JP2009176044A - Document retrieving method, device, program, and computer readable recording medium - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To time-axially retrieve a document which is changed by new generation, overwritten saving, and deletion, related to the document to be preserved in a general purpose file server without replacing an existing file system. <P>SOLUTION: A document retrieving device receives a document operation report of new document generation, updating, or deletion, and the document from a shared document server, and stores them in a temporal transposition index storage means after imparting document identifiers by each word of the document. A document list including information (operation dates) about the document identifier and the document change and the documents is stored in a document list storage means. When the plurality of document identifiers are present in a retrieval result which is obtained by retrieving the temporal transposition index storage means and the document list storage means on the basis of the retrieval conditions (a keyword, a designation date, and a document name) of an input retrieval request, the document with the older operation date is deleted from the retrieval result and outputted. <P>COPYRIGHT: (C)2009,JPO&INPIT

Description

  The present invention relates to a document search method and apparatus, a program, and a computer-readable recording medium, and more particularly, a document search method and apparatus for searching a document stored without replacing an existing file system in a general-purpose file server, and The present invention relates to a program and a computer-readable recording medium.

  By adopting a log structuring method that uses the transition of documents stored in the Linux local file system as an inverted index as a file management system, disk block overwrites are eliminated and damage in the event of a failure is minimized. There are Nilfs capable. Since this Nilfs is user-friendly, there is a technique for searching for snapshots using keywords (see, for example, Non-Patent Document 1).

Further, as a document search system, there are search engines of a full-text search system for searching an enormous number of documents, and Namazu used for small-scale searches such as personal file search (for example, see Non-Patent Document 3).
Amami et al., "Design and Implementation of Log Structured File System nilfs for Linux", IPSJ SIG Technical Report 2005-OS-99 (9) 2005/5/26, pp61-68. Namazu (http://www.namazu.org/)

  However, if a tense search is performed using the above-mentioned Namazu, there is a problem that the number of processes for creating an inverted index increases. Here, “temporal system” represents the time relationship (current, past, future, etc.) based on a point in time on the time axis, and “tense search” is based on the time axis. Say to search the search target.

  Also, when trying to search tense by Nilfs, since the transition of the document stored in the local file system is used as the inverted index, the transition of the document stored in the general-purpose file server (remote file system) is used as the inverted index. To do this, it is necessary to duplicate the document in advance and save it in the local file system.

  Also, when implementing a file server with Nilfs, it is necessary to change the file system to Nilfs in advance.

  Furthermore, even if Nilfs and Namazu are simply combined, the transition of Namazu's transposed index is stored in Nilfs and the latest Namazu's transposed index is searched, so the above problem is still not solved.

  The present invention has been made in view of the above points, and for a document (file) stored on a general-purpose file server without replacing an existing file system, it is possible to search for a document that has changed due to new creation, overwriting, and deletion. An object of the present invention is to provide a simple document search method and apparatus, program, and computer-readable recording medium.

  FIG. 1 is a diagram for explaining the principle of the present invention.

The present invention (Claim 1) is a document search method in a search device connected to a shared document server and searching for a document,
The document collecting means receives a document operation notification for creating, updating, or deleting a new document from the shared document server and copies the document, assigns a document identifier for each document, and stores it in the duplicate document storage means. (Step 1) Document collection step;
The indexing unit obtains a document from the duplicate document storage unit, assigns a document identifier to each word of the document, stores the document identifier in the time-based transposed index storage unit (step 2), and includes the document identifier and the operation date and time. Storing the document list including the information on the transition of the document and the document in the document list storage means (step 3);
The search request analyzing means analyzes the input search request and extracts search attributes and search conditions (step 4).
The search execution means searches the word from the temporally transposed index storage means based on the search condition, acquires a document identifier (transposed index) corresponding to the word (step 5), and the document list storage means based on the transposed index To obtain a search result document (step 6),
When there are a plurality of document identifiers of the document acquired in the search step, the search result combining means deletes the document with the old operation date from the search result and outputs it (step 7).

