JPH10187680A - Document repository device managed by word, sentence and grain degree of part - Google Patents

Abstract

PROBLEM TO BE SOLVED: To retrieve or edit partially necessary information by the unit of a document part concerning a structured document by managing the definition of document type definition(DTD) based on a scheme and managing a document part stuck with a tag defined by DTD based on the scheme. SOLUTION: An object-oriented data base 3 is introduced to a document repository device. Then, in a document structure managing part 1 managing the constitution (chapter number, chapter headline, paragraph, a graphic, etc.) of a structure document 4 based on DTD of standard generalized mark up language(SGML), a DTD managing part 11 manages the definition of DTD based on a prescribed scheme and a document element managing part 12 manages a document part stuck with a tag defined by DTD based on the prescribed scheme. In addition, a part linkage managing part 2 manages the referring relation of a word, a sentence and the grain degree of a part separately from the constitution of the document 4.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION The present invention relates to a method for processing words in a document,
The present invention relates to a document repository device that manages a structured document assembled according to a structure having a plurality of document components by attaching tags to sentences and portions.

As will be described later in detail, the management performed by the document repository device of the present invention is the definition of the structure between document parts of a structured document to be managed and its retention, and this management is performed. Thus, it is possible to perform a search, a maintenance of consistency, and a batch editing of document parts according to the above structure.

[0003] The document repository device of the present invention comprises:
In addition, a text file and a binary file are indexed into a binary file, and the index is treated as text, so that the index can be managed in a unified manner.

[0004]

2. Description of the Related Art ODA is a basic standard for defining a document structure.
(Open Document Architecture
re). In this method, the entire document is organized in a tree structure from two viewpoints, a logical viewpoint and a layout viewpoint.

[0005] From a logical point of view, constituent units such as paragraphs and figures are used as basic logical objects, and these are collected to form chapters and the like, and finally the logical structure of the document is formed as a tree structure that becomes the entire document. . By having a logical structure, it is possible to support editing of logical units such as paragraphs and chapters, protection of logical units at the time of exchange, and automatic numbering of logical units.

On the other hand, in the layout viewpoint, units such as pages, initials, frames (diagrams and tables), and sections are collected, and the entire document is expressed as a tree structure, similarly to the logical viewpoint.

Another standard relating to the document structure is SG.
ML (Standard Generalized)
Markup Language (Standard Generalized Markup Language). This is the document type definition DTD (DocumentType Define
ion; document type definition) defines a tag to be attached to a text document element (heading, body, list, etc.). In the DTD, the parent-child relationship, the order of appearance, the number of appearances, and the like of the tag can be defined in a tree structure. Text written in SGML according to this grammar is called a structured document.

FIG. 2 shows an example of a tree structure notation using a letter as an example. Referring to FIG. 2, the tag "lett"
The document parts surrounded by “er” include tags “date” (date), “to” (address), “body” (text), and “f”.
rom "(sender) and" ps "(postscript).

These child parts appear in the order described above, and must always appear once. The document part of the last “ps” tag is provided with a “?” Symbol, indicating that it appears 0 or 1 time.

A document part surrounded by a "body" tag has a child part having a "p" (paragraph) tag, and the number of appearances is represented by a "+" sign at least once.

[0011] A document part enclosed by "p" tags is "h".
It has a child part having a “p” (highlighted) tag, and the number of appearances is represented by “*” sign indicating 0 or more times.

Both ODA and SGML express a document in a tree structure of constituent elements. The difference is that ODA is one of the standard languages, whereas SGML is defined by the DTD tags.
It is a meta-language for languages that describe documents.

An example of the prior art relating to a structured document processing apparatus based on two standards is disclosed in, for example, Japanese Patent Application Laid-Open No.
JP-A-301721 discloses a relational database (RD) for full-text search of documents having a logical structure.
The management of the SGML document using B) is described.

In the method proposed in Japanese Patent Application Laid-Open No. 6-301721, an original document described in SGML is used as input data, and based on the SGML, each table of its structural information, tag information, and text information is converted to DDL ( Data
Definition Language). Then, each table is generated and stored in the RDB. The data of the original document data using SGML is also stored. The search of the document is performed by the search language DQL (Docume).
nt Query Language), an input of a search request designating a document structure is received, and a relational database management system (RDBMS) receives the search request.
A search request in the QL format is made. The search result is displayed by organizing the result of the search request into an SGML format including a tag using the structure information in the RDB.

