JP2000194712A - Data processor and recording medium - Google Patents

Abstract

PROBLEM TO BE SOLVED: To permit rising of a multidocument system in a short time by simplifying previous setup including keyword extraction/registration. SOLUTION: The kind of file is discriminated from the expander of a file read out of a document file 2 and when the kind of file is a drawing, a keyword is extracted from a range including a title column by a drawing keyword extracting means 5 but when it is a slip, a keyword is extracted from the item row and item column of the slip by a slip keyword extracting means 6. When it is a document, a keyword is extracted from a character string having the prescribed number of characters from the head of the document by a document keyword extracting means 8.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention has a function of automatically extracting keywords from a plurality of types of files, such as drawings created by CAD, forms created by spreadsheet software, and documents created by a word processor. Data processing device,
And a recording medium on which a computer program for keyword extraction is recorded.

[0002]

2. Description of the Related Art In order to centrally manage a large number of files, it is necessary to register a keyword that is a key for retrieval. For example, in the case of a drawing created by CAD, the right or bottom of the item name column such as “drawing number”, “drawing name”, “designer name”, “design date”, etc. The operator refers to the title box in which the data unique to the drawing is entered in the column, and inputs the unique data by key, and registers a keyword.

In the case of a form created by spreadsheet software, "year", "sales", "cost", "profit and loss" indicating attributes of numerical values inputted in each column and each row of a form in a table form.
An operator key-inputs an item name such as "1997" in each row and each column to register a keyword.

In the case of a document, for example, a method has been proposed in which the number of appearances of each word in the whole sentence is counted, and a word having a relatively high appearance frequency is automatically extracted as a keyword. However, in many cases, the search is performed in full text, and the infrequent and insignificant words may be extracted as keywords. It is.

[0005]

By the way, in the case of a drawing, it is conceivable that a range of a title box including a keyword is designated, and the keyword is automatically extracted from this range. However, if there are various drawings with various drawing sizes (A0, A1, A2, etc.) and different origin positions, the coordinate values for specifying the range of the title box (for example, the upper left coordinate value of the title box) The operator is not necessarily the same in all drawings, so the operator must set and input coordinate values for designating the range of the title box for each drawing. Further, when keywords are extracted for each item name such as “drawing number”, “drawing name”, “designer name”, “design date”, etc., the coordinate value of each item must be set and input for each drawing. Must.

If the number of rows is indefinite in the form, the “year”, “sales”, “cost”, “profit / loss”, and “1997”
It is necessary to set and input the position of the row or column in which the item name to be a keyword such as is input for each form.

Accordingly, a system (hereinafter, referred to as a unit) for integrally managing a plurality of types of files having different formats such as drawings, forms, documents, etc.
In the case of constructing a multi-document system), the setup work including the extraction (registration) of keywords is complicated, so the work is long, and it takes a long time to start up the system. Therefore, when the system is started, there is a possibility that the contents of the file have already become obsolete.

The present invention has been made to solve such a problem, and converts the coordinates of a drawing into a predetermined point, for example, coordinates having the origin at the lower left, to determine the size of the drawing and the position of the origin. Regardless of what the title block is included in, the keyword is automatically extracted from this range. Automatically extract keywords from this row and column based on the character type of the characters entered in the item, for example, to detect whether there is a character other than a number or not, based on whether the number of characters is the largest, etc. In addition, by automatically extracting keywords from a predetermined number of character strings at the beginning of a document, which is likely to include a table of contents, an outline, etc., the prior setup work including keyword extraction and registration can be performed. Simple And, allowing the launch of a multi-document system in a short period, the data processing apparatus having a keyword extraction function, and aims to provide a recording medium that keyword extraction computer program is recorded.

[0009]

A data processing apparatus according to a first aspect of the present invention is a data processing apparatus having a function of extracting a keyword from a drawing information file having a title box, wherein the data processing apparatus uses coordinates having a predetermined point as an origin. Means for storing at least one point coordinate value defining a range including a title block in a drawing defining a range for each size of the drawing, and means for converting coordinates of the drawing into coordinates having the predetermined point as an origin. Means for extracting a keyword from a range defined based on the coordinate values of the at least one point, stored in association with the size of the drawing from which the keyword is to be extracted, in the drawing after coordinate conversion. It is characterized by having.