Further, the present invention (Claim 2), in the search step,
If the date, time of creation, update, or deletion of a document is specified as a search condition corresponding to the search attribute, a document identifier that matches the date is obtained from the document list storage unit, and the document identifier Get the document to be used.

Further, according to the present invention (Claim 3), in the search step,
When a document name is specified as a search condition corresponding to the search attribute, a document identifier matching the document name is acquired from the document list storage unit, and a document corresponding to the document identifier is acquired.

  FIG. 2 is a principle configuration diagram of the present invention.

The present invention (Claim 4) is a document retrieval apparatus that is connected to a shared document server and retrieves a document,
Document collection means for receiving a document operation notification and document for creating, updating, or deleting a new document from a shared document server, copying the document, assigning a document identifier for each document, and storing it in the duplicate document storage means 170 175,
A document is acquired from the duplicate document storage unit 170, a document identifier is assigned to each word of the document, and the document is stored in the temporal displacement index storage unit 155. The document identifier, information on the transition of the document including the operation date and time, and the document Indexing means 160 for storing a document list including the document list in the document list storage means 150;
A search request analysis unit 130 that analyzes the input search request and extracts search attributes and search conditions;
A search is performed by searching for a word from the tense transposed index storage unit 155 based on the search condition, obtaining a document identifier (transposed index) corresponding to the word, and searching the document list storage unit 150 based on the transposed index. Search execution means 140 for obtaining a result document;
When there are a plurality of document identifiers of the document acquired by the search execution unit 140, the search execution unit 140 includes a search result combination unit 135 that deletes and outputs a document with an old operation date and time from the search result.

Further, the present invention (Claim 5) is provided in the search execution means 140.
If the date, time of creation, update, or deletion of a document is specified as a search condition corresponding to the search attribute, a document identifier that matches the date is obtained from the document list storage unit, and the document identifier Means for obtaining a document to be executed.

Further, the present invention (Claim 6) is provided in the search execution means 140.
Means for acquiring a document identifier corresponding to the document name from the document list storage means and acquiring a document corresponding to the document identifier when a document name is specified as a search condition corresponding to the search attribute; Including.

  The present invention (Claim 7) is a document search program for causing a computer to function as each means constituting the document search apparatus according to any one of Claims 4 to 6.

  The present invention (Claim 8) is a computer-readable recording medium storing the document search program according to Claim 7.

  The present invention obtains and stores date and time information indicating transitions such as new creation / update / deletion of a document stored in a general-purpose shared document server, and stores the document by acquiring and storing the document. This makes it possible to conduct a tense search based on a search condition including a date and time specification without replacing the file system.

  Furthermore, the date and time of deletion can also be specified as a search key by acquiring and storing the date and time of deletion from the shared document server.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings.

  FIG. 3 shows the configuration of the tense search device according to the embodiment of the present invention.

  The tense search device shown in the figure includes a user interface unit 125, a search request analysis unit 130, a search result combination unit 135, a search execution unit 140, a document list unit 150, a tense transposed index storage unit 155, an indexing unit 160, and a morphological analysis unit. 165, a duplicate document unit 170, and a document collection interface unit 175.

  The user interface 125 is connected to the user terminal 110 via a LAN (Local Area Network), and transmits / receives a search syntax 120 and a search result 115 to / from the user terminal 110. The document collection interface unit 175 is connected to the shared document (file) server 180 via the LAN.

  The document collection interface unit 175 obtains a document acquired from the shared document (file) server 180, operation types such as new document creation, update (overwrite), and deletion, and operation times thereof, and issues a Document ID for each document. The obtained document is duplicated and stored in the duplicate document unit 170 together with the Document ID.

  The indexing unit 160 passes the acquired document to the morpheme analysis unit 165, acquires a word that is the result of the morpheme analysis, assigns a Document ID to each word, and stores it in the temporal transposed index storage unit 155. Further, information acquired from the shared document (file) server 180, Document ID, operation time, and the like are additionally written in the document list unit 150.

