JP2007310617A - Structured document processing system, structured document processing method and program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To perform processing of evaluating a search expression at high speed by using differential information given, in information extraction processing by search expression evaluation from a structured document. <P>SOLUTION: A search expression 111 registered beforehand is classified into a simple structural expression, an expression with predicate, a complex expression, and the like and registered to a search expression table 108, the search expression is applied to an original document 112 registered beforehand, and its application result and calculation process information are calculated, and registered to an original document table 109. When the differential information 113 is input, it is analyzed by a differential analysis part 104 and classified into either a value difference or structure difference, and in accordance with combination of the kind of differential information 113 and the kind of search expression, the search expression is evaluated by using the differential information 113 and the calculation process information by a differential search expression evaluation part 105 when differential calculation is possible; otherwise, a structured document is restored by a full-text search expression evaluation part 106 and the search expression is evaluated. <P>COPYRIGHT: (C)2008,JPO&INPIT

Description

  The present invention mainly relates to a search process for a structured document, and in particular, evaluates a search expression for a structured document expressed by difference information from another structured document, and evaluates a difference between a search expression evaluation result for another structured document and a difference. The present invention relates to a structured document processing system, a structured document processing method, and a structured document processing program that are performed based on information.

(XML and XPath)
Structured documents such as XML and SGML are used for various purposes such as communication standards because of their advantages such as flexibility and extensibility. As a communication standard that employs a structured document as a standard, SOAP (Simple Object Access Protocol) and the like are prominent.

  A structured document constitutes a tree structure with one element as a root. A tree trunk or a leaf portion is called a node, and each node is generally composed of an element name, an element value, an attribute name, an attribute value, a child element, and the like. The relationship between the connected nodes is as follows: “parent” and “ancestor” on the side closer to the root, “child” and “descendant” on the side closer to the leaf, and “sibling” on the child node with the same parent. Expressed. Among nodes, a node that indicates a value such as an element value or an attribute value and has no descendant is expressed as a “value node” in this specification.

  XPath (XML Path Language) is used as a search expression for extracting such a structured document, particularly a specific element in XML. XPath is standardized by the standardization organization W3C (WWW Consortium), and its specifications are described in Non-Patent Document 1.

  In XPath, a search expression is expressed by listing XML elements separated by “/” to indicate the structural relationship between the elements, thereby specifying a location that satisfies the relationship in the XML document. Commonly used expressions describe a path that goes from the root of the document to the child, such as "/ A / B / C". The corresponding "C" element and its descendants in the structured document Is the result of the search expression. There can be a plurality of corresponding locations in one document with one search expression.

  It is also possible to impose additional conditions on the structural relationship with expressions called predicates. This is expressed by enclosing "[]" after the element, and further describing the search expression and the condition for the result in "[]". Common uses of predicates include the description of conditions for the contents of value nodes near the element such as "/ A / B [C = 1]" and siblings such as "/ A / B [2]" For example, the order of the node group.

  Furthermore, in XPath, it is possible to describe very complex search conditions by using functions other than parent-child relationships, function processing for specified locations in certain search expressions, and various combinations thereof.

(Existing XPath search technology)
However, this XPath search is a heavy processing, and a technique for processing this at a high speed by devising the method has been proposed. For example, in Patent Document 1, the XPath to be searched is registered in advance and converted into an automaton, and an XPath search is performed at high speed by driving the automaton while sequentially analyzing an XML document using a SAX parser. Disclosure.

  On the other hand, it is not an XPath search, but in a system that handles many similar structured documents, for example, a communication message processing system between the same system, if one document is processed, the result is retained and similar When processing other documents, several techniques have been proposed for speeding up the processing by obtaining a difference from the previous document and adapting this difference to the processing result of the previous document.

  For example, in Patent Document 2 and Patent Document 3, such a method is used in order to speed up rendering of an HTML document as structured document processing. FIG. 23 shows an example of the prior art of a structured document processing apparatus using difference information. A conventional structured document processing apparatus using difference information includes an input unit 2301, a difference calculation unit 2302, a difference reflection unit 2303, an output unit 2304, an original document storage unit 2305, and an original document processing result storage unit 2306. It is composed of

  The conventional structured document processing apparatus using the difference information having such a configuration operates as follows.

  That is, the original document is registered in the original document storage unit 2305 in advance, and the rendering result that is the processing result is registered in the original document processing result storage unit 2306. Here, when a new document to be actually processed is given to the input unit 2301, first, the difference calculation unit 2302 calculates the difference between the original document and the new document, and the difference reflection unit 2303 stores the difference in the original document processing result storage unit 2306. The processing result of the original document is determined from the difference information and the processing result of the original document, the point to be corrected of the processing result of the original document is determined, only the corresponding part is corrected, and the processing result of the original document is re-executed for the other parts. A processing result of the used new document is created and output by the output unit 2304.

  In Patent Document 4, a similar technique is used for XML document analysis processing.

  In these techniques, when the change of the original document is instructed by the difference information, the correspondence between the difference information and the processing result is always acquired in order to reflect the change in the processing result. That is, it is not assumed that the correspondence between the difference information and the processing result cannot be acquired.

  For example, in Patent Document 2 and Patent Document 3, the rendering result, which is the processing result, forms a tree structure based on the nesting relationship of the screen parts, and this corresponds to the tree structure of the original structured document as it is. When a certain part of the structure is changed, it specifies which screen part of the corresponding rendering result is to be corrected.

  On the other hand, in Patent Document 4, a structured document as a binary string is given as an input, and difference information is also a difference of the binary string. In order to reflect this in the analyzed document data that is the processing result, the state transition machine The correspondence is obtained using.

JP 2003-323429 A JP 2005-108101 A JP 2002-091730 A JP 2006-024179 A “XML Path Language (XPath)“ XML Path Language (XPath) ””, [online], [searched on December 22, 2004], Internet <URL: http://www.w3.org/TR/ xpath>

  In the speed-up method as in Patent Document 1, although the method itself is relatively fast, the difference information cannot be used effectively, so that the entire document needs to be processed every time for processing. As a result, even when a large number of similar documents have to be handled, the entire document must always be searched and the speed-up effect is limited.

  On the other hand, conventional techniques such as Patent Documents 2, 3 and 4 disclose a method for speeding up rendering and document analysis processing using difference information and previous processing results, but information extraction by search expression evaluation such as XPath is disclosed. There are no examples of effective use of difference information for processing. The reason is that in the information extraction process based on the evaluation of the search expression such as XPath, the relationship between the original document structure and the processing result is extremely diverse, and the difference information and the processing result are different from the conventional techniques such as Patent Documents 2, 3, and 4. This is because it is not always possible to accurately obtain the corresponding relationship.

  For this reason, in the information extraction process based on the retrieval formula evaluation from a structured document in a system that handles many similar structured documents, when the difference information from the original document that is a known structured document is given, the difference Development of a new method that can extract information at high speed from the original document and the information extraction processing result for the original document using information is eagerly desired.

[Object of invention]
The present invention has been proposed in view of such circumstances, and its object is to provide a novel structured document processing system and method for effectively using difference information for information extraction processing based on search expression evaluation such as XPath. Is to provide.