In the data processing apparatus according to a second aspect of the present invention, the title field of the first aspect has a plurality of item fields for inputting data unique to the drawing and fixed data indicating attributes of the data. Means for extracting fixed data entered in the column, and means for extracting data entered in the item column adjacent to the item column in which the fixed data is entered as a keyword of the fixed data. It is characterized by having.

A data processing apparatus according to a third aspect of the present invention is a data processing apparatus having a function of extracting a keyword from a file of information in a table format, wherein data indicating an attribute of the data input to each column is input. An item row composed of items having data items indicating the attributes of data items entered in the item lines and / or items in each line is detected based on the character type of the characters entered in the items. Means, and means for extracting a keyword from the item of the item row and / or the item column.

A data processing apparatus according to a fourth invention is characterized in that, in addition to the third invention, a means for extracting a keyword from a range other than information in a table format is provided.

A data processing apparatus according to a fifth aspect of the present invention is a data processing apparatus having a function of extracting a keyword from a document file, comprising: means for decomposing a character string having a predetermined number of characters from the beginning of a document into words; Means for extracting a keyword.

A data processing apparatus according to a sixth aspect of the present invention is a data processing apparatus having a function of extracting a keyword from a plurality of types of files including a drawing information file, a tabular information file, and a document file. Means for judging the file type from the extension of the file to be extracted, and at least one coordinate value defining the range including the title box in the drawing defining the range with coordinates with the predetermined point as the origin. Means for storing each drawing size; if the file type is a drawing information file, the coordinates of the drawing are converted into coordinates having the predetermined point as the origin, and keywords in the drawing after the coordinate conversion are extracted. Means for extracting a keyword from a range defined based on the coordinate values of the at least one point stored in association with the size of the drawing to be read; If the type is a file of information in the form of a table, an item row consisting of items in which data indicating the attributes of the data entered in the items of each column has been entered and / or the data of the data entered in the items of each line. An item string consisting of items in which data indicating attributes is input is detected based on the character type of the characters input in the item, and the item line and / or
Alternatively, means for extracting a keyword from the items in the item string, and when the file type is a document file, means for decomposing a character string having a predetermined number of characters from the beginning of the document into words, and extracting a keyword from the character string It is characterized by having.

A data processing device according to a seventh aspect of the present invention is the data processing device according to any one of the first to sixth aspects, wherein the means for sorting the extracted keywords and detecting duplication of the keywords is provided. And means for deleting

A recording medium according to an eighth aspect of the present invention stores a computer program including program code means for extracting keywords from a plurality of types of files including a drawing information file, a tabular information file, and a document file, A computer-readable recording medium, wherein the computer causes the computer to determine a file type from an extension of a file from which a keyword is to be extracted. In a drawing defining a range, program code means for storing at least one point coordinate value defining a range including a title box for each drawing size, and when the file type is a drawing information file, the computer Then, the coordinates of the drawing are converted into coordinates with the predetermined point as the origin, and the coordinate transformation is performed. Program code means for extracting a keyword from a range defined based on the coordinate values of the at least one point, stored in association with the size of the drawing from which the keyword is to be extracted in a later drawing, and a file type Is a file of information in a tabular format, the computer inputs the data indicating the attribute of the data input to the item of each column to the item line composed of the item to which the data is input and / or is input to the item of each line. A program code for detecting an item string composed of items in which data indicating the attribute of the data is input based on the character type of the character input in the item, and extracting a keyword from the item in the item line and / or item string Means, when the file type is a document file, the computer causes the computer to decompose a character string having a predetermined number of characters from the beginning of the document into words. Characterized in that it comprises a program code means for extracting a keyword from the character string.

In the present invention, the file type is determined from the extension of the file from which the keyword is to be extracted, and if the file type is a drawing information file, the range is defined by coordinates with the origin at the lower left of the drawing, for example. In the drawing, for example, the coordinate value at the upper left of the title box is the size of the drawing (A0, A1, A2, etc.)
Separately stores and converts the coordinates of the drawing into coordinates with the lower left of the drawing as the origin, and converts the coordinates of the drawing into the upper left coordinates of the title box, which are stored in association with the size of the drawing. Based on, for example, keywords are extracted from the range defined by the upper left coordinate value of the title box and the lower right coordinate value of the drawing, and if the file type is a table format information file, the items in each column An item row consisting of items that have data entered, such as item names that indicate the attributes of the data that is being entered, and data that has been entered, such as item names that indicate the attributes of the data that has been entered for the items in each row The item string consisting of the item is detected based on the character type of the character entered in the item, the keyword is extracted from the item of the item line and item column, and if the file type is a document file, the beginning of the document To Table of Contents Decomposing the string of characters equal can include the word, a keyword is extracted from this string.
This simplifies the pre-setup work including keyword extraction and registration, and enables a multi-document system to be set up in a short period of time.