  The search request analysis unit 130 analyzes the search syntax input from the user terminal 110, extracts a combination of search attributes and search conditions corresponding to the search attributes, and outputs the combination to the search execution unit 140.

  The search execution unit 140 obtains a Document ID, which is a document identifier for identifying a document, as a temporally transposed index from the temporally transposed index storage unit 155 based on a search condition (word as a keyword) corresponding to the search attribute. Then, based on the Document ID, a new creation date / overwrite date / time or deletion date / time of the document for the tense search is acquired from the document list unit 150, and the search condition (time axis designation and file name (document Name)) is searched for a document in the document list unit 150, and a search result (a list of documents) and a temporally transposed index are output to the search result combining unit 135.

  The search result combining unit 135 eliminates duplication of documents with the same operation date and time in the same document, and outputs the result to the user terminal 110 via the user interface unit 125.

  First, a description will be given of the processing at the time of creating a temporally transposed index and a document list in the document collection interface unit 175 and the indexing unit 160 before performing the search processing.

  FIG. 4 is a flowchart of the inverted index creation / document list creation processing according to the embodiment of the present invention.

  When the document collection interface unit 175 receives a document operation notification from the shared document server 180 (step 101), the document collection interface unit 175 determines whether the message type of the document operation notification is newly created or updated (overwritten). If none of them (step 102, No), it is determined that the message type is document deletion, and the operation date / time information of the document operation notification message is set in the file deletion date / time area of the document list storage unit 150; Further, a deletion flag for the file (document) is set, and the document list storage unit 150 is updated (step 107). If it is any (step 102, Yes), the document collection interface unit 175 duplicates the document (file) acquired from the shared document server 180, creates a Document ID corresponding to the duplicated document, and creates the Document ID The copied document is stored in the duplicate document unit 170 (step 103).

  Next, the indexing unit 160 reads the document from the duplicate document unit 170 (step 104), the morpheme analysis unit 165 performs morpheme analysis, and passes the analysis result word and the Document ID to the indexing unit 160 (step 105). . As a result, the indexing unit 160 stores the structure in the tense transposed index storage unit 155 as shown in FIG. 5 (step 106).

  Further, as shown in FIG. 6, the indexing unit 160, based on the information read from the duplicate document unit 170, changes the file identifier (number for making the Document ID a logical one file), the transition of the file (version ) Number, document identifier (Document ID) that identifies the duplicate document, new file creation date / overwrite date, file body (text), shared document server name, folder name where the file is stored A certain path name and file name are stored in the document list section 150 (step 107). The above processing is performed every time a document operation notification is received.

  Next, processing at the time of search will be described.

  FIG. 7 is a flowchart of search processing according to an embodiment of the present invention.

  Step 210) The user interface unit 125 accepts a search syntax as a search request from the user terminal 110 via the LAN.

  Step 220) The search request analysis unit 130 analyzes the search syntax and extracts a combination of search attributes and search conditions. Details of this processing will be described later.

  Step 230) The search result execution unit 140 refers to the hourly transposed index storage unit 155 and acquires a document that matches the search condition from the document list unit 150. Details of this processing will be described later.

  Step 240) The search result combining unit 135 eliminates duplication of documents with the same operation date and time in the same document in the search result, and outputs it to the user interface unit 125. Details of this processing will be described later.

  Step 260) The user interface unit 125 returns the search result 115 to the user terminal 110.

  Next, the processing of the search request analysis unit 130 in step 220 of FIG. 7 will be described in detail.

  FIG. 8 is a flowchart of search syntax analysis processing according to the embodiment of the present invention.

  Step 221) The input search syntax is decomposed into combinations of search attributes and search conditions.

  Step 222) It is determined whether or not the number of combinations of the search attribute and the search condition is 0. If it is 0, the process proceeds to Step 225. If not, the process proceeds to Step 223.

  Step 223) Retrieve search attributes.