  The present inventor has found that the reason why the relationship between the original document structure and the processing result is extremely diverse in the information extraction process by the evaluation of the search expression such as XPath is in both the search expression and the difference information. The present invention has been completed by finding that a case in which a search formula can be evaluated based on difference information can be specified by a combination of the search formula and the type of difference information if classified according to the nature. The document processing system includes a search expression management unit that classifies and holds search expressions based on their properties, and an original document management that stores the original document together with a result of applying the original document to the held search expression and processing process information at the time of application. A difference analysis unit that classifies difference information of a structured document based on its properties, determines a search expression evaluation method based on a difference type and a search expression type, and a search expression based on difference information A difference search expression evaluation unit that evaluates the search expression using the difference information, the search expression and its classification result, the original document, and the evaluation result of the search expression for the original document when the price is possible, and the difference information A full-text search expression evaluation unit that generates a new document in which difference information is applied to the original document and applies the search expression to the original document, a differential search expression evaluation unit, and a full-text search expression evaluation unit And an output unit that outputs the evaluation result of the retrieval formula for the document in which the difference information is applied to the original document.

  According to a second structured document processing system of the present invention, in the first structured document processing system, the search expression management unit includes a search expression when the search expression includes only a parent-child relationship of nodes of the structured document. It is classified as a simple structural formula, and when the search formula is a simple structural formula with a predicate including the simple structural formula added, it is classified as a predicated expression, and the other formula is classified as a complex formula, The difference analysis unit classifies the difference information according to whether the change part is a value or a document structure, and classifies the difference information as a value difference when the change part is only a value, and when the difference information is a value difference and a search formula Is a simple structural formula or a predicate-added formula, it is judged that the retrieval formula evaluation based on the difference information is possible.

  According to a third structured document processing system of the present invention, in the second structured document processing system, the original document management unit is a result of the processing process information being a list of nodes included in the application result of the search expression. The node set is stored as a conditional node set that is a list of nodes directly evaluated by the search expression in order to derive the application result of the search expression.

  According to a fourth structured document processing system of the present invention, in the third structured document processing system, the original document management unit further includes a predicate in the search expression as the processing step information, and an expression other than the predicate Is a candidate node set that is a list of nodes that are applicable and determined not to be applicable according to the evaluation result of the predicate.

  According to a fifth structured document processing system of the present invention, in the fourth structured document processing system, the original document management unit stores the original document by assigning a unique ID to each node and processing process information nodes The set is designated by the ID, and the difference analysis unit designates the change location of the difference information by the ID.

  According to a sixth structured document processing system of the present invention, in the fourth or fifth structured document processing system, the difference search expression evaluation unit is configured such that the difference is the value difference and the search expression is the simple structure expression. If there is a change to the output of the processing result for the original document by the result node set, if there is no change, the processing result is directly used as the evaluation result of the search expression. And applying the difference to the processing result of the original document is used as the evaluation result.

  According to a seventh structured document processing system of the present invention, in the sixth structured document processing system, the difference search expression evaluation unit further includes the difference as the value difference and the search expression as a pre-descriptive word-added expression. If it is, it is checked whether there is a possibility that an output result of the search expression evaluation for the original document may be invalidated by a change due to the difference by the condition node set. Re-evaluate the search formula for the part by applying only the relevant differences, and it is determined that there is no possibility in the search of the condition node set, or the search formula is the result of the re-evaluation of the search formula In this case, it is checked whether or not there is a change in the output of the processing result for the original document by the result node set. If there is no change, the processing result is directly used as the evaluation result of the search expression. Applies the result and applies the difference to the processing result of the original document as the evaluation result, and at the same time, the node determined by the candidate node set that the change due to the difference does not correspond to the search expression evaluation for the original document If there is a possibility, only the difference related to the candidate node set is applied and the search expression is reevaluated for that part. When it is determined that it corresponds, it is checked whether there is a change to the output result of the corresponding part. If there is no change, the output result is directly used as the evaluation result of the search expression. It is characterized by having an evaluation result.

  According to an eighth structured document processing system of the present invention, in any one of the first to seventh structured document processing systems, a difference calculation unit that compares an input structured document with an original document and calculates a difference; An input document storage unit that stores the input structured document, and the difference search expression evaluation unit and the full-text search expression evaluation unit store the difference in the original document instead of applying the difference to the original document. The input document is referred to.

  According to a ninth structured document processing system of the present invention, in any one of the first to seventh structured document processing systems, a new document that manages a new document in which a part or all of differences are applied to an original document. The difference search expression evaluation unit and the full-text search expression evaluation unit request the new document management unit to apply and obtain the result when applying the difference to the original document, and the new document management unit The difference applied part is not subjected to difference application again, and the application result is returned from the stored new document.

  A tenth structured document processing system according to the present invention provides one or more search expressions and their types, an original document, a storage unit for storing a result of applying the search expression to the original document, and processing process information, The difference information of the structured document is classified based on its property, and the difference information is used for each search formula according to the type of the difference information and the type of the search formula stored in the storage unit. A difference analysis unit that determines whether or not to perform the evaluation, and a result and processing of applying the search formula to the original document stored in the storage unit for the search formula determined to be evaluated using the difference information A difference search expression evaluation unit that performs evaluation using process information, and a search expression that is determined to be evaluated without using difference information, and a document in which the difference information is applied to the original document stored in the storage unit. Full-text search expression evaluation unit Characterized in that it comprises a.

  According to an eleventh structured document processing system of the present invention, in the tenth structured document processing system, the storage unit includes a simple structural formula search formula including only a parent-child relationship of nodes of the structured document, and the search The result of applying the formula to the original document and the process information are stored, and the difference analysis unit, when the difference information of the given structured document is a value difference whose change part is only a value, The retrieval formula for the simple structural formula is determined to be evaluated using the difference information, and the differential retrieval formula evaluation unit stores the simple structure in the original document stored in the storage unit for the retrieval formula for the simple structural formula. It is characterized in that evaluation is performed using a result of applying an expression search expression and processing process information.

  According to a twelfth structured document processing system of the present invention, in the tenth structured document processing system, a predicate including a simple structural formula in a simple structural formula including only a parent-child relationship of nodes of the structured document in the storage unit. Is stored, a result of applying the search expression to the original document, and processing process information, and the difference analysis unit changes the difference information of the given structured document. If the location is a value difference that is only a value, it is determined that the search expression with the predescript word-added expression is evaluated using the difference information, and the difference search expression evaluation unit determines the search expression with the pre-description word add-on expression. Is evaluated using a result obtained by applying a search expression with a pre-description word to the original document stored in the storage unit and processing process information.

  A first structured document processing method of the present invention is a method for evaluating a search expression for a structured document using a computer, wherein the computer adds one or more search expressions to the storage unit and adds the types thereof. A first step of adding the result of applying the search formula to the original document and processing process information to the original document and storing the original document in the storage unit; Whether to classify information based on its properties, and to perform evaluation using difference information for each of the search formulas according to the type of difference information and the type of search formula stored in the storage unit A second step of determining the search result, and a result of applying the search formula to the original document stored in the storage unit and processing process information for the search formula determined to be evaluated by the computer using the difference information; The And a third step of evaluating the search expression determined to be evaluated without using the difference information using a document in which the difference information is applied to the original document stored in the storage unit. It is characterized by including.

  A second structured document processing method of the present invention is a method for evaluating a search expression for a structured document using a computer, wherein the computer searches for a simple structure expression including only parent-child relationships of nodes of the structured document. A first step of adding an expression, a result of applying the search expression to the original document, and processing process information to the storage unit, and storing the difference in the given structured document If the information is classified based on its property and the changed part is a value difference whose value is only a value, the second step of determining to evaluate using the difference information for the search formula of the simple structural formula, A third step in which the computer evaluates the simple structural formula retrieval formula using the result of applying the simple structural formula retrieval formula to the original document stored in the storage unit and the process information; And wherein the Mukoto.