Further, in the present invention, fixed data such as item names such as “drawing number”, “drawing name”, “designer name”, “design date” are extracted from the title box of the drawing, and the fixed data is input. The data input to the item columns such as the right side and the lower side of the item column is extracted as a keyword of the fixed data. As a result, a search can be performed for each fixed data, and a target file can be searched in a short time.

In the present invention, a keyword is extracted from a range other than information in a table format. Thus, when the name of the form is input above the item row of the form, to the left of the item column, below the bottom line, etc., this name can be extracted as a keyword.

Further, according to the present invention, the extracted keywords are sorted, for example, for each file to detect keyword duplication, and the number of duplicate keywords is reduced to one. Thereby, the search time can be reduced.

According to the present invention, a data processing device such as a personal computer or a workstation reads a computer program on a recording medium, determines a file type from an extension of a file from which a keyword is to be extracted, and determines whether the file type is a drawing. In the case of an information file, for example, in a drawing whose range is defined by coordinates with the origin at the lower left of the drawing, for example, the coordinate value of the upper left of the title box is stored for each drawing size (A0, A1, A2, etc.). , Convert the coordinates of the drawing into coordinates with the lower left of the drawing as the origin,
Based on the coordinates of the upper left corner of the title block stored in association with the size of the drawing, for example, a keyword is extracted from a range defined by the upper left coordinates of the title block and the lower right coordinates of the drawing. If the file type is a table format information file, an item line consisting of items in which data such as an item name indicating the attribute of the data entered in the item of each column is entered, and an item in each line Detects an item string consisting of items in which data such as an item name indicating the attribute of the data entered in item is entered based on the character type of the character entered in the item, and detects the item row and item column If the file type is a document file, a character string having the number of characters that can include the table of contents, outline, etc. is decomposed into words from the beginning of the document, and the keyword is extracted from this character string. This allows computer programs to extract keywords from multiple types of files to CD-ROM, MO
And the like, or can be provided via a line.

[0022]

DESCRIPTION OF THE PREFERRED EMBODIMENTS FIG. 1 is an overall configuration diagram of the present invention. The document designation unit 1 designates a file name of a file from which a keyword is to be extracted from among the file names (also referred to as documents) displayed on the screen by allowing the user to select the file name by clicking the mouse, depressing an execution key, or the like. .

The document file 2 stores a plurality of types of files for constructing a multi-document system. FIG. 9 is a conceptual diagram of the document file 2, which includes a drawing (extension “.CAD”), a form (extension “.EXL”), a document (extension “.DOC”), and an image (extension “.TIF”). Multiple types of files such as ") are stored and managed by file names.

The document type determining means 3 determines the type of the file specified by the document specifying means 1 from the file name extension. In the extraction method determination file 4, this method is executed in association with a keyword extraction method such as drawing keyword extraction, form keyword extraction, and document keyword extraction, and a method name to be performed before the keyword extraction method determination such as character information extraction. The extension of the file from which the keyword is to be extracted is stored.
FIG. 10 is a conceptual diagram of the extraction method determination file.

The drawing keyword extracting means 5, the form keyword extracting means 6, and the document keyword extracting means 8 automatically extract keywords from the drawing, the form, and the document, respectively, by executing a procedure described later.

The character information extracting means 7 determines whether or not character information can be written out from a file such as a drawing, a form, or an image other than a document, for example, by a macro. The character information is written in such a work area, the document keyword extracting means 8 is started, and the keyword is automatically extracted from the character information of the character information work file. On the other hand, if the character information cannot be written, this file is stored in the character information unextracted file 9. FIG. 11 is a conceptual diagram of the character information unextracted file 9.

The keyword extraction file 10 includes, among the keywords extracted by the drawing keyword extraction means 5, the form keyword extraction means 6, and the document keyword extraction means 8,
For example, a keyword having a designated number of characters, such as two or more characters, is stored including an overlapping portion. FIG. 12 is a conceptual diagram of the keyword extraction file 10.

