Structured document retrieval system

Claim 2

A structured document retrieval apparatus for storing structured documents and for retrieving the stored documents in accordance with a retrieval expression, each of said structured documents comprising a plurality of components and logical structures therebetween that define a hierarchical relationship between components, said apparatus comprising: a registration unit for conducting a preprocess for storing a structured document received from an input unit; a logical structure extraction unit for incorporating knowledge of the logical structures of the structured documents, and for extracting a select logical structure, a document attribute, and a document content from the structured documents in accordance with the knowledge of the logical structures; a document information storage unit for storing the extracted logical structure, document attribute, and document content while grouping them by kind; a logical structure decision unit for incorporating knowledge of logical structures of the structured documents, and for deciding whether or not a logical structure read out from said document information storage unit coincides with a logical structure written in the retrieval expression; and a retrieval unit for conducting a retrieval by reading out from said document information storage unit only information which is an ultimate retrieval object of a structured document in which the logical structure coincides with that of the retrieval expression, for rearranging a structured document obtained from the retrieval, and for outputting the rearranged structured document.

Claim 3

A structured document retrieval apparatus as recited in claim 1, further comprising: an independently disposed logical structure knowledge base for storing the knowledge of logical structures of the structured documents, wherein said logical structure extraction unit and said logical structure decision unit commonly use said logical structure knowledge base.

Claim 4

A structured document retrieval apparatus comprising: a document input unit for inputting a structured document, said structured document comprising a plurality of components and logical structures therebetween that define a hierarchical relationship between components; a logical structure and document content extraction unit for extracting a logical structure and a document content from the structured document received from said document input unit; a document information storage unit for storing the logical structure and the document content extracted by said logical structure and document content extraction unit, while adding an identifier to the logical structure and the document content; a retrieval expression input unit for inputting a retrieval expression having retrieval conditions based on a logical structure and a document content; and a retrieval unit for conducting a retrieval by reading out from said document information storage unit only information which is an ultimate retrieval object of a structured document in which the logical structure coincides with that of the retrieval expression wherein said retrieval unit comprises: a logical structure decision unit for retrieving logical structure from said document information storage unit which coincides with the logical structure conditions of the retrieval expression, a fetch unit for fetching from said document information storage unit a document content corresponding to an identifier of a retrieval result of the logical structure decision unit, an evaluation unit for retrieving and evaluating the fetched document content based on document content conditions of the retrieval expression, and an output unit for combining the document content obtained as a retrieval and evaluation result from the evaluation unit with a logical structure corresponding to an identifier of the document content, and for outputting the combination.

Claim 5

A structured document retrieval apparatus comprising: a document input unit for inputting a structured document, said structured document comprising a plurality of components and logical structures therebetween that define a hierarchical relationship between components; a logical structure and document content extraction unit for extracting a logical structure and a document content from the structured document received from said document input unit; a document information storage unit for storing the logical structure and the document content extracted by said logical structure and document content extraction unit, while adding an identifier to the logical structure and the document content; a retrieval expression input unit for inputting a retrieval expression having retrieval conditions based on a logical structure and a document content; and a retrieval unit for conducting a retrieval by reading out from said document information storage unit only information which is an ultimate retrieval object of a structured document in which the logical structure coincides with that of the retrieval expression wherein said retrieval unit comprises: a logical structure decision unit for retrieving logical structure from said document information storage unit which coincides with the logical structure conditions of the retrieval expression, a fetch unit for fetching from said document information storage unit a document content corresponding to an identifier of a retrieval result of the logical structure decision unit, an evaluation unit for retrieving and evaluating the fetched document content based on document content conditions of the retrieval expression, and an output unit for combining the document content obtained as a retrieval and evaluation result from the evaluation unit with a logical structure corresponding to an identifier of the document content, and for outputting the combination.

Claim 6

A structured document retrieval apparatus as recited in claim 1, wherein said document information storage unit collectively stores in one area document contents belonging to document components which are included in the structured document and are of the same kind.