JPH08171565A - Document structure converting device - Google Patents

Abstract

PURPOSE: To provide the device which can convert a document structure without a change in the appearance order of elements according to a conversion rule for the document structure that is written by a writer for a conversion rule of a rule without the need to consider the appearance order of the elements. CONSTITUTION: The document structure converting device, which converts the 1st document structure A of a structured document to the 2nd document structure B, has a conversion table 1 which holds the correspondence relation between elements included in two document classes prescribing the document structure, an element kind search part 4 which receives an element included in the 1st document structure A, searches the conversion table 1, and then determines a corresponding element in the 2nd document structure B, an element generation part 5 which generates the element determined by the element kind search part 4, and an element connection part 6 which connects elements generated by the element generation part 5 according to the order relation of the 1st document structure A to generate the 2nd document structure B.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an apparatus for converting the structure of an electronic document processed by a personal computer, a workstation or the like having a word processor or a document processing function.

[0002]

2. Description of the Related Art Since an electronic document created by a word processor or the like is represented as digital data, it is possible to easily add, delete, change, etc., and improve the document creation efficiency. Further, when creating a document similar to an existing document, a new document can be easily created using an existing electronic document. However, even though they are similar documents, there are differences in style and form between the existing document and the document to be created now. It was necessary to proceed with the work while considering the style, and it was quite troublesome even if it was not like handwriting.

Therefore, it has been proposed to structure the electronic document so that the document creation work can be effectively performed. The structured document will be described below.

Information about elements such as chapters, headings, and paragraphs that make up a document and the relationship between the elements, for example, information about a chapter having headings and paragraphs as a substructure is called a document structure.

ODA (Office Document Arc), which is an international standard that embodies the concept of structured documents
title) and SGML (Standard)
Generalized Markup Langua
Due to the spread of ge), documents are often represented as structured data on a computer.

In addition, Maestro (Ian A. M
acknowledge, “Storage and Retrie
val of Structured Documente
ts ", Information Process
g and Management 1990, Vo
l. 26 No. 26. 2, pp. Database management system (D) for documents such as 197-208)
BMS: Database Management S
Also, the document is stored as having a document structure and the document structure is used for retrieval.

A document having the above document structure is called a structured document. Usually, the document structure is
The restrictions on the types of elements that can appear in the document structure, the types of contents associated with each element, the types of attributes, and the connection relationships between elements are determined. This is called a document class. ODA has a common logical structure, and SGML has a DTD (Document).
t Type Definition), Maestr
In o, Document Type corresponding to the schema
Corresponds to the document class.

In many cases, processing for a structured document is premised on that the document structure conforms to some document class.

For example, the common layout structure expressing the layout rule of the ODA document is defined as a pair with the document class, and the same common layout structure cannot be used for documents having different document classes.

In Maestro, all the stored documents match with any document class, and a search is performed using the element types and attributes defined by the document class. It is difficult to search documents with different document classes under the same search condition.

In order to use the same common layout structure for documents having different document classes or to search for documents having different document classes under the same search condition, the document structure of one document class is converted to the document structure of another document class. There is a need to.

For example, in Japanese Unexamined Patent Publication No. 4-177455, which has been applied for a patent by the applicant of the present application, the document structure appears in the document structure in order to convert the document structure of one document class into the document structure of another document class. A rule that describes the relationship of the substructure of the element to be prepared is prepared, and the document structure is converted by evaluating the rule.

[0013]

According to the rule of the document structure generation method described in the above publication, increase / decrease of elements connected directly below,
A change in the order can be described, and even the logical configuration can be changed by changing the order of appearance of document elements. However, on the other hand, the rule writer must always write in consideration of the order of appearance of elements, which complicates the work.

When the same common layout structure is used for documents with different document classes, or when the purpose of searching documents with different document classes under the same search condition is satisfied, it is not necessary to change the order of appearance of the elements, but rather the elements. If the order of appearance of is changed, the meaning of the contents may change, which may be harmful.