The keyword storage unit 11 sorts the keywords stored in the keyword extraction file 10 for each file, detects keyword duplication, deletes one or more duplicate keywords, and deletes only one. It is stored in the keyword file 12. In the keyword file 12, the keywords of each file are stored in association with the file names, with each keyword separated by a specific symbol such as ",". FIG. 17 is a conceptual diagram of the keyword file 12.

When the keyword input means 13 determines that character information cannot be written out of the file in the course of the keyword extraction process, or after a series of keyword extraction processes is completed, the operator inputs an appearance frequency file described later by the operator. The keyword input by reference or the like is stored in the keyword extraction file 10.

Next, an example of a specific procedure of drawing keyword extraction will be described with reference to the flowchart of FIG. 2, the explanatory diagram of the relative coordinate conversion of FIG. 3, the conceptual diagram of the upper left coordinate file of FIG. 13, and the fixed diagram of FIG. A description will be given based on a conceptual diagram of the item keyword extraction file. In the title block upper left coordinate file (Fig. 13), the origin (0,0),
When the upper left is (0,1), the lower right is (1,0), and the upper right is (1,1), the relative coordinate value at the upper left of the title box that defines the range including the title box is the drawing size ( Paper size).

The fixed item keyword extraction file (FIG. 14) is associated with item names such as “drawing number”, “name”, “(designer) name”, and “(design) date” in the title column of the drawing. Then, keywords of each file of the drawing are stored.

First, a drawing is read from the document file 2 (step S2-1). The coordinate values of the lower left and upper right of the drawing are determined, for example, as (0,0) and (1,1) (step S2-2),
As shown in FIG. 3, for example, the absolute coordinates of the drawing in which the lower left coordinate value is (-1, -1) and the upper right coordinate value is (2,1) are represented by (0,0) ), The upper right coordinate value is converted into a relative coordinate of (1,1) (step S2-3).

Next, referring to the upper left coordinate file (FIG. 13), a range including the title box (a range of the title box) is determined (step S2-4), and keywords in the title box are extracted (step S2-4). Step S2-5). Keyword extraction file for keywords with a specified number of characters (for example, two or more characters)
Write to 10 (step S2-6).

It is determined whether the keyword written in the keyword extraction file 10 is a fixed item such as an item name (step S2-7). In the case of the fixed item, the fixed item is extracted, and for example, the character string of the item on the right side of the item is written as a keyword in the fixed item keyword extraction file (FIG. 14) (step S2-8). It is determined whether it is not the last item or not (step S2-9). If it is not the last item, the process returns to step S2-7 to shift to the determination of the next item, and if it is the last item, the process ends. On the other hand, if it is not a fixed item,
Determine whether this item is the last item (step S2-9),
If it is not the last item, the process returns to step S2-7 to shift to the determination of the next item, and if it is the last item, the process ends.

In the above example, the absolute coordinates of the drawing are represented by (0,0) in the lower left, regardless of the drawing size such as A0, A1, A2, etc.
, The upper right is converted to relative coordinates with (1,1), but the lower left of the drawing is (0,0) (the upper right depends on the drawing size). May be stored, and the absolute coordinate values of the drawing may be simply converted into coordinate values with the origin at the lower left (0,0).

Next, the procedure for extracting the form keyword will be described with reference to FIG.
Will be described with reference to the flowchart of FIG. The form is read from the document file 2 (step S4-1), and the first line containing the most kanji, hiragana and katakana is determined as the item line, and the first line containing the most kanji, hiragana and katakana is determined as the item column. (Steps S4-2, S4-3). Next, as shown in FIG. 5, keywords such as a form name and an item name of each column are extracted from the range from the first line to the item line (step S4-4).
The keyword of the item string of each line below the item line is extracted (step S4-5). Among the extracted keywords, keywords having a designated number of characters, such as two or more characters, are stored in the keyword extraction file 10 (step S4-6).

Next, the procedure of character information extraction will be described with reference to the flowchart of FIG. 6 and the conceptual diagram of the character information extraction work file of FIG. The document is read from the document file 2 (step S6-1). The character information is written using a macro or the like (step S6-2), and it is determined whether or not the character information can be written (step S6-3).
If writing of the character information is not possible, the process is terminated and the document is stored in the character information unextracted file 9.

On the other hand, if the writing of character information is possible, the character information is written into a character information extraction work file as shown in FIG. 15 (step S6-4). A keyword is extracted from the character information written in the character information extraction work file by a document keyword extraction procedure described later.