For example, Japanese Patent Application Laid-Open No. 6-314281 discloses a common logical structure having a small data amount from a specific logical structure of a structured document used when document data is exchanged between different document processing systems. An apparatus for generating a structured document having the same has been proposed, and an object-oriented idea is introduced in generating a common logical structure described in ODA or the like.

That is, in the document editing apparatus proposed in Japanese Patent Application Laid-Open No. 6-314281, a class corresponding to each object is generated on a one-to-one basis based on a specific logical structure, and an attribute of each class is generated. Describes a document editing apparatus that copies from a corresponding object and generates a class of a common logical structure by associating the hierarchical structure between the classes with the hierarchical structure between the corresponding objects as it is. The class is given, for example, in correspondence with each document part such as a chapter number, a chapter heading, a paragraph, and a figure. In addition, a class part-aggregation relationship is introduced such that the four parts are set as lower classes, which are aggregated to give a higher class chapter class.

For example, Japanese Patent Application Laid-Open No. Hei 6-119331 proposes a document part search apparatus which enables a structured document to be searched for by using a frame or a headline in the document.
In this method, the logical structure of the document part is specified, and the specified logical structure is managed by attaching an index. The search based on the frame or the heading is performed by conditionally matching the document part corresponding to the frame or the heading with information on the attribute of the document part obtained by searching the added index. As a specific example, 3
Each of the documents A, B, and C has a frame inside the document,
A search for a storage device holding a table frame, a graphic frame, and a graph frame as an index is shown. As a search,
A document part “frame” is included in the document, and a search is made for a document that is a graph frame.

In addition to document structuring techniques such as ODA and SGML, there is hypertext as a technique related to the present invention. This is WWW (World Wide Wed)
b; HTML (Hyper Text Ma) used for displaying resources on a World Wide Web (BW) browser
rkup Language (Hypertext Markup Language) is typical, and HTML is realized by one DTD with SGML.

HTML can create a part on the WWW browser that can be clicked with a mouse called an anchor. Specify a link to this anchor that points to another HTML resource or another location of the same HTML resource.
It can be moved to that location and displayed.

The designation of this link is possible for resources of any server machine on the Internet (except for server machines and resources whose access is restricted by security), thereby enabling the current WWW world to be established. Building.

[0021]

However, the above-mentioned prior art has the following problems.

(1) First, in the method proposed in JP-A-6-301721, RD is stored in a database.
B (relational database) has been introduced. The structure of the document parts in the structured document is a tree structure representing a parent-child relationship and an inclusion relationship, and a network structure representing a reference relationship. However, the table format handled by the RDB has a large gap in expressing such a structure.

For a search request for a document part extending between tables, a join operation between tables must be performed in the RDB, and this operation generally requires a lot of calculation time.

(2) Next, in the document editing apparatus proposed in the above-mentioned Japanese Patent Application Laid-Open No. 6-314281, the relationship between parts is limited to only the classification of partial and aggregated relationships, and the general relationship between classes is described. However, it does not describe the modeling of the class relating to the reference relation. For example, one document part is related to another document part over a plurality of structured documents, and generation of a common logical structure such as linking the relation in search is not considered. For this reason, it is not possible to generate an object of a structured document that enables a navigation search to follow a link between structured documents.

(3) Further, the above-mentioned Japanese Patent Application Laid-Open No. Hei 6-119331
The document part search device proposed in Japanese Patent Application Publication No. JP-A-2005-26095 is for searching for a document including the attribute by designating the attribute attached to the frame, for example, a geometric figure, a raster figure, a table, a mathematical expression, or the like. It is.

However, the search range does not extend to the inside of the frame. For example, it is impossible to search for a document that also includes a table including the heading (search key) by using what is the heading of the table column as a search key. is there.

This is because the above-mentioned Japanese Patent Application Laid-Open No. Hei 6-119331 is used.
In the document part search apparatus disclosed in Japanese Patent Application Laid-Open No. H10-209, only one index is attached to the entire frame, and the index also pays attention to the type of the frame. .

(4) HT represented by WWW
The great advantage of ML's hypertext technology is that you can freely specify the link, but on the other hand,
Not structural.

Here, "not structured" means that no link is set on the assumption that a certain resource and another resource have a certain relationship.

In an environment where links can be freely set,
Intensive management of link relationships is extremely difficult. That is, even if all the link destinations from a certain resource at a certain point in time are grasped, it is expected that the link destination resource will disappear or a link to a new resource will be added. It means that the pipe cannot be divided.

