be described.

[0004] One of the errors peculiar to the creation of a text with reference to a drawing is an error in assigning a code for reference to the drawing (generally composed of numerals or alphanumeric symbols). For example, in the case where the item (name of the symbol) “control device” with the symbol “100” is used in the patent drawings, the control device is referred to as “control device 100”. In the present specification, this is referred to as a drawing reference. Due to the change in the drawing, the code “100” is changed to “1”.
01 ", the" control device 10 "
"0" must be changed to "control device 101". However, the description “control device 100” is often used a plurality of times in the specification, and there is a possibility that the description may be left as “control device 100” due to omission of change.

As a means for minimizing the above error, a means for using a character string search function designating a limited character type such as a number in the above-mentioned "Elaboration" and allowing a person to silently confirm the extracted portion can be considered. However, when this function is used, a part which is not clearly a sign, such as a chapter number or a number in a sentence, is also searched at the same time, and the inspection efficiency is poor.

[0006]

SUMMARY OF THE INVENTION The present invention provides a text inspection apparatus which overcomes the above-mentioned drawbacks of the prior art and supports the detection of a drawing reference code error existing in a text. According to the present invention, input means for inputting a sentence, sentence storage means for storing the input sentence, numeral string information extracting means for extracting a digit string in the sentence and a position in the sentence, Numeric string storage means for storing a number string, a word dictionary, word dictionary search means for searching the word dictionary based on a character string immediately before the number string in the text, and search by the word dictionary search means Drawing reference word candidate extracting means for extracting a set of numeral strings corresponding to the extracted word strings as drawing reference word candidates in the text, and drawing reference word candidate storage means for storing the drawing reference word candidates together with the positions in the text A drawing reference word storage means for storing a drawing reference word composed of a code and the name of the code; and comparing the drawing reference word candidate with the drawing reference word so that the code and the numeric string match and correspond. The sign Can be obtained and drawings reference word collating means for detecting what names and the word sequence does not match, the writing test apparatus characterized by comprising a display means for displaying the collation result.

[0007]

The operation of the present invention is as follows. First, a sentence to be input is stored in a sentence storage device, and then a digit string in the sentence and a position in the sentence are extracted. However, in the specification of the present application, a numeral string means a character string that can be used as a reference for drawing, and a continuous alphabetic symbol immediately after the numeral string (a delimiter such as a punctuation mark or a parenthesis) such as 100a or 100-a. Is expanded to allow the combination of
Next, for each of the extracted numeric strings, a word dictionary is searched based on the immediately preceding character string in the text to extract the word strings. In the present specification, a single word is regarded as a kind of word string. The extracted word string is temporarily stored as a drawing reference word candidate in association with the immediately following numeral string. And
By comparing and collating with a drawing reference word (constituted by a code and a name of the code) stored in advance, a fraudulent drawing reference word candidate is detected and displayed.

[0008]

DESCRIPTION OF THE PREFERRED EMBODIMENTS The present invention will be described below with reference to the drawings. FIG. 1 is a block diagram showing one embodiment of the present invention.
In the figure, 1 is an input means for inputting a sentence, 2 is a sentence storage means for storing the input sentence,
Numeral 3 denotes numeral string information extracting means for extracting a numeral string in a text in the text storage means 2 and a position in the text.
Is a number string storage means for storing the extracted number string and position information, 5 is a word dictionary containing at least an independent word that may be a code name,
Reference numeral 6 denotes a word dictionary search unit for searching the word dictionary 5, and 7 extracts a set of numeral strings corresponding to one or more words searched by the word dictionary search unit 6 as drawing reference word candidates. Drawing reference word candidate extracting means 8 is a drawing reference word candidate storage means for storing the drawing reference word obtained by the drawing reference word candidate extracting means 7, and 9 is composed of a code and the name of the code. Drawing reference word storage means for storing a drawing reference word desired by the creator. Reference numeral 10 denotes a drawing reference word candidate in the drawing reference word candidate storage means 8 and a drawing reference word in the drawing reference word storage means 9. Reference numeral 11 is a drawing reference word collating means for collating and outputting a collation result, and reference numeral 11 is a display means for displaying the collation result by the drawing reference word collating means 10. A control unit 12 controls the entire operation. In FIG. 1, an input means 1 is a floppy disk device or a communication line, and a sentence storage means 2, a numeric string storage means 4, a word dictionary 5, a drawing reference word candidate storage means 8, and a drawing reference word storage means 9 are a memory device or the like. The display means 11 is a CRT display device or the like, and the numeral string information extraction means 3, the word dictionary search means 6, the drawing reference word candidate extraction means 7, the drawing reference word matching means 10, and the control unit 12 are realized by a CPU and a memory. it can.

FIG. 2 is a part of an example of a patent specification sentence input from the input means 1 and stored in the sentence storage means 2. In the sentence, "result storage means" and "morphological analysis means" This is an example of an error given the same code “4”. FIG. 3 is an example of a correct drawing reference word which is also input from the input means 1 and stored in the drawing reference word storage means 9. Hereinafter, the operation of this embodiment will be described with reference to the text of FIG. 2 and the drawing reference words of FIG. 3 as examples.

The Japanese sentence input from the input means 1 is
It is stored in the text storage means 2. Thereafter, under the control of the control unit 12, the digit string information extracting unit 3
Are extracted and stored in the numeral string storage means 4. The details of the digit string extraction procedure are omitted because they are introduced as "character type character string search commands" in the above-mentioned document. FIG. 4 shows the contents of the numeral string storage means 4 at this point. In FIG. 4, “position” is represented as the number of characters from the beginning of the text.