Next, the procedure of document keyword extraction will be described with reference to the flowchart of FIG. The document is read from the document file 2 or from the character information extraction work file as shown in FIG. 15 (step S7-
1) A character string having a predetermined number of characters calculated from the number of characters considered to include the title and the "table of contents" from the beginning is extracted from the total number of specified characters, decomposed into words and written into a keyword extraction file ( Step S7-
2). It is determined whether the word “contents” is present in the extracted character string (step S7-3).

If there is a word "table of contents", a keyword is extracted from the contents of the table of contents and a keyword extraction file 10 is extracted.
(Step S7-4). Further, a character string of the remaining designated number of characters obtained by subtracting the above-mentioned predetermined number of characters from the entire designated number of characters is extracted, decomposed into words and written into the keyword extraction file 10 (step S7-5). On the other hand, if the word “table of contents” is not included in the character string having the predetermined number of characters, the process proceeds to step S7-5, and the character string having the remaining specified number of characters obtained by subtracting the predetermined number of characters from the entire specified number of characters Is extracted, decomposed into words, and written into the keyword extraction file 10 (step S7-5).

Next, the procedure for storing the keyword will be described with reference to the flowchart of FIG. 8, the conceptual diagram of the appearance frequency file of FIG. 16, and the conceptual diagram of the keyword file of FIG. Read the keyword extraction file 10 (Step S8
-1), sort keywords by file (step
S8-2). If the keywords are duplicated, the number of keywords is reduced to one (step S8-3). At the same time, the number of appearances of the sorted keywords in each file or all files is counted and written to the appearance frequency file as shown in FIG. 16 (step S8-4). For example, in the case of a drawing, keywords for each file excluding duplicates are selected according to a rule such as "Only two or more kanji, hiragana, and katakana are used as keywords except for the figure number and document number". Then, each keyword is written in the keyword file 12 by being separated by a specific symbol such as "," (step S8-5).

The keyword input means 13 stores in the keyword extraction file 10 a keyword that has been key-input by the operator by referring to an appearance frequency file, which will be described later. When the operator determines that character information cannot be written out of the document in the course of the keyword extraction process, or after a series of keyword extraction processes is completed, the operator extracts the keyword of the document stored in the character information unextracted file 9. For example, if the documents managed by the multi-document system belong to the same field, key input is performed by referring to the appearance frequency file.

After automatically extracting keywords from documents such as drawings, forms, documents, etc., and starting up the multi-document system in a short period of time as described above, when the documents are registered and replaced, the keywords are input. Re-registering by key input improves search accuracy.

The computer program for keyword extraction as described above can be provided by being preinstalled on a computer, or can be provided on a portable recording medium such as a CD-ROM or MO. Further, it can be provided via a line.

[0045]

As described above, the data processing apparatus and the recording medium of the present invention convert the coordinates of a drawing into a predetermined point, for example, coordinates with the origin at the lower left, and determine the size of the drawing and the position of the origin. Regardless, we detect a range that includes the title block and automatically extract keywords from this range,
In addition, based on the character type of the character entered in the item, for example, whether or not there is a character other than a numeral, and whether or not the number of characters is the largest, based on the character type of the character entered in the item. And automatically extract keywords from these rows and columns. Therefore, keywords are extracted from a predetermined number of character strings at the beginning of the document, which are likely to contain a table of contents, summary, etc. Because it extracts automatically, it simplifies the pre-setup work including keyword extraction and registration,
It has an excellent effect that a multi-document system can be started up in a short time.

[Brief description of the drawings]

FIG. 1 is an overall configuration diagram of the present invention.

FIG. 2 is a flowchart of drawing keyword extraction.

FIG. 3 is an explanatory diagram of relative coordinate conversion.

FIG. 4 is a flowchart of form keyword extraction.

FIG. 5 is an explanatory diagram of form keyword extraction.

FIG. 6 is a flowchart of character information extraction.

FIG. 7 is a flowchart of document keyword extraction.

FIG. 8 is a flowchart of keyword storage.

FIG. 9 is a conceptual diagram of a document file.

FIG. 10 is a conceptual diagram of an extraction method determination file.

FIG. 11 is a conceptual diagram of a character information unextracted file.

FIG. 12 is a conceptual diagram of a keyword extraction file.

FIG. 13 is a conceptual diagram of a title column upper left coordinate file.

FIG. 14 is a conceptual diagram of a fixed item keyword extraction file.

FIG. 15 is a conceptual diagram of a character information extraction work file.

FIG. 16 is a conceptual diagram of an appearance frequency file.