Accordingly, the present invention has been made in view of the above problems, and has as its object to manage structured documents in a prescribed structure and to partially manage the managed structured documents in units of document parts. It is an object of the present invention to provide a document repository device capable of searching for or editing information necessary for a document.

Another object of the present invention is to provide a document repository device capable of assigning and managing an index to a document such as a binary file whose contents cannot be tagged with a tag. is there.

[0033]

Means for Solving the Problems To achieve the above object, the first invention of the present application is an SGML (Standard G).
energyalized Markup Langua
Ge) DTD (Document Type Def)
In order to manage a structured document based on the DTD, the definition of the DTD is managed based on a schema.
It is characterized by including TD management means and document element management means for managing document parts to which a tag defined by the DTD is attached based on a schema.

Further, the second invention of the present application is a DTD of SGML.
A meta-model managing unit that manages a model of the relation of the link as a schema in order to manage a link with another structured document at a granularity of a word, a sentence, and a part included in the structured document based on Tag management means for defining and managing a tag for newly linking according to a schema, a document part to which the tag is attached and its link relation,
Document part management means for managing the appearance position of the document part in the structured document.

The third invention of the present application provides the DTD management means and the document element management means of the first invention of the present application, and the meta model management means, the tag management means and the document part management means of the second invention of the second invention. Using the object-oriented database that manages the schema definition of document parts and manages them as instances, manages consistency between document parts, deletes, adds and updates document parts, and batch edits document parts Is made possible.

The fourth invention of the present application assigns an index to a binary external file outside a structured document that is not a text format, and records the relationship between the contents of the index and the binary external file indicated by the index. Also, the present invention is characterized in that management can be performed in a binary format in the same framework as the document part management system described in any one of the first to third embodiments of the present application.

[0037]

Embodiments of the present invention will be described below with reference to the drawings. FIG. 1 is a block diagram showing a basic configuration of a document repository device according to an embodiment of the present invention.

Referring to FIG. 1, the present invention, in its preferred embodiment, comprises an object-oriented database (Object Oriented Data Base).
e; “OODB”) 3 is introduced into the document repository device. This OODB has a small semantics gap between a conceptual model representing the connection of document parts and an object-oriented schema implementation.

In the embodiment of the present invention, OO
Since a method of storing the connection of objects in a database in advance using a DB is adopted, the navigation search between objects is performed more efficiently than the method of calculating the connection between tables when a search is executed by issuing SQL as in an RDB (relational database). Be faster. As a result, the above-mentioned problem (1) relating to Japanese Patent Application Laid-Open No. 6-301721 is solved.

The above-mentioned problem (2) relating to the above-mentioned Japanese Patent Application Laid-Open No. 6-314281 and HTML represented by WWW
In order to solve the above-mentioned problem (4) of the hypertext technology of the present invention, the present invention provides, in a preferred embodiment thereof,
Referring to FIG. 1, a document structure management unit 1 that manages a document structure (chapter number, chapter heading, paragraph, graphic, etc.) based on the SGML DTD, a word in the document, A component link management unit 2 is provided for managing the reference relationship between the granularity of a sentence and a part.

The document structure management section 1 comprises a DTD management section 11 and a document element management section 12.

The component link management unit 2 includes a meta model management unit 21, a tag management unit 22, and a document component management unit 23.
And

The structured document 4 is registered in OODB through these five management devices.

FIG. 3 is a diagram for explaining the operation of the embodiment of the present invention. In the registration to the structured document OODB, five management devices (ie, the DTB management unit 1)
1, the document element management unit 12, the meta model management unit 21, the tag management unit 22, and the document component management unit 23) show the overall flow and individual functions.

Referring to FIG. 3, both the document structure management unit 1 and the component link management unit 2 first generate a class of a common structure such as the DTD determination 111 and the metamodel definition 211.

This approach differs from the conventional hypertext in that a link structure in a document is defined in advance by a meta model, and the model is managed as a schema of OODB. That is, it does not allow links between resources beyond the scope of the schema.

As a result, the degree of freedom of the hyperlink is reduced. However, as described above, the document repository device of the present invention can centrally manage the link relation of the resource, which is difficult with the hypertext. And

Further, the navigation utilizing the structure of the meta model defined in advance can be performed by the document repository device of the present invention.

The above-mentioned approach is described in Japanese Patent Application Laid-Open No.
This is also different from the approach described in Japanese Patent No. 14281 for generating a class of a common logical structure from each existing structured document. This is because the approach of the present invention is to first generate a common logical structure of the structured document to be generated, and to use this to generate and manage objects of each structured document.