  According to a third structured document processing method of the present invention, in a method for evaluating a search expression for a structured document using a computer, the computer simply converts to a simple structure expression including only a parent-child relationship of nodes of the structured document. A predicate-added search expression to which a predicate including a structural formula is added, a first step of adding the result of applying the search formula to the original document and processing process information and storing the result in the storage unit; When the computer classifies the difference information of the given structured document based on its property and the change part is a value difference whose value is only the value, the difference information is used for the search expression with the pre-descriptive expression A second step of determining that the evaluation is performed, and a result of the computer applying the search expression with the predescription word to the original document stored in the storage unit for the search expression with the predescription word; With process information Characterized in that it comprises a third step of evaluation are.

"Action"
In the structured document processing system of the present invention, a search expression to be evaluated and an original document to be subjected to difference are registered in advance. Here, the search expression management unit holds the registered search expressions, and at the same time classifies the search expressions based on their properties, and the original document management unit registers the registered original document and the original document. The result of applying the retrieval formula and the processing process information are both held.

  Next, when the difference information between a new document and the original document is given, the difference analysis unit analyzes the difference information of the given structured document, and performs each search based on the difference type and the search expression type. The retrieval formula evaluation method is determined for each formula. When the search formula evaluation using the difference information is possible, the differential search formula evaluation unit searches using the difference information, the search formula and its classification result, the original document, and the evaluation result of the search formula for the original document. When the evaluation of the expression is performed and the search expression cannot be evaluated using the difference information, the full-text search expression evaluation unit generates a new document in which the difference information is applied to the original document, applies the search expression to the original document, and outputs the output unit. However, the results of the difference search expression evaluation unit and the full text search expression evaluation unit are combined and output as the evaluation result of the search expression for the document in which the difference information is applied to the original document.

  In this way, the search expression and the difference information are classified according to their properties, and an appropriate search expression evaluation method is determined from the combination of the search expression and the difference information type. In particular, the difference information can be applied to the search expression evaluation result of the original document. In the case of the combination, by performing the search expression evaluation using the difference information and the search expression evaluation result of the original document, the application of all the difference information to the original document and the evaluation of the entire document of all the search expressions can be omitted. The object of the present invention can be achieved.

  According to the present invention, when the search formula and the difference information of the structured document are given, depending on the nature of the search formula and the difference information, the original document to be subjected to the difference and the information extraction processing result for the original document are used. Thus, the information extraction process can be performed at high speed.

  The reason for this is to analyze the properties of the search expression and difference information, use an appropriate search expression evaluation method for each property, especially when the difference information is applicable to the search expression evaluation results of the original document. This is because the evaluation of the entire search expression document can be omitted.

  Next, the best mode for carrying out the present invention will be described in detail with reference to the drawings.

  Referring to FIG. 1, the structured document processing apparatus 101 according to the first embodiment of the present invention includes a search expression management unit 102, an original document management unit 103, a difference analysis unit 104, and a difference search expression evaluation unit 105. And a full-text search expression evaluation 106 and an output unit 107.

  The search formula management unit 102 has a classification unit 102a and a storage unit 102b therein, and the storage unit 102b has a search formula table 108. FIG. 2 is a configuration diagram of the search expression table 108. The search expression table 108 has a search expression and a search expression type for each search expression.

  The original document management unit 103 includes a search expression evaluation unit 103a and a storage unit 103b, and the storage unit 103b includes an original document table 109. FIG. 3 is a configuration diagram of the original document table 109. The original document table 109 has an original document, a search expression, a processing result, a processing result node set, a condition node set, and a candidate node set. Each original document can have multiple search expressions. Each search expression can have multiple processing results. Each processing result can have one result node set and one condition node set. Each search expression can have one candidate node set. Here, in the present embodiment, corresponding to a certain original document and search expression in the original document table 109, a node group appearing in the application result is separated from the result of applying the search expression to the original document. The reason why the ID set is stored as a result node set is to make it possible to easily check whether there is an influence of a difference by comparing IDs. As another embodiment, an embodiment in which the result node set is omitted and the result node set is dynamically generated every time from the application result is also conceivable.

  The difference analysis unit 104 includes a classification unit 104a and an evaluation unit determination unit 104b.

  Each of these means generally operates as follows.

  The search formula management unit 102 analyzes and classifies the search formula 111 registered from outside by the classification unit 102a, and stores the result together with the search formula 111 in the search formula table 108 through the storage unit 102b. Here, the search formula 111 is classified based on the ease of difference calculation.

  The original document management unit 103 applies the search expression held in the search expression table 108 by the search expression management unit 102 to the original document 112 registered from the outside, and the original document 112, the application result of the search expression, Information on the application process is stored in the original document table 109 through the storage unit 103b.

  The difference analysis unit 104 analyzes and classifies the input difference information 113 by the classification unit 104a, and for each search formula registered in the search formula management unit 102 by the evaluation unit determination unit 104b, the difference information classification result and the search formula Whether or not a search using difference information is possible based on the classification result, and let the differential search formula evaluation unit 105 evaluate the search formula that can be used and the full-text search formula evaluation unit 106 to evaluate the search formula that cannot be used. .

  The difference retrieval formula evaluation unit 105 determines a method of using the difference information from the difference information 113 and the type of the retrieval formula, the change target of the difference information 113, the processing result stored in the original document management unit 103, and the processing process information. The search expression is evaluated using the difference information.

  The full-text search expression evaluation unit 106 creates a new document in which the difference information 113 is applied to the original document, and evaluates the search expression for the new document.

  The output unit 107 outputs the result of the difference search expression evaluation unit 105 and the result of the full text search expression evaluation unit 106 together.

  FIG. 4 is a flowchart showing the operation of the search expression management unit 102. The search expression management unit 102 receives an external registration request for the search expression 111 in S401, examines the description contents included in the search expression 111 in S402, classifies the search expression 111, and classifies the search expression 111 in S403. The result is added to the search expression table 108, and the original document management unit 103 is notified that the expression is added in S404. If the search formula 111 includes only the simple parent-to-child reference relationship of the element that is the structure information of the structured document, the classification is described by only the simple reference formula and the simple reference formula. If there is a predicate that is a condition for the value to be processed, it is classified as a predicate expression, and other expressions are classified as complex expressions.

  FIG. 5 is a flowchart showing the operation of the original document management unit 103. In step S501, the original document management unit 103 receives an external request. If the original document 112 is registered in step S502, the original document management unit 103 extracts the target original document 112 from the registration request in step S503. The original document 112 is stored in the original document table in step S504. 109, and a list of search formulas is acquired from the search formula management unit 102 in S505, the processing results and processing steps of the search formulas are calculated for each search formula in S506, and the processing results and processing step information are calculated in S507. Save.

  Here, when storing the original document, an ID unique to the document is assigned to each node of the original document 112, and the original document 112 is stored together with this ID. In addition, the processing result of the search expression and the calculation of the processing process are specifically, the result applied to the original document 112 is stored as the processing result in the original document table 109, and the set of node group IDs appearing in the application result is processed. As a result node set, the set of node group IDs directly evaluated by the search expression to derive the application result is stored as a condition node set of the process, and if the search expression is a predicate expression, Candidate node set is the set of node group IDs corresponding to the conditional node set when a node that is applicable only if it is a simple structure excluding predicates but is no longer applicable by the predicate part is regarded as a temporary application result. Save as.

  If the request is a request for adding a new search formula from the search formula management unit 102 in S502, the target search formula is extracted from the search formula addition request in S511, and all the registered search formulas are registered in S512. The original document is taken out, the application result of the search formula and the processing process are calculated for it in S513, and added to the original document table 109 in S514.