FIG. 17 is a conceptual diagram of a keyword file.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Document designation means 2 Document file 3 Document type judgment means 4 Extraction method judgment file 5 Drawing keyword extraction means 6 Form keyword extraction means 7 Text information extraction means 8 Document keyword extraction means 9 Text information unextracted file 10 Keyword extraction file 11 Keyword storage Means 12 Keyword file 13 Keyword input means

 ──────────────────────────────────────────────────続 き Continued on the front page F term (reference) 5B075 ND06 NK32 NK37 NR05 NR14 UU21 5B082 EA08

Claims (8)

[Claims] 1. A data processing device having a function of extracting a keyword from a drawing information file having a title box, wherein a drawing defines a range by coordinates with a predetermined point as an origin.
Means for storing coordinate values of at least one point defining a range including the title box for each size of the drawing; means for converting the coordinates of the drawing into coordinates having the predetermined point as the origin; Means for extracting a keyword from a range defined based on the coordinate value of the at least one point, stored in association with the size of the drawing from which the keyword is to be extracted. . 2. The title column has a plurality of item columns for inputting data unique to the drawing and fixed data indicating the attribute of the data, and extracts the fixed data input in the item column. Means and the data entered in the item column at a position adjacent to the item column in which the fixed data is entered,
2. A data processing apparatus according to claim 1, further comprising means for extracting the fixed data as a keyword. 3. A data processing apparatus having a function of extracting a keyword from a file of information in a tabular format, wherein an item row including an item in which data indicating an attribute of data input in an item of each column is input. And / or means for detecting, based on the character type of the character input to the item, an item string consisting of the item in which data indicating the attribute of the data input to the item in each line is input, and Or a means for extracting a keyword from the items in the item string. 4. The data processing apparatus according to claim 3, further comprising means for extracting a keyword from a range other than the information in a table format. 5. A data processing apparatus having a function of extracting a keyword from a document file, comprising: means for decomposing a character string having a predetermined number of characters from the beginning of the document into words; and means for extracting a keyword from the character string. A data processing device comprising: 6. A data processing apparatus having a function of extracting a keyword from a plurality of types of files including a drawing information file, a tabular information file, and a document file, wherein a file from which the keyword is to be extracted is extended. Means for determining the type of file from a child, and a drawing in which a range is defined by coordinates with a predetermined point as an origin,
Means for storing at least one point coordinate value defining a range including the title box for each drawing size, and, if the file type is a drawing information file, drawing coordinates are set to coordinates with the predetermined point as an origin. Means for extracting a keyword from a range defined based on the coordinate values of the at least one point, wherein the keyword is stored in association with the size of the drawing from which the keyword is to be extracted, in the drawing after the coordinate conversion. When the file type is a file of the information in the table format, the file is entered in the item line of the item in which the data indicating the attribute of the data entered in the item of each column is entered and / or the item of each line. An item string consisting of items in which data indicating data attributes is input is detected based on the character type of the character input in the item, and a key is detected from the item in the item line and / or item column. And a means for extracting a character string having a predetermined number of characters from the beginning of the document into words and extracting a keyword from the character string when the file type is a document file. Data processing device. 7. The apparatus according to claim 1, further comprising: means for sorting the extracted keywords to detect duplication of the keywords; and, when the keywords are duplicated, means for deleting a keyword other than one. Data processing equipment. 8. A computer program including program code means for extracting a keyword from a plurality of types of files including a drawing information file, a tabular information file, and a document file is recorded, and can be read by a computer. A program code means for causing the computer to determine a file type from an extension of a file from which a keyword is to be extracted; and A program code means for storing at least one point coordinate value defining a range including the title box for each drawing size; and when the file type is a drawing information file, the computer stores the drawing coordinates in a file. The coordinates are converted to coordinates with the predetermined point as the origin, and the Program code means for extracting a keyword from a range defined based on the coordinate values of the at least one point, stored in association with the size of the drawing from which the word is to be extracted; In the case of (1), the computer is provided with an item row composed of items in which data indicating the attributes of the data input to the items of each column are input, and / or data indicating the attributes of the data input in the items of each row. Is an item string consisting of items for which
Program code means for detecting based on the character type of the character input to the item and extracting a keyword from the item in the item line and / or the item column; and when the file type is a document file, the computer A recording medium comprising program code means for decomposing a character string having a predetermined number of characters from the beginning into words and extracting a keyword from the character string.