In order to register a structured document in OODB, referring to FIG.
Then, a DTD definition of SGML is prepared. Then, in the schema and schema definition processing 112 of the tag and its tree structure, the tag is represented by an object model regarded as a class, and based on this,
Generate OODB schema.

FIG. 4 shows the schema.
It models the tags included in the DTD definition, the attributes of the tags, and the parent-child relationship of the tags. FIG. 4 shows OMT (O
project Modeling Technology;
Object modeling techniques).

The DTD management unit 11 stores the tag 3 shown in FIG.
The schema definition (including the association between classes) of 1 and the tag attribute 32, and individual tags and tag attributes that are instances thereof are managed using OODB.

Next, document instance creation processing 121
Then, a document instance according to the DTD determined in the DTD determination processing 111 is created. DB registration of document parts 12
In step 2, the document parts in the document instance are associated with the tag and the tag attribute schema generated in the schema definition processing 112 of the tag and its tree structure.

At this time, image processing is performed in which the tags and their attributes defined in the DTD determination processing 111 are determined on attribute values and attached to document parts.

An example of the relationship (schema) between a tag and a tag attribute, a document part and the value of a tag attribute of the document part is shown in FIG.
Shown in

The document element management unit 12 (see FIG. 1)
Document part 33 and document part attribute 3 based on the schema shown in FIG.
4 is generated and managed by OODB.

By this instance management, the document parts corresponding to the instance of the schema in the actual document instance are managed.

The structured document stored in the steps of the DTD determination processing 111, tag and tree structure schema definition 1 processing 12, document instance creation processing 121, and document part DB registration processing 122 is structured. When a link is managed based on the granularity of a word, a sentence, or a part in a document, the management is performed using a component link management unit 3 described below.

First, in the meta model definition 211, a model of the link relation of the entire structured document is defined. An object-oriented analysis technique represented by OMT or the like can be used to define the link relation model. That is, tags can be associated with one class, and a general reference relation, a part, an aggregation relation, a generalization-specialization relation, a multiplicity of relations, and the like can be defined using the class.

Further, by adding attributes and operations to a class, an operation common to document parts belonging to the class can be described. For example, an operation for performing batch editing on document parts with the same tag (that is, belonging to the same class) is described, restrictions on operations are added by adding special attributes, and navigation from a certain document part is performed. Various definitions can be given to the metamodel, such as describing operations for gate search. If a class for a model of the link relation of the tag is referred to as “tag link” abstractly, the link tag inherits the tag 31 of the schema shown in FIG. 5 and inherits the schema shown in FIG. Is the component class of.

Referring to FIG. 6, id which is an attribute of tag link 35 is one of tag attributes 32, and has a name of id and a positive type. The id is for identifying the instance of the link tag 35 (that is, the actual tag) by attaching a serial number attribute thereto.

Referring again to FIG. 3, in the meta model schema definition 212, the meta model (object model) defined by the meta model definition 211 is converted into an OODB schema.

The meta model management unit 21 generates a schema of the converted schema and registers it in OODB.

In the tag definition 221 based on the meta model, the name (character string) of the tag actually attached to the document instance is defined for each class defined in the meta model.

In the tag DB registration 222, the tag management unit 2
2 generates an instance of each class corresponding to the link tag of the meta model shown in FIG. As an argument for this instance generation, a tag definition 22 based on the meta model
Gives the name (character string) of the tag defined in 1. As a result, each tag whose name is determined is registered in OODB.

In attaching the tag 231 to the document,
The tag defined by the tag definition 221 based on the meta model is attached to an actual document instance, and a document part having a granularity of words, sentences, and parts is created. In the DB registration 232 of the document part, the instance of the link part 36 which is inherited and added to the document part 33 shown in FIG. 7 is generated.

Referring to FIG. 7, an instance of link component 36 has a pointer (an instance of relationship 37) to an instance of link tag 35 defining a tag, and this pointer is used as an instance of link component 36 (ie, document). Identify which tag is attached to the document part in the instance).

As indicated by the relation 38, the instance of the link component 36 records which document component the parent has in accordance with the structure tree of the entire structured document.

The document part management unit 23 stores the relations 37, 3
An instance of the link component 36 including the link 8 is registered and registered in the OODB, and a document component having a link relationship based on the granularity of a word, a sentence, or a part is managed.