In order to avoid such an inconvenience, in the document structure generation method described in the above publication, rules must be described with care so that the order of appearance of elements does not change, which imposes a burden on the operator. There is a problem that it becomes heavy.

According to the present invention, a device for converting a document structure in which the appearance order of elements does not change by the conversion rule of the document structure described without the consideration of the appearance order of elements by the writer of the rule conversion rules. Is provided.

[0017]

As shown in FIG. 1, a document structure conversion apparatus according to the present invention converts a first document structure A in a structured document into a second document structure B. A device, which receives a conversion table 1 that holds a correspondence relationship between elements included in two document classes that define a document structure and an element included in the first document structure A, and searches the conversion table 1 Then, the element type searching means 4 for determining the corresponding element included in the second document structure B, the element generating means 5 for generating the element determined by the element type searching means 4, and the element generating means. 5 and the element connecting means 6 for connecting the elements generated by the step 5 according to the order relation of the first document structure A to generate the second document structure B.

In the document structure conversion device of the present invention, the second document structure B generated by the element connecting means 6 is
It is characterized by further comprising a document structure inspection means 3 for inspecting whether the second document structure B complies with a document class which restricts the second document structure B.

[0019]

When a document having the first document structure A is input, the element type searching means 4 receives the elements included in the first document structure A, searches the conversion table 1 and outputs the second table.
Then, the corresponding element included in the document structure B is determined. The element generating means 5 generates the element determined by the element type searching means 4, and the element connecting means 6 generates the element generating means 5
The second document structure B is generated by connecting the elements generated by the above-mentioned method according to the order relation of the first document structure A.

The information in the conversion table 1 has no information about the appearance order of the elements of the document.
When the user describes the contents of, it is not necessary to consider the order of appearance of elements.

When the elements of the second document structure B are connected, the appearance order of the elements determined in the first document structure is stored and connected, so that the appearance order of the elements may change due to conversion. Absent.

The document structure inspecting means 3 can inspect whether the converted document structure is a document structure that matches the document class that constrains the second document structure B, and inspect the correctness of the conversion result. it can.

[0023]

1 is a diagram showing the outline of a document structure conversion device to which the present invention is applied.

In this embodiment, a document structure restricted by a certain document class is converted into a document structure restricted by another document class. The document structure used in this embodiment is a tree structure. In the figure, A shows the document structure before conversion, and B shows the document structure after conversion.

The document structure conversion apparatus of this embodiment comprises the following three parts. That is, according to the conversion table 1 that stores rules for determining what kind of element each element of the input document structure becomes after conversion, and a new document from the input document structure according to the rules stored in the conversion table 1. A document structure generation unit 2 that generates a structure and a document structure inspection unit 3 that inspects whether the generated document structure complies with a document class that restricts the document structure.

Table 1 shows an example of the contents of the conversion table 1.

[0027]

[Table 1] The conversion table 1 stores conditions for the elements of the input document structure and what kind of elements the elements matching the conditions become in the converted document structure. The conversion table 1 is composed of a nonvolatile memory or a memory backed up by a battery, and the contents of the conversion table can be rewritten by the user.

The conditions for the elements of the input document structure in the conversion table 1 are pointers to various conditions.
The condition that the pointer points to is a tag that indicates the condition type,
The structure is made up of data indicating condition contents (called condition data).

The condition data differs depending on the condition type, as shown below.

If the condition type is about whether or not the element is a specific element type, the condition data becomes the element type.

If the condition type is for the attribute of the element, the condition data is a condition for the attribute of the element, for example, a pair of attribute name and attribute value.

If the condition type is for whether the content that the element has is of a particular type, such as a string, geometry, table, etc., then the condition data will be the content type.

If the condition type is about whether the character content of the element includes a character string of a specific pattern, the condition data is a character string pattern.