  FIG. 6A is a flowchart showing the operation of the difference analysis unit 104. The difference analysis unit 104 designates reference to the original document stored in the original document management unit 103 in S601 and each node of the original document by the ID given by the original document management unit 103, and instructs to change the node. When the expressed difference information 113 is received, a list of change destination nodes described in the difference information 113 is extracted in S602, and whether a node other than a value node such as an element value or an attribute value is included in the change destination node in S603. The difference is classified by checking whether or not. In the difference classification, a difference in which a node other than the value node is not included in the change destination node is a value difference, and a difference in which at least one node other than the value node is included is a structure difference.

  Next, in S604, a list of search expressions and their types registered from the search expression management unit 102 is acquired, and in S605 and S606, difference information is used for each search expression from the difference type and the search expression type. If the difference information is a value difference in S607 and the search expression is a simple structural expression or a predicate expression in S608, it is determined that the difference information can be processed. In S610, the search formula is assigned to the differential search formula evaluation unit 105. In S607 and S608, if the difference information is a structural difference or the search formula is a complex formula, the search formula is determined to be impossible. The search expression is evaluated by passing the search expression, difference information, and type of difference information to the search expression evaluation sections 105 and 106 assigned to the evaluation section 106 and assigned to each search expression in S611.

  That is, in the present embodiment, as shown in FIG. 6B, the difference information is classified into two types, that is, a value difference and a structure difference, depending on its property, and the search expression is expressed by a simple structure formula and a predicate-added expression based on its property. If the combination is classified into three types with other complex expressions, the difference information type is a value difference, and the search expression type is a simple structural expression and a predicate expression, a search based on the difference information is performed. If the combination is marked with an X, a full text search is performed.

  FIG. 7 is a flowchart showing the operation of the difference search expression evaluation unit 105. When the difference search expression evaluation unit 105 receives the set of search expressions, the difference information, and the type of the difference information, the difference search expression evaluation unit 105 acquires the process information and the processing result of the corresponding search expression from the original document management unit 103 for each search expression. The search information is evaluated using the difference information as follows, and the result is sent to the output unit 107.

  (1) When the retrieval formula is a simple structural formula, the difference is applied to the processing result of the original document, and this is used as the result. Specifically, in S710, it is checked whether even one change target node ID is included in the result node set. If it is not included, the processing result of the original document is directly used as a result in S712. A result obtained by rewriting the portion corresponding to the processing result of the document with the difference information is the result.

  (2) If the search expression is a predicate expression, both of the following processes are performed.

  (2.1) In S720, it is checked whether even one change target node ID is included in the condition node set. If even one change target node ID is included in the condition node set, the original document indicated by the condition node set in S721 Is rewritten with the difference information, and the search expression is evaluated again only for this part in S722. If it is not included in the conditional node set in S720, or if the search expression is satisfied again in S722, the difference is applied to the processing result of the original document in S723, S724, and S725, and this is the result.

  (2.2) In S731, it is checked whether even one change target node ID is included in the candidate node set. If it is included, in S732, a part of the original document indicated by the candidate node set is rewritten with the difference information, and in S733 If the search expression is evaluated again only for this part, and the search expression is newly applicable as a result, it is checked whether there is a difference with respect to the result in S734. If there is, the difference is applied again to the result in S735, and this is used as the result.

  FIG. 8A is a flowchart showing the operation of the full-text search expression evaluation unit 106. Upon receiving the search expression set and difference information 113 in S801, the full-text search expression evaluation unit 106 extracts the original document from the original document management unit 103 in S802, and generates a new document by applying the difference information to the original document in S803. In step S804, the search expression designated for the new document is evaluated, and in step S805, the result is output to the output unit.

  The output unit 107 outputs the evaluation result for each search expression.

(Algorithm explanation)
Here, a supplementary description will be given of the relationship among the types of search expressions, the types of differences, the process information, and the operation of the difference search expression evaluation unit 105. The impact of document changes due to differences on the processing results of the original document can be classified into the following four categories.

(1) It is a change of a part unrelated to the evaluation of the search expression.
(2) As an influence on the part corresponding to the search expression in the original document,
(2-1) The corresponding part in the original document is no longer applicable.
(2-2) The corresponding part in the original document continues to be applicable, but the corresponding part to be output is changed.
(3) The part that was not applicable in the original document becomes applicable.

  However, the influence of one change on the result of one search expression may be (2) and (3) at the same time (S734 to S736 are provided in consideration of this). Cases that do not fall under either (2) or (3) are (1).

  Since the contents of the retrieval formula and the difference information are extremely diverse and it is difficult to clarify the general relationship, in the present invention, the difference and the retrieval formula that clearly indicate which of the above the influence of the difference corresponds to. In the case of the above combination, calculation using the difference information is performed, and in other cases, that is, when the influence of the difference on the search result with respect to the original document is unknown, normal search expression evaluation in the conventional technique is performed.

  In order to check whether or not it falls under (2-1), it is first checked whether the node that is the condition for deriving the location corresponding to the search expression in the original document has been changed. Is the condition node set. However, in the case of simple structural formulas, predicates that restrict values are not included, so the change in value does not affect the result of the search formula.As a result, (2-1) There is no such thing. If the predicate-added expression falls under (2-1), the result in the original document may not be applicable in the new document, so the search expression is re-evaluated only at that location (S721, S722).

  Next, when (2-1) is not satisfied, it is checked whether or not it corresponds to (2-2). Processing result information for performing this is a result node set. In this case, it is simply checked whether or not the changed part is included in the node included in the output. If applicable, only the result in the original document is rewritten with the difference and output (S711, S724).

  On the other hand, in order to check whether or not it corresponds to (3), it is necessary to select a part that may be applicable by the difference from the parts that were not applicable in the original document search and re-evaluate only that part. There is. However, in the case of simple structural formulas, predicates that restrict values are not included, so the change in value does not affect the result of the search formula.As a result, (2-1) There is no such thing. However, as described above, in the case of a simple structural formula, since the predicate which is a restriction on the value is not included, the corresponding part does not change by changing the value, and therefore the simple structural formula does not become (3). In addition, in the predicate expression, the part corresponding to the simple structural expression excluding the predicate part does not change when the value is changed, and the part that was not applicable in the original document may be changed by the evaluation of the predicate. Only. The process information used to check this possibility is a candidate node set (S731, S733).

  FIG. 8B shows a summary of the contents of the above search algorithm in a table format.

  Next, the effect of this embodiment will be described.

  In this embodiment, the search formulas registered in advance are classified, and the result of applying the search formula to the original document registered in advance and the information on the processing process are held, and the input difference information The search method evaluation method that can be used is determined from these classification results and processing process information, and when the labor saving of the calculation is possible using the difference information, the difference is obtained by using the labor-saving method. The retrieval formula of the structured document can be evaluated at high speed using the information.

  Further, in this embodiment, the value difference and the simple structural formula are further classified by classifying the search formula into a simple structural formula, a predicate-added formula, and other complex formulas, and classifying the difference information into a value difference and a structural difference. Alternatively, in the case of a combination of predicates, the search result can be obtained using the processing result of the search expression for the original document, the processing process information, and the difference information.

(Second embodiment)
Next, a second embodiment of the present invention will be described in detail with reference to the drawings.

  Referring to FIG. 9, in the second embodiment of the present invention, in addition to the configuration of the first embodiment, a difference calculation unit 901 that calculates difference information between the input structured document 114 and the original document 112. And an input document storage unit 902 that holds the input document.

  Each of these means generally operates as follows.

  Unlike the first embodiment, the structured document processing apparatus 101 according to the present embodiment receives the structured document 114 instead of the difference information, and the difference calculation unit 901 registers the input structured document 114 in advance. Calculates the difference from the original document and outputs it to the difference analysis unit 104 as difference information, and saves the input document together with the correspondence information with the input document node considered to match the original document at the time of the difference calculation Stored in the part 902. Correspondence relationship information can be realized, for example, by assigning the same ID as the node of the corresponding original document to the node of the input document.