In order to solve the problem related to the management of the link relation in the HTML hypertext technology represented by WWW, as described above, an instance including the link relation of the document parts is generated and registered in the OODB. .

When a document part is deleted or added, the document part must follow the underlying metamodel.

There is an existence restriction on deletion. If there is a document part of a reference destination and a reference source with respect to a link, and these are registered in OODB as instances, deleting the instance of the reference destination will update the attributes of the instance of the reference source. OODB can sense the effect of deleting one instance on other instances (this feature is implemented).
For this reason, the document repository apparatus of the present invention using OODB can alert a user who operates a database or prohibit deletion.

As for the addition of document parts, there is the presence / absence of association indicated by the meta model, the cardinality restriction, and the like. This is because classes cannot be linked between classes that are not defined in the metamodel, and no instance can be created beyond the number of instances that other classes associated with a class can create. is there.

If it is necessary to store a certain document part in the document repository device, it is necessary to modify the meta model so as to correspond to the addition of the document part and to extend the schema of OODB.

Next, in order to solve the above-mentioned problem (3) of the prior art described in the above-mentioned JP-A-6-119331,
In the embodiment of the present invention, the index for the frame is a document part in the document instance, a necessary number of links to the frame are provided from this document part, and the internal information of the frame is stored in the document part as the link source. Use a method of expressing the content. Some document editing editors, such as word processors, render figures, tables, and graphs in text format.However, in general, figures, tables, and graphs are edited using another dedicated editor, and externally edited as binary files. These are referred to in the document.

In the embodiment of the present invention, the relationship between the tag attached to the reference portion of the external file and the tag representing the index portion is represented by a meta model. FIG. 8 is a diagram showing an example of this meta model, in which 41, 42,
43 is a meta model. Note that the meta model corresponds to the link tag 35 in FIG.

Referring to FIG. 8, an instance is generated from each metamodel class (41 and 42), and a tag name is determined (<IDX id = "1"> 441,
<IDX id = “2”> 442, <GAIBU id =
“1”> 45).

At the time of instance generation, a link between the instances is established by the relation 43 between the classes of the meta model. The tag generated in this way is attached to the index part and the external file reference part in the document instance, and a new document part is created. FIG.
FIG. 1 shows an example of a new document component.

Referring to FIG. 9, link component 3 shown in FIG.
Document part "room layout" 51, which is an instance of 6
Are tagged by a tag 441 (<IDX id = “1”>), and the document part “2DK” 52 and “two kitchens in one room” 53 are also tagged 442 (<IDX id = “1X”).
“2”>).

Similarly, the document part “<IMGSRC =“ m
adori. gif ">" 54 is the tag 45 (<GAI
BU id = “1”>).

In FIG. 9, “2DK” 52 and “two rooms and one kitchen” 53 have the same tag (<IDX id
= “2”>) means that the same treatment is performed at the level of the link tag 35 even if the content of the attribute of the link component 36 is different. For example, the same treatment of synonyms is used.

Each instance of the link tag 35 and the link component 36 is managed by the component link management (2 in FIG. 1), which enables indexing and searching of an external file such as a binary file.

As described above, the document repository device according to the embodiment of the present invention manages a structured document in a certain prescribed structure, and searches the managed structured document for necessary information partially in units of document parts. It is suitable for use when editing or editing. Also, it is suitable for use in a case where an index is given to a document such as a binary file to which a tag cannot be attached to its contents and managed.

[0084]

DESCRIPTION OF THE PREFERRED EMBODIMENTS In order to describe the above-mentioned embodiment of the present invention in more detail, an embodiment of the present invention will be described below.

[Document Database Registration] Registration to a document database (OODB) will be described as an embodiment of the present invention. FIG. 10 shows DT as an embodiment of the present invention.
4 shows an example of document registration based on a D-tree structure. Here, there is no particular attribute of the tag. The DTD definition is registered in the object-oriented database (OODB) using the schema model shown in FIG.

Referring to FIG. 10, tags 61 to 68 have names of “doc (document)”, “sec (chapter)”,
“Subsec”, “p” (paragraph), “hp” (emphasis), “fig (figure)”, “list”,
This is an “item” and is registered in the OODB 3 by the DTD management unit 11 (see FIG. 1) as an instance of the tag 31 shown in FIG.

The parent-child relationship of the tag is also registered in OODB3 (see FIG. 1) in the relationship as shown in FIG.

Next, the tag defined above is attached to an actual document, an instance of a document part is generated, and OODB3
Register with. FIG. 11 is a diagram illustrating an example of the document instance.

