JPS6177955A - Automatic number application system for sentence number of document processing - Google Patents

Abstract

PURPOSE:To apply numbers for each sentence mechanically and effectively to attain an application system of easily associable numbers, by giving sentence numbers automatically and effectively to each sentence of documents supplied continuously after recognizing the special symbols and sentence end symbols which are added additionally. CONSTITUTION:A display device 1 containing an input device like a keyboard, etc. is connected to a processor 3 of an automatic application system for document numbers. While an external memory 6 having a memory 4, a screen buffer 5 and an English sentence file is connected to the processor 3. The English sentence file part of the memory 6 includes an index part which gives positioning to an identification number and a data part which stores the identification number and the data on English sentences or Japanese sentences. Then the processor 3 recognizes the special symbols and sentence end symbols supplied additionally, and sentence numbers are given automatically and effectively to each sentence of documents supplied continuously from the device 1. Thus the numbers are applied to each sentence mechanically and effectively. This ensures the easily associable numbers.

Description

[Detailed Description of the Invention] [Field of Application of the Invention] The present invention provides an efficient way to assign an identification number to each sentence for a field in which it is necessary to identify each sentence of a natural language document (for example, a machine translation system). This relates to a method for automatically assigning numbers to.

[Background of the invention]

Traditionally, sentences written in natural language were
The text is processed by a text editing device (word processor) as shown in 765, which processes multiple sentences as one processing unit.

Therefore, each - sentence is not recognized and processed. Furthermore, in electronic translation devices such as those shown in Japanese Patent Application Laid-Open No. 56-42876, a word or a single sentence is immediately translated and displayed, so there is no need to assign a recognition number to each input sentence or word. .

However, in machine translation from English to Japanese using the method disclosed in Japanese Patent Application Laid-Open No. 58-40684, there are systems that have to identify and process each sentence.Such systems Now, the input sentence (
An identification number must be assigned to each sentence (hereinafter referred to as "sentence"), but conventionally, an inputter had to consciously assign an identification number to each sentence to create keyed data.

[Purpose of the invention]

In view of the above points, it is an object of the present invention to provide a method for mechanically and efficiently adding numbers to each sentence when inputting a document written in a natural language.

[Summary of the invention]

In order to achieve the above object, the present invention adds special symbols and end-of-sentence symbols to a document written in a natural language in accordance with the structure of the document, thereby adding numbers to each sentence that match the physical concept. That is, it provides a method for automatically assigning numbers that are easy to associate with each other when searching for a sentence.

[Embodiments of the invention]

Examples of the present invention will be described in detail below.

FIG. 1 represents a block diagram showing one embodiment of a device for implementing the automatic numbering system according to the present invention.

In the figure, 1 is a display device, 2 is an input device such as a keyboard, 3 is a processor, 4 is a memory, 5 is a screen buffer, and 6 is an external memory such as a magnetic disk. In this embodiment, for convenience of explanation, the numbering method and input method will be explained using English as the natural language.

Documents written in natural language often exist in the form of books, newspapers, etc. as shown in Figure 2. When explaining these documents, you might say something like, ``What book, what page, what line is the text written in? What newspaper, what page, what column is the article?'' K. I often position documents.

Using these physically assigned names (for example, the names of document, page, paragraph, and sentence as shown in Figure 2), the document name (9 digits), page number (4 digits), page number (4 digits), Paragraph number (2 digits)
, a sentence number (2 digits), a total of 17 digits of identification number are assigned, and an example in which this identification number is stored in an English file in the external memory 6 is shown in Figure Wc3. This file is
Sentence data can be input and output by specifying an identification number, and data can also be input and output in ascending or descending order of identification numbers. The details are shown in FIG.

The English file in the external memory 6 is as shown in FIG.
It is a file consisting of an identification number, an index section consisting of a group of pointers that locates in which data section the data is stored, and a portion of data storing the identification number and English or Japanese data.

The index set (I8) 1101 in FIG. 4 stores the maximum identification number stored in each sequence set (38) 1102 and a pointer indicating which sequence set. is the maximum identification number in the control interval (CI) in each control area (CA) 1103,
Stores a pointer indicating which control interval. This control interval is
This corresponds to one track of a magnetic disk device. By searching the index section, arbitrary data can be found and read out in ascending order or descending order. Also, when data is read in ascending order or descending order, the record 1104 of the data section has an identification number, so it is possible to know the identification number of the read data.

Further, this file has management information as shown in FIG. 5 as a record in the data record section of FIG. 4. Page management information 1201iJ There is page management information for each document that is stored.

a record identification number page to indicate that the
O is written in all the bar graph and sentence areas. The bar graph management information 1202 is prepared for each page, and is identified by the bar graph and sentence areas being 0.

Similarly, sentence management information 1203 is prepared.

These stores include the maximum number of currently stored pages, paragraphs, and sentences, and the number of pages, paragraphs, and sentences that have been stored (registered in the file).