  The full-text search expression evaluation unit 106 takes out a document stored in the input document storage unit 902 instead of the process of generating a new document from the difference information and the original document in the first embodiment, and uses this.

  Also, the difference search expression evaluation unit 105 is a process in the case where the search expression in the first embodiment is a predicate expression and the difference is a value difference, and a part of the original document indicated by the condition node set or the candidate node set is determined based on the difference information. Instead of rewriting, using the correspondence information between the input document stored in the input document storage unit 902 and the node of the original document, a part of the input document corresponding to a part of the original document to be rewritten Decide and evaluate the search formula only there.

  According to the present embodiment, even when the input to the structured document processing apparatus 101 is not a difference information but a structured document, if the load of difference calculation is sufficiently smaller than the load of application of the entire search expression document, the search It is possible to speed up the processing of expressions.

(Third embodiment)
Next, a third embodiment of the present invention will be described in detail with reference to the drawings.

  Referring to FIG. 10, the third embodiment of the present invention includes a new document management unit 1001 that stores a new document in which difference information is applied to an original document, in addition to the configuration of the first embodiment.

  Each of these means generally operates as follows.

  In the first embodiment, when evaluating a search expression, a part or all of the difference information is applied to the original document as needed separately for each search expression. The application of difference information to the original document is centrally managed by 1001 so that the application result is shared across all search expressions, and the same difference is not applied multiple times.

  Specifically, the difference analysis unit 104 passes the difference information to the full-text search expression evaluation unit 106 and the difference search expression evaluation unit 105, and also transfers the difference information to the new document management unit 1001.

  On the other hand, the full-text search expression evaluation unit 106 and the difference search expression evaluation unit 105 request the new document management unit 1001 for necessary portions when a part or all of the new document in which the difference information is applied to the original document is necessary. To get.

  When the new document management unit 1001 takes out the original document from the original document table 109 and requests a part or all of the new document from the full-text search expression evaluation unit 106 and the difference search expression evaluation unit 105, the new document management unit 1001 Check whether a part of the document exists, if not, generate a part of the new document by applying only the difference information related to the requested part to the original document, and save this as a request source Returns that information.

  As a result, when it is necessary to apply the difference information to the same part in the difference application to the original document required for the individual search formula, the application process can be completed once for each difference information, and the overlapping difference application Processing can be omitted.

  In the first embodiment, all combinations of differences and search formulas that cannot be evaluated by the differential search formula evaluation unit 105 are processed by the full-text search formula evaluation unit 106. It is not limited that the evaluation unit 106 has to process.

  For example, a third retrieval formula evaluation unit using a method different from the present invention is further provided, and in combination with the differential retrieval formula evaluation unit 105 and the full-text retrieval formula evaluation unit 106 according to the present invention, the difference analysis unit 104 includes a first retrieval formula evaluation unit. A general engineer assigns a part of the combination of the difference and the search expression assigned to the difference search expression evaluation unit 105 and the full-text search expression evaluation unit 106 in the embodiment to the third search expression evaluation unit. You can easily imagine if you have.

  Next, the operation of the best mode for carrying out the present invention will be described using specific examples.

  FIG. 11 is a configuration diagram of the message filtering apparatus according to the present embodiment. The message filtering apparatus applies a filter condition given in advance to the received message and outputs the result to the outside. The filter condition is given as an XPath expression, and the corresponding node and its descendants are output as a result.

  According to FIG. 11, the message filtering apparatus 1101 of this embodiment includes a search expression management unit 102, an original document management unit 103, a difference analysis unit 104, a difference search expression evaluation unit 105, and a full-text search expression evaluation unit 106. An output unit 107, a new document management unit 1001, a communication unit 1102, and a management unit 1103.

  The communication unit 1102 receives the external message 115 and passes the difference information included therein to the difference analysis unit 104. In the management unit 1103, the administrator of the message filtering device 1101 registers the filter condition (search formula 111) and the original document 112 in the message filtering device 1101 through the user interface.

  XML is used for structured documents, and XPath is used for all search expressions.

(Example operation)
(Example operation / pre-registration)
Here, when two of "/ A / B / C" and "/ A / B [C = t1] / D" are registered as search formulas 111 through the management unit 1103, the search formula management unit 102 will both be registered. The contents of the search expression are checked, and "/ A / B / C" includes only a simple parent-to-child reference relationship, so "/ A / B [C = t1] / D" is simple Since it includes a predicate part in addition to the structural formula, it is classified as a predicate-added formula, and the result is stored in the search formula table 108. The search formula table 1201 in FIG. 12 is an example of the search formula table 108 in which search formulas are registered.

  Here, "/ A / B / C" means "node of tag C in tag B in tag A in root node" in XPath, and "/ A / B [ "C = t1] / D" is "a node with a tag of D in a tag of B in a tag of A under the root node, and a value of the tag of C in B is t1" Meaning.

  Next, as a template of frequently used messages (original documents), when the document 1301 that is the structured document in FIG. 13 is registered as an original document from the management unit 1103, the original document management unit 103 stores the document in each form. IDs are assigned to nodes in the order of appearance, and the ID and original document are stored in the original document table 109. Here, FIG. 14 shows an original document 1401 in which an ID is expressed as a tree structure, and numbers in parentheses indicate the ID.

  Next, the original document management unit 103 applies two already registered XPath expressions to the original document, and saves the result and processing process information in the original document table 109. In this case, two nodes satisfying the condition of “/ A / B / C” are ID = 3 and ID = 8. The node evaluated as a condition for deriving these results is a search expression “/ Nodes corresponding to 'A', 'B', and 'C' in A / B / C, respectively. The condition node set for ID = 3 is ID = 1, ID = 2, ID = 3, ID = The condition nodes for 8 are ID = 1, ID = 7, ID = 8, and the resulting output nodes are ID = 3 and below ID = 3 and ID = 4, ID = 8 and below ID = 8 And ID = 9. Since “/ A / B / C” is a simple structure, it has no candidate node set.

  On the other hand, a node satisfying the condition of “/ A / B [C = t1] / D” has ID = 5, but the node set evaluated as this condition is “/ A / B [C = t1] / ID = 1, ID = 2, ID = 3, ID = 4, ID = 5, which are text nodes corresponding to 'A', 'B', 'C', 'D' and 't1' The output nodes are ID = 5 and ID = 6, where ID = 5 or less. Furthermore, "/ A / B [C = t1] / D" is a predicate expression, and the expression "/ A / B / D" excluding the predicate [C = t1] is applicable, but the predicate [C = t1] The node that is no longer applicable by the part of the right half of the tree where the part of ID = 9 is not applicable, the candidate node set is equivalent to the conditional node set of this part, that is, ID = 1, ID = 7, ID = 8, ID = 9, ID = 10.

  An original document table 1501 in FIG. 15 is an example of the original document table 109 in which original documents, results, and processing process information are registered.

  FIG. 16 and FIG. 17 illustrate images of processing process information of “/ A / B / C” and “/ A / B [C = t1] / D”. That is, the process information indicates what relationship each node of the original document has with respect to the evaluation result of the XPath expression.

(Execution example / Search processing for difference information)
(Difference example 1)
Next, consider a case where a message 115 including difference information “replace document 1 node ID = 11 with“ t5 ”” arrives at communication section 1102. This corresponds to a representation of the structured document of the new document 1801 in FIG. 18 as difference information from “Document 1”. First, the communication unit 1102 extracts difference information included in the message 115 and passes it to the difference analysis unit 104. The difference analysis unit 104 first retrieves the document 1 stored in the original document table 109, and checks whether or not the node with ID = 11 is a value node. Here, since the node with ID = 11 is a text node, this is a value node, and the difference information includes only this difference, so this difference information is classified as a value difference.