When attaching a tag to a document, the document element management unit assigns the tagged document part as an instance of the document part 33 according to the schema model shown in FIG.
Register in ODB3. In addition, parent-child relationship 3 at the document parts level
8 is also registered in OODB3 (see FIG. 1).

[Registration of Component Link of Structured Document]
As an embodiment of the present invention, registration of a component link of a structured document will be described. Taking the structured document shown in FIG. 11 as an example,
A procedure for registering a link relationship between document parts having the granularity of a word, a sentence, or a part will be described.

FIG. 11 is a part of a basic specification of a system for realizing an electric commerce order processing.
Upon receiving an order from a user through a GUI (graphical user interface) displayed on a computer terminal, the system creates a slip for the order and updates the stock of the product.

The basic specifications include words representing classes, associations, attributes, and operations used in object-oriented analysis / design, and these words are assumed to have a certain structural relationship.

Further, apart from FIG. 11, an image diagram of a GUI for performing order processing as an external file in a binary format,
Assuming that there is an object diagram representing a relationship such as a class or an attribute, a relationship between these binary diagrams and words in the structured document shown in FIG. 11 is provided, and a structural relationship is provided as a whole.

First, a meta model is defined. The schema class (class modeled for defining the tab) used for defining the meta model is a link tag 35 shown in FIG.
, And all classes of the metamodel to be defined inherit the tag 31.

FIG. 12 is a diagram showing a meta-model between structured documents relating to the order processing of electric commerce. Referring to FIG. 12, reference numerals 71 to 74 represent "class", "association", "attribute", and "operation" used in object-oriented analysis / design, and their structural relationships are modeled. . Reference numerals 75 to 78 denote models related to binary files, which are modeled using a generalization-specialization relationship.

Class 71, Association 72, and Attribute 7 in FIG.
3. The relationship between the operation 74 and the “drawing” 76 is the same as the relationship 43 between the index 41 and the external file 42 shown in FIG. 8, and 71 to 74 are the indexes of the “drawing” 76.

The meta model shown in FIG. 12 is converted into an OODB schema. The meta model management unit generates a schema of the converted schema and registers it in OODB.

Next, a tag name is given to each class of the meta model, an instance is generated, and registered in OODB. The names of the tags are "cl", "r" in order from 71.
l "," at "," mt "," bf "," fig ",
“Gui” and “ob”.

The created tag is attached to FIG. 11, and an instance of the link component 36 in FIG. 7 is generated.

FIG. 13 shows a part of an example in which a tag is attached to the document instance of FIG. The arrangement of the tagged document parts appearing in FIG. 13 is as shown in FIG. However, FIG. 14 is a table arranged for convenience of explanation, and is not stored in the OODB in such a table format in the document repository device of the present invention.

Referring to FIG. 14, id in the tag is an attribute of link tag 35 shown in FIG. In this example, the document parts “product name”, “price”, and “description” are all instances of the same class “attribute” 73 (see FIG. 12) of the metamodel, but these are identified at the instance level. Is represented by a serial number of id.

The document parts "product name" and "information" have the same tag <at id = "1">. This is used to indicate the same meaning at the instance level. This is effective, for example, when performing synonymous or clustering of document parts in a certain sense. This relationship is an example of the implementation of the schema of the relationship 37 in FIG. 7 in which a plurality of link components exist for one link tag.

Next, these document parts are registered in the OODB as instances of the schema shown in FIG. 7 and the metamodel shown in FIG. Based on the meta model, an instance relationship as shown in FIG. 15 is set as an example.

Regarding the parent-child relationship of document parts (relationship 38 in FIG. 7), according to the example of the document instance in FIG. 13, all of the document parts are included in the first paragraph <p> order form of the section <subsec> system scenario. Included in. Therefore, all of the document parts have a parent part in this paragraph <p> order form. This relationship is registered in OODB, and the location of the document component in the document instance is recorded.

The link relationship with the “drawing” 76 in FIG. 12 is set by the method described with reference to FIG.

Of the instances in FIG. 15, only the tags to be used as indexes of the GUI image diagram or the object diagram, which are binary files, are linked with the tags of FIG.
Register in DB.

[Navigation Search Based on Document Component Structure] Next, as one embodiment of the present invention, a navigation search based on the document component structure will be described. FIG.
Taking 2 as an example, for example, it is possible to search for a document part according to the structure of the meta model listed below.

List all "attribute" and "association" instances of a certain "class" instance.

List an instance of a certain “class” and an instance of a “related” of a certain “class”.