The condition is a connection condition, that is, the condition type is for a parent element, for a child element, for an ancestor element, for a descendant element, and for a sibling element. Then, the condition data is the parent element, child element, ancestor element, descendant element,
It is a pointer to the conditions that apply to sibling elements.

Next, the document structure generator 2 will be described. The document structure generation unit 2 is composed of the following three parts.
That is, by referring to the conversion table 1, the element type searching unit 4 that determines which kind of element the input document structure becomes after conversion, and the element type searching unit 4 from the input document structure element An element generation unit 5 that generates an element of the determined type and an element connection unit 6 that connects the elements generated by the element generation unit 5.

Further, the element type searching unit 4 is composed of the following five parts. That is, the element type condition evaluation unit 7 that evaluates the condition for the type of element of the input document structure, and the connection condition (parent, ancestor, child, descendant,
Connection condition evaluation unit 8 for evaluating (conditions regarding sibling elements) and content type conditions for evaluating conditions regarding the type of contents (characters, geometric figures, rasters, mathematical formulas, tables, etc.) held by the elements of the input document structure An evaluation unit 9, an attribute condition evaluation unit 10 that evaluates a condition about an attribute of an input document structure element, and a content pattern condition evaluation unit that evaluates a condition about a character content pattern of an input document structure element Eleven.

Next, the operation of the document structure generator 2 of this embodiment will be described.

The element connecting unit 6 drives the entire document structure generating unit 2. The element type search unit 4 and the element generation unit 5 are called by the element connection unit 6 as necessary.

The movement of the element connecting portion 6 will be described with reference to the flowchart of FIG.

In the element connecting unit 6, any element of the document structure on the input side is an input node InputNode, and any element of the document structure being generated has already been generated is an output node O.
The process is performed as an output node.

At the beginning of generation of the document structure, the element that is the root of the document structure on the input side is the Input Nod.
Since e and Output Node are undecided, the process of the element connecting unit 6 starts.

The element connection unit 6 first calls the element type search unit 4 to determine the element type after conversion of the Input Node (steps 101 and 102).

If the element type cannot be determined, Input
If Node is the root of the document structure on the input side (step 10
4) The generation of the document structure causes an error. Input N
If ode is not the root of the document structure on the input side, Input
The processing for the child element of the Node is started (steps 108 and 109).

When the element type after conversion can be determined, the element generation unit 5 is called to generate the element of the type (step 1
03).

After creating the new element, Input Nod
e is not the root of the document structure on the input side (step 10
5) Concatenate the generated element to the Output Node as the youngest child (step 106), and make the generated element a new Output Node (step 10).
7). If the Input Node is the root of the document structure on the input side, the concatenation is not performed and the generated element is replaced with the new Out.
Put node (step 107).

Thereafter, if the Input Node does not have a child element (step 108), the processing ends.

When the Input Node has a child element, the eldest son of the Input Node is replaced with a new Inp.
ut Node (Step 109), Input
For Node and Output Node, the same processing as in the flowchart of FIG. 2 is recursively called (step 110).

After the processing, if the Input Node does not have the younger brother element (step 111), the process ends.

When the Input Node has the younger brother element, the younger brother immediately after the Input Node is newly replaced with the new In element.
Put Node (Step 112), Input
For Node and Output Node, the same processing as in the flowchart of FIG. 2 is recursively called (step 110).

Next, the operation of the element type search unit 4 will be described. The element type search unit 4 checks in order from the first rule of the conversion table 1 whether or not the condition of the rule matches the Input Node, and sets the element type of the first matching rule as the element type after conversion. For example, in the conversion table of Table 1, the element type PARAGRAPH is rule 1
~ Only if none of Rule 5 is matched, but Rule 6 is matched.

When evaluating the conditions, the element type condition evaluation unit 7, the connection condition evaluation unit 8, the attribute condition evaluation unit 10, the content type condition evaluation unit 9, and the content pattern condition evaluation unit 11 are selected according to the type of condition. To call and evaluate.