  Next, the difference analysis unit 104 obtains a list of registered expressions and their types from the search expression management unit 102, classifies the difference information value difference, and the search structure simple structure expression and predicate expression classification result Therefore, it is determined that both formulas can be processed using the difference information, and the search formula, the difference information, and the type of the difference information are passed to the differential search formula evaluation unit 105.

  Next, when the difference search expression evaluation unit 105 receives the set of search expressions, the difference information, and the type of the difference information, first, the search expression “/ A / B / C” is obtained from the original document management unit 103 with “/ A / B”. The process information for the original document of “/ C” is acquired, and this is compared with ID = 11, which is the rewrite target node of the difference information. In this case, since the result node set and the condition node set do not include ID = 11, it is determined that the application result of “/ A / B / C” for the new document 1801 is the same as the original document, and the original document The portion of the processing result after ID = 3 and ID = 8 below the two nodes is sent to the output unit 107, and as a result, the corresponding part of the search expression “t1” and “t3” is output. Similarly, since “/ A / B [C = t1] / D” does not include ID = 11 in both the result node set and the condition node set, the portion below the node of ID = 5 that is the processing result of the original document As a result, “t2” is output.

  Therefore, in this case, since the difference information is irrelevant to the evaluation and the result of the XPath expression, the result is output without applying any difference or re-evaluating the search expression.

(Difference example 2)
Next, consider a case where the message 115 including the difference information “Replace the node with ID = 11 with“ t5 ”and replace the node with ID = 4 with“ t6 ”in document 1” reaches the communication unit 1102. . This corresponds to a representation of the structured document of the new document 1901 in FIG. 19 as difference information from “Document 1”.

  When this difference information is passed to the difference analysis unit 104, the difference analysis unit 104 checks whether ID = 4 and ID = 11 are both value nodes, and both are value nodes. are categorized.

  Next, the difference analysis unit 104 obtains a list of registered expressions and their types from the search expression management unit 102, and processes both of the expressions using difference information from the difference information classification result and the search expression classification result. It is determined that it is possible and is passed to the difference search expression evaluation unit 105.

  Next, when the difference search expression evaluation unit 105 receives the set of search expressions, the difference information, and the type of the difference information, first, the search expression “/ A / B / C” is obtained from the original document management unit 103 with “/ A / B”. The processing process information for the original document of “/ C” is acquired, and this is compared with ID = 4 and ID = 11, which are rewrite target nodes of the difference information. In this case, since the result node set includes ID = 4, the difference search expression evaluation unit 105 instructs the new document management unit 1001 to apply the difference for ID = 4.

  Since the new document management unit 1001 has not applied any difference yet, the new document management unit 1001 takes out the original document from the original document management unit 103, applies the difference of “replace node with ID = 4 to“ t6 ””, and ID = The content of 4 is rewritten to "t6" and the result is returned, and the part below the two nodes of ID = 3, ID = 8, which is the processing result of the original document, is sent to the output unit 107. As a result, "t6", The corresponding part of the search expression "t3" is output.

  Next, the processing process information of “/ A / B [C = t1] / D” is checked. Here, since the condition node set includes ID = 4, the difference search expression evaluation unit 105 is a new document management unit. Instruct 1001 to apply the difference for ID = 4. Since the new document management unit 1001 has already applied the difference for ID = 4, the new document management unit 1001 returns the result with reference to the stored content. Next, the difference search expression evaluation unit 105 re-evaluates the expression “/ A / B [C = t1] / D” for ID = 1, 2, 3, 4, 5 in the condition node set. Here, since the condition “C = t1” is no longer satisfied, the result of ID = 5 for the original document no longer satisfies the condition and is excluded from the result for the new document. In addition, since the candidate node set does not include ID = 4, the expression is not re-evaluated for the nodes with ID = 1, 7, 8, 9, 10, and as a result, "/ A / B [C = t1] "/ D" is output as there is no corresponding part.

  Therefore, in this case, only the difference “Replace node with ID = 4 with“ t6 ”” related to the evaluation and result of XPath expression is applied, and the difference with “Replace node with ID = 11 with“ t5 ”” is applied. Does not apply. In addition, re-evaluation of the expression is performed only for the expression of “/ A / B [C = t1] / D” with respect to the nodes of ID = 1, 2, 3, 4, and 5. "/ A / B / C" does not re-evaluate the expression, only the output result is changed.

(Difference example 3)
Next, let us consider a case where a message 115 including difference information “Replace node with ID = 9 with“ t1 ”in document 1” reaches communication unit 1102. This corresponds to a representation of the structured document of the new document 2001 in FIG. 20 as difference information from “Document 1”. The difference information is transferred from the difference analysis unit 104 to the difference search expression evaluation unit 105, and is the same as the difference example 1.

  Next, when the difference search expression evaluation unit 105 receives the set of search expressions, the difference information, and the type of difference information, the difference search expression evaluation unit 105 checks the processing process information for the search expression “/ A / B / C” and the result node set is ID Since = 9 is included, the new document management unit 1001 is instructed to apply the difference with respect to ID = 9 as in the case of the difference example 2, and outputs ID = 3 and nodes with ID = 8 or less as a result.

  Next, the processing process information of "/ A / B [C = t1] / D" is checked, but here the condition node set and result node set do not include ID = 9, so the processing result in the original document ID = 5 is output as a result.

  On the other hand, since the candidate node set includes ID = 9, a difference application instruction for ID = 9 is issued to the new document management unit 1001 and the difference has already been applied. Re-evaluate "/ A / B [C = t1] / D" for the nodes with ID = 1, 7, 8, 9, 10 described in the set. In this case, the condition described in the predicate part is satisfied by the substitution of ID = 9, and the result is ID = 10. Therefore, a node having ID = 10 or less is output as a result, and as a result, The difference search expression evaluation unit 105 outputs two ID = 5 and ID = 10 as the result of “/ A / B [C = t1] / D” together with the result in the previous original document.

  Therefore, in this case, the part of the evaluation result in the original document is not re-evaluated, and only the part of the search expression that may satisfy only a part of the new document that may satisfy the search expression due to the change by difference information. Re-evaluation is performed.

(Difference example 4)
Next, consider a case where a message 115 including difference information “Replace node with ID = 3 with“ t1 ”in document 1” reaches communication unit 1102. This corresponds to a representation of the structured document of the new document 2101 in FIG. 21 as difference information from “Document 1”.

  When this difference information is passed to the difference analysis unit 104, since the node with ID = 3 is not a value node but an element node, the difference is classified as a structure difference, and as a result, it is determined that calculation using the difference cannot be performed. Then, the difference information and the search formula are passed to the full-text search formula evaluation unit 106.

  The full-text search expression evaluation unit 106 instructs the new document management unit 1001 to apply all the differences, and evaluates each of the two search expressions registered for the entire completed new document, and "/ A / B / C As a result, only ID = 8 is output for “/ A / B [C = t1] / D”, and “not applicable” is output as a result.

  Therefore, in this case, even when the difference information is complicated and the difference between the processing result / processing process information and the difference information in the original document cannot be calculated, the processing efficiency cannot be improved, but the evaluation result of the search expression is output. Things will be possible.

(Difference example 5)
Next, consider a case where an XPath “/ A / B / * [2]” is further registered as a search expression. The search formula table 2201 in FIG. 22 is an example of the search formula table 108 to which “/ A / B / * [2]” is added. This expression means "the node that appears second in any tag that appears in the tag B in the tag A under the root node", and a simple parent-child "second in any tag" Includes relationships between nodes that are not relationships. In this case, the search formula management unit 102 classifies the search formula as a complex formula and stores the result in the search formula table 2201.