List the "operation" instances of instances of "classes" that have instances of "attributes" contained (indexed) in instances of certain "object diagrams".

Referring to FIG. 15 as an example, if all the instances of the attribute of the class “product” can be listed, the attributes “product name”, “price”, and “description” are returned. In FIG. 14, if all the document part instances having the tag <at id = "1"> are to be listed, the document parts "product name" and "information" are returned.

[Consistency Management Based on Document Component Structure]
Next, as one embodiment of the present invention, consistency management based on the structure of a document component will be described. In the example shown in FIG. 12, when an instance of a certain "class" is deleted from the existence constraint, the instances of "attribute" and "operation" related thereto are affected. When deleting an instance of the "class", the link between the "attribute" and "operation" instances that refer to the instance of the class must be deleted. In some cases, it is necessary to delete the instances of the “attribute” and the “operation”.

In the example shown in FIG. 15, when the class “customer” is deleted, the operation “purchase check box...” Is an operation that is no longer necessary, and an instance of this operation is also deleted at the same time. .

With regard to the restrictions on the addition of an instance, referring to FIG. 12 as an example, tags that can be attached to the “FIG.” 76 as indexes are <cl>, <rl>, and <a>.
t> and <mt> only, and other tags such as <bf> cannot be linked as an index of “figure” 76.

A constraint on instance updates is whether it is guaranteed that an update of an instance is a valid association for its associated instance.

Referring to FIG. 12 as an example, when an instance of the “GUI image diagram” 77, that is, the GUI image diagram itself is largely rewritten, the class 71 assigned as the index and the association 7
2. It is necessary to reconsider whether each instance of the attribute 73 and the operation 74 is valid as an index. In this case, the fact that an update constraint exists is recorded in the association between the “drawing” 76 and 71 to 74, and when an update operation occurs, all instances that affect the update are listed. Check for updates one by one.

[Batch Editing of Document Parts Based on Structure of Document Parts] Next, as one embodiment of the present invention, batch editing of document parts based on the structure of document parts will be described. The document repository device of the present invention manages document components by providing a metamodel, tags, document components, and three abstract hierarchies. When batch editing document components in a structured document according to a certain rule, editing operations can be performed from any of the three layers.

At the metamodel level, it is possible to change the structure of the metamodel, change the values of the document part attributes of all instances (document parts) belonging to a certain schema class, or delete all instances. is there.

The same operation can be performed on the document parts assembling at the tag level and the document part level. In addition, these operations can be performed based on the consistency management.

Referring to the example shown in FIG. 14, an example of the batch editing will be described.
If the document part of the tag with the d-number is to be rewritten to "product name", all the document parts "information" with <at id = "1"> in the structured document are replaced with "product name".

[0121]

As described above, according to the present invention, the following effects can be obtained.

According to the first invention of the present application (claim 1), SG
Structured documents based on the ML DTD can be managed by OODB.

As described above, by managing in the OODB, the parent / child relationship of the document parts in the structured document, the tree structure representing the inclusion relationship, the network structure representing the reference relationship, etc.
Bridging the semantic gap with human intuitive images. Regarding the navigation search, the RDB can perform the search faster than the join operation method between tables.

According to the second invention of the present application (claim 2), the SGML DTD definition, which focuses on the definition of the parent-child relationship and the inclusion relationship of the document parts in the structured document, is represented by the hypertext technology. It is possible to manage the link relation between the documents to be created with the granularity of words, sentences and parts. Furthermore, a structural management element is added to the hypertext link mechanism to solve the difficulty of maintaining the link relationship due to the disappearance of resources and the like.

According to the third invention of the present application (claim 3), a navigation search between document parts for a structured document possessed by the document repository device, consistency management regarding deletion, addition and updating of document parts, batching of document parts The editing function facilitates editing of the structured document according to a predetermined rule and extraction of necessary information.

According to the fourth invention of the present application (claim 4), a framework is provided which treats a binary file in the same manner as a text file by giving an index representing the contents thereof, and provides a framework for managing structured documents. It gives unity.

[Brief description of the drawings]

FIG. 1 is a diagram showing a configuration of a document repository device according to an embodiment of the present invention.

FIG. 2 is a diagram illustrating a DTD tree structure notation of a letter.

FIG. 3 is a diagram illustrating a processing operation according to the embodiment of the present invention, and is a diagram illustrating a flow of document database registration by a document structure management unit and a document link management unit.