If the condition type is for the parent, the connection condition evaluation unit 8 evaluates the condition indicated by the condition data for the parent node of the InputNode. If the result of evaluating the condition for the parent node is true, then the rule has been matched.

If the condition type is child, In
The condition indicated by the condition data is evaluated for all child nodes of the put Node. If there is a child node for which the result of the evaluation of the condition indicated by the condition data is true, the rule has matched.

If the conditional species is of an ancestor, then I
The condition indicated by the condition data is evaluated for all the ancestor nodes of the nput Node. If there is an ancestor node such that the result of the evaluation of the condition indicated by the condition data is true, the rule has matched.

If the condition type is for offspring, then I
The condition indicated by the condition data is evaluated for all descendant nodes of the nput Node. If there is a descendant node for which the result of the evaluation of the condition indicated by the condition data is true, it means that the rule has matched.

If the condition type is for siblings, then I
The condition indicated by the condition data is evaluated for all sibling nodes of the nput Node. If there is a sibling node for which the result of the evaluation of the condition indicated by the condition data is true, then the rule has matched.

As for the conditions indicated by the condition data, the element type condition evaluation unit 7, the connection condition evaluation unit 8, the attribute condition evaluation unit 10, the content type condition evaluation unit 9, and the content pattern condition are also selected according to the type of the condition. The evaluation unit 11 is called and evaluated. If the condition indicated by the condition data is a connection condition,
The condition indicated by the condition data of the connection condition also includes the element type condition evaluation unit 7, the connection condition evaluation unit 8, the attribute condition evaluation unit 10, the content type condition evaluation unit 9, and the content pattern according to the type of the condition. The condition evaluation unit 11 may be called and evaluated.

Next, the operation of the element generator 5 will be described. The element generation unit 5 generates an element of the type determined by the element type search unit 4. At this time, Input N
The contents of ode are copied and become the contents of the new element. At this time, the attribute value of one of the attributes of the Input Node may be copied to be the attribute value of the attribute of the new element. The conversion table 1 may store the rule of which attribute value has which attribute value.

The document structure inspection unit of this embodiment may be constructed by using an existing technique.

When the document class is defined as SGML DTD, for example, Anne Bruggem
an-Klein, “Regular express
ions into finite automat
a ”, Right Notes in Comp
uter Science, Vol 583, p
p. 87-98, 1992, or Anne B
ruggeman-Klein and Derick
Wood, “Deterministic regu
lar language ", Lecture No
tes in Computer Science,
Vol. 577, pp. By the method shown in 173-184, 1992, an automaton for inspecting the document structure may be created and used for inspecting.

An example of document conversion in this embodiment is shown below.

FIG. 3 shows a document class that restricts the document structure that is an input of this conversion example.

"Root" and "se" surrounded by a rectangle
"ct", "fig", and "par" are the types of elements that appear in the document structure, "root" is the root, and "sec" is the root.
“T” is a chapter, “fig” is a figure, and “par” is a paragraph.
Links connecting element types and OPT, REP, CH
O and SEQ limit the tree structure that the document structure can have. The meaning is "ISO / IS 8613, Inf.
orientation Processing-Tex
t and Office System-Off
ice Document Architecture
(ODA) and Interchange Fo
Since it is the same as that described in rmat, Part 2, Annex B ″, a detailed description thereof will be omitted here, but OPT indicates that a lower structure appears 0 or 1 times, and REP indicates a lower structure. The structure indicates that the structure appears one or more times, CHO indicates that any one of the structures shown immediately below appears, and SEQ indicates that all of the structures shown immediately below appear in the indicated order.

In the document class according to this embodiment, the types of contents that can be possessed and the attributes are defined for each type of element. In the document class of FIG. 3, it is assumed that root and set have no content, par has character content, and fig has geometric figure content. Also par is
It is assumed that an attribute par-type having a character string type attribute value is specified.