For example, the page management information includes the maximum page number and number of pages registered in the DokiToriment.

By using this management information, it is possible to know the maximum number of pages, maximum number of paragraphs, maximum number of sentences, etc. without directly searching for sentences using the automatic sentence numbering described below. Note that the management information record does not need to exist in the English file. That is, it may exist as a single file as an external file.

In order to make the document usable for machine processing, the document to be processed is input into an English file in an external memory from the input device 2 shown in FIG. In this case, there are two possible cases: a new input and an additional input. In inputting an English document, the break between one sentence is recognized by adding a space of two characters to the end-of-sentence symbol shown in FIG. (For example, T11 is a p
en, ΔΔ...Δ indicates a blank. ) The operation of the embodiment having the above configuration will be explained using the processing flow diagram of FIG. 7 and the screen showing the processing process of FIG. 8.

First, the processor 3 displays the guidance screen shown in FIG. 8 1al on the display group 'B1. The person who entered the information should use this 11
5L) 1).

Based on this instruction data, the processor 3 sets the mode (new input mode or additional input mode) (502
).

In the new input mode, the processor 3 creates a screen for inquiring about the name of the document and transfers it to the screen buffer 5. The screen displayed on the display device (Fig. 8(b)
)), the inputter specifies only the Dokishimament name from the keyboard 2 (503), and the processor 3 checks whether the specified Dokishimament name exists in the English file in the external memory 6 (504). ), if the document name is specified incorrectly, the process returns to step 503. If it does not exist, the process returns to step 507 as a new input.

In the case of the additional input mode, the processor 3 creates a screen that inquires about the document name, page number, and paragraph number, and transfers it to the screen buffer 5.

On the screen displayed on the display device (FIG. 8(b)), the inputting person specifies the document name and page number bar graph number using the keyboard 2. Set the page number and paragraph number depending on the position where additional input is made (
505). The processor 3 checks whether data having the specified document name, page number, and bar graph number as an identification number exists in the English file of the external file (506), and if it does not exist, it indicates a specification error (505).
Return to processing.

If it exists, control is transferred to processing in step 507. The processor 3 frees the English language screen shown in FIG. 8(C) and transfers it to the screen buffer 5. Figure 8 (C) shows the dokiement name,
An index display area that displays page numbers, paragraph numbers, etc., a previous screen information display area that displays part of the information on the previous screen or part of the information in the English file in the external memory 6, and an English text area for inputting English data. data entry area,
It consists of an error message display area that displays error information during input.

In the previous screen information display area, nothing is displayed in the case of the first English manual screen at the time of new input, but in the case of the English manual screen at the time of additional input, additional input of the English file in the external memory 6 is attempted. Display information about the sentence immediately before the sentence. The inputter inputs an English document into the English data input area.

(507) At this time, at the end of each sentence, the sentence-final symbol 1 shown in FIG. 6 and two or more spaces are added and input. The processor 3 recognizes sentence breaks using this end-of-sentence symbol and two or more blank spaces, and assigns an identification number to each sentence. For English documents, the number of pages can be determined by entering English text using the input method described in the input method column of the numbering method for new input in Figure 9 and the numbering method for additional input in Figure 10. or change the bar graph number.

The meanings of the symbols in Figures 9 and 10 are as follows.

Δ: 1 or more blanks, ^20 or more blanks p: Specified page number (within 4 digits = 1 to 9999) h: Specified paragraph number (within 2 digits: 1 to 99) po: l Previous English page Number: lBaragraph number before the English sentence jg:l Sentence number before the English sentence Pl: Set of existing page numbers in the English file in external memory P2: Set of existing page numbers in the #1 side backup memory H
1: Set of existing paragraph numbers in the English file in external memory H2: Set of existing paragraph numbers in the screen Bakuno memory hI: Maximum paragraph number of page p in the English file in external memory Band 1: Unless the numbering conditions are satisfied Explanation of error conditional expression: a5A...a is the element of the bill collector.

A to a...A is not an element of set A 0 The explanation of FIGS. 9 and 10 will be made using FIGS. 11 and 12. This example is an example of new registration.

Figures 11 and 12 are excerpts from pages 135 and 136 of a certain manual, and Figure 13 shows the sentences of this manual entered using the numbering method shown in Figure 9. This is an example of input.

First, we will take as an example the case where input is made from the fifth row from the bottom of FIG. 11 to FIG. 12. %DOCφφφ1” has already been specified as the document name.
In order to specify that the English sentence at the beginning of the English human power screen in Figure 3 is the first sentence of the third paragraph on page 135, we add the special symbol 1・“ (at mark) and write it as '...135Δ @3 ΔScheduling jo
The input format of b -...“ is.This 1@・“
(The double at mark 1i!!> is a symbol that specifies a page change, and the number 135 immediately after it indicates that the page number has been changed to 135. This is the page management information in Figure 5. 1201. 'e'' (one at mark) is a symbol specifying a paragraph change,
The immediately following number t3 means that the paragraph number has been changed to 3. This is also stored in the paragraph management information 1202. Δ means blank. With the above specification, the identification number of the first English sentence becomes the identification number shown in the example of FIG. 3. In addition, the first English sentence has a 1 :“ (colon) and two spaces that allow the processor 3 to recognize the sentence breaks.The next English sentence has the special symbols ゝ and “ added to it. However, in this case, an identification number is given by adding the number 1 to the sentence number of the previous sentence.