FIG. 4 is a diagram for describing an embodiment of the present invention, and is a diagram illustrating a schema model representing a DTD tag, a tag attribute, and a parent-child relationship of the tag.

FIG. 5 is a diagram for explaining the embodiment of the present invention, and is a diagram showing a schema model representing a relationship among tags, tag attributes, document components, and document component attributes.

FIG. 6 is a diagram for explaining the embodiment of the present invention, and is a diagram illustrating a schema of a meta model management unit to which a tag link is added inherited from a schema (framework) of a document structure management unit. .

FIG. 7 is a diagram for explaining the embodiment of the present invention, and is a diagram illustrating a schema of a document component management unit in which a link component is added and inherited from a schema (framework) of the document structure management unit. .

FIG. 8 is a diagram for describing the embodiment of the present invention, and is a diagram illustrating an example of a meta model representing a relationship between an external file and an index.

FIG. 9 is a diagram for describing the embodiment of the present invention, and is a diagram illustrating an example of attaching a tag to a document instance;

FIG. 10 is a diagram for explaining an embodiment of the present invention;
It is a figure showing an example of a DTD tree structure.

FIG. 11 is a diagram for explaining an embodiment of the present invention;
FIG. 9 is a diagram illustrating an example of a document instance.

FIG. 12 is a diagram for explaining an embodiment of the present invention;
FIG. 5 is a diagram illustrating a meta model between structured documents related to an electric commerce order processing.

FIG. 13 is a diagram for explaining an embodiment of the present invention;
FIG. 12 is a diagram illustrating a part of an example in which a tag is attached to the document instance in FIG. 11.

FIG. 14 is a diagram for explaining an embodiment of the present invention;
FIG. 9 is a diagram illustrating a list of tagged document parts.

FIG. 15 is a diagram for explaining an embodiment of the present invention;
FIG. 11 is a diagram illustrating an example of an instance-level link relationship of a tag according to a meta model.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Document structure management part 2 Parts link management part 3 Object-oriented database 11 DTB management part 12 Document element management part 21 Meta model management part 22 Tag management part 23 Document part management part 31 Tag 32 Tag attribute 111 DTD determination processing 112 Schema definition Process 121 Document instance creation process 122 DB registration 211 Metamodel definition

──────────────────────────────────────────────────続 き Continued on the front page (51) Int.Cl. ⁶ Identification code FIG06F 15/40 380E 15/419 320

Claims (5)

[Claims] 1. An SGML (Standard Generator)
DTD (Document Type Define) of allied Markup Language
ION) based on a predetermined schema, and a DTD management means for managing the definition of the DTD based on a predetermined schema, and a document component attached with a tag defined by the DTD based on a predetermined schema. And a document element management means for managing the document element. 2. A link model is managed as a schema in order to link and manage another structured document with the granularity of words, sentences, and parts included in a structured document based on the SGML DTD. Meta-model managing means, tag managing means for defining and managing tags for newly linking according to the schema, document parts to which the tags are attached and their link relationships, and the document parts in the structured document And a document component management means for managing an appearance position of the document repository device. 3. An SGML (Standard Generator)
DTD (Document Type Define) of allied Markup Language
and a DTD management means for managing the definition of the DTD based on a predetermined schema, and a document component attached with a tag defined by the DTD in a predetermined schema. A document structure management unit comprising: a document element management unit that manages documents based on a document structure; a document structure management unit including: a document structure management unit including: a structured document based on the SGML DTD; In order to manage, a meta model managing means for managing the model of the link relation as a schema, a tag managing means for defining and managing a tag for newly establishing a link according to the schema, and the tag is attached. A document part management unit that manages a document part and its link relationship, and an appearance position of the document part in a structured document; By using the object-oriented database that manages the schema of document parts and manages them as instances, it is possible to search for navigation between document parts, manage consistency of deletion, addition, and updating of document parts, and perform batch editing of document parts. A document repository device characterized by being made possible. 4. An index is assigned to an external file in a binary format outside the structured document that is not a text format, and the relationship between the contents of the index and the external file in a binary format indicated by the index is recorded. A document repository device characterized in that management can be performed in the same framework as the document component management method described in any one of Items 1 to 3. 5. The DTD management means manages a schema definition of tags and tag attributes, and individual tags and tag attributes which are instances thereof, using an object-oriented database. The document component and an instance of the document component attribute are generated based on a tag attribute, a document component, and a schema of a value of the tag attribute of the document component, and are managed using the object-oriented database. 4. The document repository device according to 1 or 3.