Next, the drawing reference word candidate extracting means 7 activates the word dictionary searching means 6 under the control of the control section 12, and activates the word dictionary searching means 6 in the sentence storage means 2 immediately before each digit string in the Japanese sentence. A word dictionary 5 is searched based on a character string (which can be referred to by using the position information in FIG. 4) to extract a word string, and combined with a corresponding numeral string as a drawing reference word candidate storage means. 8
To be stored. FIG. 5 shows an example of realizing the word dictionary 5. FIG.
Indicates that each word in the word dictionary 5 has notation, reading, part of speech information, and unnecessary word information indicating that it is not a drawing reference word candidate. Since the word dictionary 5 and the word dictionary search means 6 are generally used in kana-kanji conversion processing of a Japanese word processor, detailed description thereof will be omitted.

The drawing reference word candidate extraction means 7 performs processing in the following procedure. (Step 7-1) The number string and its position information are extracted from the number string storage means 4. When all are taken out, the process ends. (Procedure 7-2) Referring to the text storage means 2, based on the character string immediately before the numeric string indicated by the position information, the procedures (7-2a) to (7-2a)
The word dictionary 5 is searched by the word dictionary search means 6 based on the procedure (7-2c). (7-2a) If a word cannot be extracted, it is determined that extraction has failed (to step 7-1). (7-2b) If the extracted word string contains information indicating that the word is unnecessary, the extraction is determined to have failed and the procedure proceeds to (procedure 7-1). (7-2c) A word string that forms the longest compound word that can be formed by the character string immediately before the number string is extracted. (Procedure 7-3) A set of the obtained word string and numeral string is set as a drawing reference word candidate, stored in the drawing reference word candidate storage unit 8, and then to (Step 7-1).

FIG. 6 shows the contents of the drawing reference word candidate storage means 8 in this embodiment. Comparing the numeric expressions in the text of FIG. 2 with the drawing reference word candidates of FIG. 6, “claim 1”, “3.”, “between two words”, and “third” in FIG. 2 are excluded. This means that “3.” and “between two words” are (7-2a)
This is because "claim 1" and "third" are excluded in (7-2b).

Next, the drawing reference word matching means 10 checks the contents of the drawing reference word candidate storage means 8 against the contents of the drawing reference word storage means 9, and detects an invalid one in the drawing reference word candidate storage means 8. To detect. The specific procedure is as follows. (Step 9-1) One drawing reference word candidate is extracted from the drawing reference word candidate storage means 8. When all processing is completed, the process ends. (Procedure 9-2) It is detected whether or not the same drawing reference word candidate as extracted exists in the drawing reference word storage means 9. If it exists (procedure 9-1), if it does not exist, go to (procedure 9-3). (Step 9-3) The drawing reference word candidate is sent to the display means 11, and
Go to (Procedure 9-1).

The contents of the drawing reference word candidate storage means 8 are shown in FIG.
Then, if the contents of the drawing reference word storage means 9 is FIG.
It is detected that the "result storage means 4" in the drawing reference word candidate storage means 8 is invalid. The obtained drawing reference word candidate which is considered to be illegal is displayed to the user by the display means 11. FIG. 7 shows a display example of the display means 11 in this example.

FIG. 7 shows only display of drawing reference word candidates which seem to be illegal, but as another display form,
Based on the corresponding position information in the numeral string storage means 4 and the sentence information in the sentence storage means 2, it is also possible to display how the incorrect drawing reference word candidate is used in the sentence.

[0017]

According to the present invention, it is possible to easily detect an illegality of a reference symbol for a drawing in a text and present it to a user.

[Brief description of the drawings]

FIG. 1 is a block diagram showing one embodiment of the present invention.

FIG. 2 is an explanatory diagram showing a part of an example of a Japanese sentence in a sentence storage means 2;

FIG. 3 is an explanatory diagram showing an example of a drawing reference word in a drawing reference word storage means 9;

FIG. 4 is a diagram showing an example of contents stored in a number string storage means 4;

FIG. 5 is a diagram showing a part of the contents of a word dictionary 5;

FIG. 6 is a diagram showing an example of contents stored in a drawing reference word candidate storage means 8;

FIG. 7 is a diagram illustrating a display example of a display unit 11;

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Input means 2 Sentence storage means 3 Numeric string information extraction means 4 Numeric string storage means 5 Word dictionary 6 Word dictionary search means 7 Drawing reference word candidate extraction means 8 Drawing reference word candidate storage means 9 Drawing reference word storage means 10 Drawing reference words Matching means 11 Display means

Claims (1)

(57) [Claims] An input unit for inputting a sentence; a sentence storage unit for storing the input sentence; a digit string information extracting unit for extracting a digit string in the sentence and a position in the sentence; Number string storage means for storing a sequence, a word dictionary, and word dictionary search means for searching the word dictionary based on a character string immediately before the number string in the text,
Drawing reference word candidate extraction means for extracting a set of a number string corresponding to the word string searched by the word dictionary search means as a drawing reference word candidate in a sentence, and storing the drawing reference word candidate together with the position in the sentence Drawing reference word candidate storage means, a drawing reference word storage means for storing a drawing reference word composed of a code and the name of the code, and comparing the drawing reference word candidate and the drawing reference word with the code A sentence inspection apparatus, comprising: a drawing reference word matching unit that detects a case where the number of strings matches and the name of the corresponding code does not match the word string; and a display unit that displays a matching result.