  Here, as in the difference example 1, consider a case where the message 115 including the difference information “replace the node with ID = 11 of document 1 with“ t5 ”” reaches the communication unit 1102.

  In this case, the difference analysis unit 104 classifies the difference information as a value difference, and as a result, for the search expressions “/ A / B / C” and “/ A / B [C = t1] / D”, the difference example 1 In the same manner as above, it is determined that the search based on the difference information is possible, and it is passed to the difference search expression evaluation unit 105 and evaluated in the same manner. The output results of these two search expressions are also the same as in Difference Example 1.

  On the other hand, since “/ A / B / * [2]” is a complex expression, it is always passed to the full-text search expression evaluation unit 106 regardless of the type of difference information. The full-text search expression evaluation unit 106 instructs the new document management unit 1001 to apply all the differences, searches only “/ A / B / * [2]” for the entire new document, and outputs the result. . In this case, nodes with ID = 5 and ID = 10 are output as results.

  Similarly, in the case of other difference information examples, since "/ A / B / * [2]" is a complex expression, "/ A / B / * [[ 2] "is always evaluated by the full-text search expression evaluation unit 106.

  Therefore, in this case, even if the search expression is complex and the difference calculation cannot be performed based on the processing result / process process information and difference information in the original document, other simple search can be performed by evaluating only such a search expression. If possible, the calculation using the difference information is possible.

  Although the embodiments and examples of the present invention have been described above, the present invention is not limited to the above examples, and various other additions and modifications can be made. The structured document processing apparatus 101 in each embodiment and the message filtering apparatus 1101 in the examples can be realized by a computer and a program in addition to being realized by hardware. The program is provided by being recorded on a computer-readable recording medium such as a magnetic disk, read by the computer when the computer is started up, and the operation of the computer is controlled to thereby execute the above-described embodiment on the computer. Search formula management unit 102, original document management unit 103, difference analysis unit 104, difference search formula evaluation unit 105, full-text search formula evaluation unit 106, output unit 107, difference calculation unit 901, input document storage unit 902, new document management unit 1001, a communication unit 1102, and a management unit 1103 are realized, and the computer executes the processes in the above-described embodiments.

  INDUSTRIAL APPLICABILITY According to the present invention, the present invention can be applied to applications such as a communication device that filters and monitors communication messages according to rules specified by a search expression.

It is a block diagram which shows the structure of the 1st Embodiment of this invention. It is a figure which shows an example of a structure of the search expression table in the 1st Embodiment of this invention. It is a figure which shows an example of a structure of the original document table in the 1st Embodiment of this invention. It is a flowchart which shows operation | movement of the search expression management part in the 1st Embodiment of this invention. It is a flowchart which shows operation | movement of the original document management part in the 1st Embodiment of this invention. It is a flowchart which shows operation | movement of the difference analysis part in the 1st Embodiment of this invention. It is a figure which shows the correspondence of the combination of the type | formula of a search formula and difference information, and a search method in the 1st Embodiment of this invention. It is a flowchart which shows operation | movement of the difference retrieval type | formula evaluation part in the 1st Embodiment of this invention. It is a flowchart which shows operation | movement of the full-text search expression evaluation part in the 1st Embodiment of this invention. It is explanatory drawing of the search algorithm in the 1st Embodiment of this invention. It is a block diagram which shows the structure of the 2nd Embodiment of this invention. It is a block diagram which shows the structure of the 3rd Embodiment of this invention. It is a block diagram which shows the structure of the Example of this invention. It is a figure which shows an example of a structure of the search expression table in the Example of this invention. It is explanatory drawing of the original document in the Example of this invention. It is explanatory drawing of the original document which provided ID in the Example of this invention. It is a figure which shows an example of a structure of the original document table in the Example of this invention. It is explanatory drawing of the process process information in the Example of this invention. It is explanatory drawing of the process process information in the Example of this invention. It is explanatory drawing of the new document in the Example of this invention. It is explanatory drawing of the new document in the Example of this invention. It is explanatory drawing of the new document in the Example of this invention. It is explanatory drawing of the new document in the Example of this invention. It is a figure which shows an example of a structure of the search expression table in the Example of this invention. It is a block diagram which shows the structure of the structured document processing apparatus of a prior art.

Explanation of symbols

DESCRIPTION OF SYMBOLS 101 ... Structured document processing apparatus 102 ... Search formula management unit 102a ... Classification unit 102b ... Storage unit 103 ... Original document management unit 103a ... Search formula evaluation unit 103b ... Storage unit 104 ... Difference analysis unit 104a ... Classification unit 104b ... Evaluation unit Determination unit 105 ... difference search expression evaluation unit 106 ... full-text search expression evaluation unit 107 ... output unit 108 ... search expression table 109 ... original document table 111 ... search expression 112 ... original document 113 ... difference information 114 ... structured document 115 ... message 901 ... Difference sentence calculation unit 902 ... Input document storage unit 1001 ... New document management unit 1101 ... Message filtering device 1102 ... Communication unit 1103 ... Management unit 1201 ... Search formula table 1301, 1401 ... Original document 1501 ... Original document table 1601, 1701 ... original document 1801 ... new document 1
1901 ... New Document 2
2001 ... New Document 3
2101 ... New document 4
2201 ... Search formula table 2301 ... Input unit 2302 ... Difference calculation unit 2303 ... Difference reflection unit 2304 ... Output unit 2305 ... Original document storage unit 2306 ... Original document processing result storage unit

Claims (18)