This can be obtained by reading the maximum page number (sentence management information 120y) in FIG. 5 and adding 1 to it. This updated number is also written to the sentence management information @1203. A table summarizing the numbering method described above is shown in FIG. 9 and No. 1011.

Next, the processor 3 processes the input data of the English text transferred from the screen buffer 5 to the memory 4 based on the numbering method shown in FIGS. 9 and 10 from the special symbol and the end of sentence symbol (FIG. 6) K. Assign an identification number to each sentence. Page, paragraph, and sentence numbers are updated by reading the management information in response to the page change, bar graph change, and sentence change symbols shown in Figures 9 and 10, unless a number is specified as described above. The maximum number stored in the system plus 1 is given as the identification number.

On the other hand, the updated maximum number is stored in the management information.

At this time, logic such as whether there is a sentence with the same identification number in the English file in the external memory 6 (an error occurs if there is the same one), and whether the identification numbers of each sentence in the English Human Power screen are duplicated is checked. Check (508
). The numbering conditions and numbering check conditions for input English text are described in FIGS. 9 and 10. If there is an error in the numbering check of the input English text, control is transferred to step 510; if there is no error, control is transferred to step 511 (509
).

If there is an error in the input English data, information specifying the first position where the error was found (cursor position information, etc.) and an error message are transferred to the screen buffer 5. The English expert corrects the input data from the position specified on the screen, referring to the error message (510).

If there is no error in the input English sentence, each English sentence is written as keyed data to the English file in the external memory 6 along with the identification number automatically assigned in the process of step 508 (511).

When all English manual processing is completed, the manual operation ends. However, if the input is to be continued, control is transferred to 513 (512).

The processor 3 uses part of the information on the currently processed screen to create the next English human screen and transfers it to the screen buffer (513). A concrete example of this is shown in FIG. When creating the English Human Power screen, display the final sentence displayed on the previous screen in the information display area of the previous screen (re-
Figure 4(a)). The previous screen is indicated by 1401 and the next screen is indicated by 1402. Before! If the entire sentence to be displayed does not fit in the information display area, only the data of the length that fits is displayed from among the sentences.

If the sentence after Nk displayed in the previous 1fiI direction does not have the final symbol B or two or more spaces, that sentence will not be marked as the final sentence, will be treated as unentered, and will be displayed in the English data input area. (Figure 14(b)).

The previous screen is shown as 140 and the third screen is shown as 1404. In this way, by displaying a part of the previous screen information on the English human-powered screen, smooth input operations are possible.

〔Effect of the invention〕

As explained above, according to the present invention, it is possible to use a natural language input method (more efficient automatic numbering for each sentence), and to use a physical input method. It is a format that can be easily associated with oJ.

[Brief explanation of drawings]

FIG. 1 is a block diagram showing an embodiment according to the present invention, FIG. 2 is a diagram showing an example of a document written in one natural language, and FIG. 3 is a diagram showing an example of an English file in an embodiment according to the present invention. Explanatory diagram, Figure 4 is a diagram showing the structure of a file, Figure 5 is a diagram explaining management information, Figure 6 is a diagram showing end-of-sentence symbols in English, and Figure 7 is a processing flow of an embodiment according to the present invention. FIG. 8 is a diagram showing a display screen in an embodiment of the present invention, FIGS. 9 and 10 are diagrams showing an automatic numbering method in an embodiment of the present invention, FIG. 11, FIG. 12 Figure 13 shows a document that is an excerpt from a certain manual.
FIG. 14, which is a diagram showing an example of an English human-powered screen according to an embodiment of the present invention, is an explanatory diagram showing a processing outline of the human-powered method according to an embodiment of the present invention. [Explanation of symbols] 1...Display device 2...Input device such as keyboard 3...Processor 4...Memory 5...Screen buffer 6...External memory such as magnetic disk 1st (! 1 2nd circle (Q-) (+W hesisinon (t) 3rd name vJ4唖t2o; 1 躬 6I￥1 も70 Procedural amendment (method) Indication of the case 1988 Patent Application No. 198556 Invention 0 people Ya Relationship between automatic pattern number type correction of paper name in document processing Patent applicant name 15101 Manufactured by Hitachi Co., Ltd.
Manufacturer (1 person) Agent

Claims (1)

[Claims] 1. In natural language document input, by recognizing additionally input special symbols and end-of-sentence symbols, it automatically and efficiently assigns sentence numbers to each consecutively input document. Automatic numbering method for sentence numbers in document processing.