FIG. 4 illustrates a document class that restricts the document structure of the conversion result of this conversion example. The meaning of the contents of the figure is the same as that of FIG.

In the document class shown in FIG. 4, ROOT and SEC are used.
TION has no contents, TITLE, PARAGRA
It is assumed that PH has character contents and FIGURE has geometric figure contents.

FIG. 5 illustrates an example of a document structure which is an input of this conversion example.

The document structure may be implemented, for example, as a set of structures representing elements. The structure that represents an element is, for example, a character string indicating the element type, data about the attribute of the element (for example, a set of attribute name and attribute value pairs), information about the content of the element (content type is It should be implemented so that it consists of a character string that represents the content, a pair of data that represents the content, and a string of pointers that point to the child element of the element.

In FIG. 5, each element of the document structure is shown by connecting the element type and a number for identification. Figure 5
In the document structure, the value of the attribute par-type of par1 is a character string "title". Also pa
It is assumed that the character content of r4 includes the character string "Chapter 2".

The result of converting the document structure of FIG. 5 by the present embodiment having the conversion table 1 shown in Table 1 is shown in FIG.
Is the document structure of.

In FIG. 6, each element of the document structure is shown by connecting the element type and a number for identification. Also, what is enclosed by () shows which element of the document structure before conversion corresponds to that element.

The conversion does not generate an element corresponding to sect2 that does not match any of the rules in Table 1.
Therefore, the document structure of FIG. 6 has a smaller depth of hierarchy than the document structure of FIG. However, with respect to the other elements, corresponding elements have been generated, and the generation result is an inter-element grandchild relationship defined by the document structure of FIG. The order of appearance is preserved.

Since the document structure of FIG. 6 satisfies the constraint of the document class of FIG. 4, the inspection of the document structure inspection section normally passes.

[0074]

As described above, in the present invention, since the appearance order of elements is not changed by the conversion, it is possible to prevent the change of the logical structure of the document due to the change of the appearance order of elements. it can. Moreover, since the element concatenation unit guarantees the preservation of the appearance order and is not described in the conversion table itself, the user can describe the contents of the conversion table without considering the preservation of the appearance order.

[Brief description of drawings]

FIG. 1 is a diagram showing an outline of a document structure conversion apparatus to which the present invention is applied.

FIG. 2 is a flowchart for explaining the movement of the element connecting portion.

FIG. 3 is an explanatory diagram illustrating a document class that restricts a document structure before conversion.

FIG. 4 is an explanatory diagram illustrating a document class that restricts a document structure after conversion.

FIG. 5 is an explanatory diagram showing a document structure before conversion.

FIG. 6 is an explanatory diagram showing a document structure after conversion.

[Explanation of symbols]

1 ... conversion table, 2 ... document structure generation unit, 3 ... document structure inspection unit, 4 ... element type search unit, 5 ... element generation unit, 6 ... element connection unit, 7 ... element type condition evaluation unit, 8 ... connection condition Strengthening department,
9 ... Content type condition Condition evaluation unit, 10 ... Attribute condition evaluation unit, 1
1 ... Content pattern condition evaluation unit

Claims (2)

[Claims] 1. A document structure conversion device for converting a first document structure in a structured document into a second document structure, the correspondence relationship between elements included in two document classes that define the document structure. A conversion table that holds the element, and an element type searching unit that receives an element included in the first document structure, searches the conversion table, and determines a corresponding element included in the second document structure, An element generation unit that generates the element determined by the element type search unit, and an element that generates the second document structure by connecting the elements generated by the element generation unit according to the order relation of the first document structure. A document structure conversion device comprising: a connecting unit. 2. The second generated by the element connecting means
The document structure conversion device further comprising a document structure inspection means for inspecting whether or not the document structure of is according to a document class that restricts the second document structure.