  A search expression management unit that classifies and holds search expressions based on their properties, an original document management unit that stores an original document together with a result of applying the original document to the held search expression and processing process information at the time of application, and a structure A difference analysis unit that classifies document difference information based on its properties and determines a search expression evaluation method based on the type of difference and the type of search expression; A search formula evaluation unit that evaluates a search formula using information, a search formula and its classification result, an original document, and a search formula evaluation result for the original document, and a search formula evaluation using differential information is impossible In this case, a new document in which difference information is applied to the original document and a search expression is applied to the new document, and the results of the difference search expression evaluation section and the full-text search expression evaluation section are combined into the original document. Retrieval formula for documents with difference information applied Structured document processing system characterized by comprising an output section for outputting the evaluation result.   The search formula management unit classifies the search formula as a simple structural formula when the search formula includes only the parent-child relationship of the nodes of the structured document, and the search formula adds a predicate including the simple structural formula to the simple structural formula. If it is a thing, it is classified as a predicate expression, and other expressions are classified as complex expressions. The difference analysis unit classifies the difference information according to whether the changed part is a value or a document structure, and the changed part is If only the value is classified as a value difference, and if the difference information is a value difference and the search formula is a simple structure formula or a predicate-added formula, it is determined that the search formula evaluation using the difference information is possible The structured document processing system according to claim 1, wherein the structured document processing system is one.   The original document management unit includes a result node set that is a list of nodes included in the application result of the search expression and a list of nodes that are directly evaluated by the search expression to derive the application result of the search expression. 3. The structured document processing system according to claim 2, wherein the structured document processing system is stored as a conditional node set.   The original document management unit further includes a candidate node that is a list of nodes that are determined to be not included in the search expression, including a predicate included in the search expression, an expression other than the predicate being applicable, and an evaluation result of the predicate 4. The structured document processing system according to claim 3, wherein the set is stored.   The original document management unit stores the original document with a unique ID assigned to each node and designates a node set of processing process information by the ID, and the difference analysis unit identifies the change information by the ID. 5. The structured document processing system according to claim 4, wherein the system is designated.   The difference search expression evaluation unit determines whether or not there is a change to the output of the processing result for the original document by the result node set when the difference is the value difference and the search expression is the simple structural expression. If there is no change, the processing result is used as the evaluation result of the search formula as it is, and if there is a change, it is applied and the result obtained by applying the difference to the processing result of the original document is used as the evaluation result. Item 6. The structured document processing system according to Item 4 or 5.   The difference search expression evaluation unit further outputs a search expression evaluation for the original document when the difference is the value difference and the search expression is a pre-descriptive expression, and the change according to the difference is performed by the condition node set. Check whether there is a possibility that the result is invalid, and if there is a possibility, apply only the difference related to the condition node set and reevaluate the search expression for the part, and the condition node Whether or not there is a change in the output of the processing result for the original document by the result node set when there is no possibility in the search of the set or when it is determined that the search expression corresponds to the search expression as a result of the re-evaluation of the search expression If there is no change, the processing result is used as the evaluation result of the search expression as it is, and if there is a change, it is applied and the result of applying the difference to the processing result of the original document is used as the evaluation result. , Whether or not there is a possibility that a node determined by the candidate node set that the change due to the difference is not applicable in the search expression evaluation for the original document may be applicable. Apply only the differences related to the set and re-evaluate the search expression for that part, and if the result is determined to be applicable to the search expression, check whether there is a change to the output result of the corresponding part, 7. The structured document processing system according to claim 6, wherein when there is no change, the output result is directly used as an evaluation result of a search expression, and when there is a change, an evaluation result is applied.   A difference calculation unit that calculates a difference by comparing the input structured document with the original document; and an input document storage unit that stores the input structured document. The difference search expression evaluation unit and the full-text search 8. The structured document processing according to claim 1, wherein the expression evaluation unit refers to the input document stored in the input document storage unit instead of applying the difference to the original document. system.   A new document management unit that manages a new document in which a part or all of the differences are applied to the original document is further included, and the difference search expression evaluation unit and the full-text search expression evaluation unit are configured to apply the difference to the original document. Requests to the new document management unit to apply and obtain the result, and the new document management unit returns the application result from the saved new document without applying the difference again for the already applied difference. Item 8. The structured document processing system according to any one of Items 1 to 7. A storage unit that stores one or more search expressions and their types, an original document, a result of applying the search expression to the original document, and processing process information;
The difference information of a given structured document is classified based on its property, and the difference information is classified for each search expression according to the type of the difference information and the type of the search expression stored in the storage unit. A difference analysis unit that determines whether or not to perform an evaluation, and
A difference search expression evaluation unit that performs evaluation using a result of applying the search expression to the original document stored in the storage unit and processing process information for the search expression determined to be evaluated using difference information; ,
A full-text search expression evaluation unit that performs evaluation using a document in which the difference information is applied to an original document stored in the storage unit for a search expression determined to be evaluated without using difference information. A structured document processing system. The storage unit stores a simple structural formula search formula including only the parent-child relationship of the nodes of the structured document, a result of applying the search formula to the original document, and processing process information.
The difference analysis unit determines that when the difference information of the given structured document is a value difference in which the changed portion is only a value, the search formula for the simple structural formula is evaluated using the difference information. ,
The differential retrieval formula evaluation unit evaluates the simple structural formula retrieval formula using a result of applying the simple structural formula retrieval formula to the original document stored in the storage unit and processing process information. The structured document processing system according to claim 10. In the storage unit, a predicate-added search expression obtained by adding a predicate including a simple structural formula to a simple structural formula including only the parent-child relationship of the nodes of the structured document, and a result and processing of applying the search formula to the original document Process information is stored,
The difference analysis unit determines that when the difference information of the given structured document is a value difference in which the changed part is only a value, the search expression with the predescription word is evaluated using the difference information. And
The difference search expression evaluation unit evaluates a search expression with a pre-descriptive word using a result obtained by applying the search expression with a pre-description word to the original document stored in the storage unit and processing process information. 11. The structured document processing system according to claim 10, wherein the structured document processing system is performed. In a method for evaluating a search expression for a structured document using a computer,
The computer adds one or more types of search expressions to the storage unit and stores the result in the storage unit by adding a result of applying the search expression to the original document and processing process information. A first step of storing;
The computer classifies the difference information of a given structured document based on its property, and for each search formula according to the type of the difference information and the type of search formula stored in the storage unit A second step of determining whether to perform the evaluation using the difference information;
For the retrieval formula determined by the computer to be evaluated using the difference information, the evaluation is performed using the result of applying the retrieval formula to the original document stored in the storage unit and the process information, and the difference information And a third step of evaluating, using a document obtained by applying the difference information to the original document stored in the storage unit, for a search expression determined to be evaluated without using Document processing method. In a method for evaluating a search expression for a structured document using a computer,
The computer adds a simple structural formula search formula including only the parent-child relationship of the nodes of the structured document, a result of applying the search formula to the original document, and processing process information to the storage unit. A first step of storing;
When the computer classifies the difference information of a given structured document based on its property and the change part is a value difference whose value is only a value, the retrieval formula of the simple structural formula is evaluated using the difference information A second step of determining to perform
A third step in which the computer evaluates the retrieval formula of the simple structural formula using a result of applying the retrieval formula of the simple structural formula to the original document stored in the storage unit and processing process information; A structured document processing method comprising: In a method for evaluating a search expression for a structured document using a computer,
A result of applying the search formula to the original document by adding a predicate-added search formula in which the computer adds a predicate including the simple structural formula to a simple structural formula including only the parent-child relationship of the nodes of the structured document; And a first step of adding the process information and storing it in the storage unit;
When the computer classifies the difference information of a given structured document based on its property, and the change part is a value difference whose value is only a value, the difference information is used for a search expression with a pre-descriptive expression A second step for determining to perform the evaluation;
A third evaluation in which the computer evaluates a search formula with a pre-description word using a result of applying the search formula with a pre-description word to an original document stored in the storage unit and processing process information; A structured document processing method comprising: steps. On the computer,
A first type of one or more retrieval formulas is added and stored in the storage unit, and a result of applying the retrieval formula to the original document and processing process information are added and stored in the storage unit. And processing
The difference information of a given structured document is classified based on its property, and the difference information is classified for each search expression according to the type of the difference information and the type of the search expression stored in the storage unit. A second process for determining whether or not to use the evaluation;
For the retrieval formula determined to be evaluated using the difference information, the evaluation is performed using the result of applying the retrieval formula to the original document stored in the storage unit and the process information, and the difference information is not used. A program for causing a search expression determined to be evaluated to perform a third process of performing an evaluation using a document in which the difference information is applied to an original document stored in the storage unit. On the computer,
A first search formula that includes only a parent-child relationship between nodes of a structured document, a result obtained by applying the search formula to the original document, and processing process information are added to the storage unit and stored in the storage unit. And processing
The difference information of a given structured document is classified based on its property, and when the changed part is a value difference whose value is only a value, it is determined that the retrieval formula of the simple structural formula is evaluated using the difference information A second process to
To perform a third process for evaluating the simple structural formula search formula using the result of applying the simple structural formula search formula to the original document stored in the storage unit and the process information. Program. On the computer,
A search expression of a predicate expression in which a predicate including a simple structure expression is added to a simple structure expression including only a parent-child relationship of nodes of a structured document, a result of applying the search expression to the original document, and processing process information And a first process of storing in the storage unit,
If the difference information of a given structured document is classified based on its properties, and the change part is a value difference whose value is only the value, the search expression with a pre-descriptive word expression is evaluated using the difference information A second process to determine;
For the search expression with a pre-descriptive word, a third process for performing an evaluation using the result of applying the search expression with the pre-description word to the original document stored in the storage unit and processing process information is performed. Program to let you.

Images