  Step 224) The search condition is extracted and stored in a memory (not shown) together with the search attribute extracted in Step 223, and the process proceeds to Step 222.

  Step 225) When the combination of the search attribute and the search condition is 0 in Step 222, it is determined whether or not the search attribute includes a full-text search specification. The process proceeds to 230, and if not included, the process ends.

  FIG. 9 shows an example of search attributes and search conditions extracted by the upper processing.

  The retrieved search attributes and search conditions are a keyword, date and time, and a file name. In the example of FIG. 9, the search is specified using the AND condition of the specified search keyword. Also, the date and time is specified by the date and time, and the search is limited to the files that existed at the specified date and time, or the files that existed in the specified period before and after the specified date and time are limited. To search. By specifying a part of the file name, a document including the specified file name can be searched.

  Next, the search execution process in the search execution unit 140 in step 230 of FIG. 7 will be described.

  FIG. 10 is a flowchart of search processing according to an embodiment of the present invention.

  Step 231) The search execution unit 140 acquires the Document ID paired with the searched word from the tense transposed index storage unit 155 based on the acquired search condition keyword of the search attribute, and stores it in a memory (not shown). Store. At this time, if a plurality of keywords are specified, the search is performed using an AND condition.

  Step 232) The tense transposed index storage unit 155 is searched for all the words corresponding to the keyword, and the corresponding Document ID is stored in a memory (not shown).

Step 233) It is determined whether or not there is a date specification in the search condition. To specify the date and time, as shown in FIG.
・ Specified date and time;
・ Before specified date and time;
・ After specified date and time;
・ Specified period;
There is a method of designation. If there is a date and time designation, the process proceeds to step 234, and if not, the process proceeds to step 235.

  Step 234) Based on the date / time designation acquired in step 232, the document list storage unit 150 is searched, a Document ID corresponding to the date / time designation is obtained, and stored in a memory (not shown).

  Step 235) It is determined whether or not a document name (file name) is specified in the search attribute. If there is, the process proceeds to Step 236, and if not, the process proceeds to Step 240.

  Step 236) The document list portion 150 is searched based on the document name (file name), and the Document ID corresponding to the document name (file name) is acquired and stored in the memory (not shown).

  Next, the processing of the search result combining unit 135 in step 240 of FIG. 7 will be described in detail.

  FIG. 11 is a flowchart of search result combining processing according to an embodiment of the present invention.

  Step 241) The document ID acquired from the search execution unit 140 in the above step 232 and stored in the memory (not shown) as the first key is acquired in the above step 234 and step 236, and the memory (see FIG. Document IDs stored in (not shown) are arranged in descending order on a memory (not shown) using the Document ID as the second key.

  Step 242) It is determined whether there are a plurality of second keys for the first key. If there are more than one, the process proceeds to step 243, and if not, the process proceeds to step 249.

  Step 243) The operation date and time corresponding to the Document ID corresponding to the second key is acquired from the document list portion 150.

  Step 244) It is determined whether or not there is an operation date and time to be compared. If there is, the process proceeds to Step 245, and if not, the process proceeds to Step 248.

  Step 245) If there is the same operation date and time for one Document ID, the process proceeds to Step 246, and if there is not, the process proceeds to Step 247.

  Step 246) If there is the same operation date and time in Step 245, it is duplicated, so the Document ID of the old operation date and time is deleted from the search result, and the process proceeds to Step 244.

  Step 247) If there is no overlap in Step 245, the next operation date and time is set, and the process proceeds to Step 244.

  Step 248) If there is no operation date and time to be compared in Step 244, the document to be compared next is set as the comparison source, and the process proceeds to Step 243.

  Step 249) If there are not a plurality of Document IDs as the second key in Step 242, the ranking for rearranging the search list is calculated. For the ranking calculation, for example, TF / IDF (Term Frequency / Inverted Document Frequency) is used.

  Step 250) The search result list is rearranged based on the ranking value.

  Note that the operation of the components of the above tense search device can be constructed as a program and installed in a computer used as the tense search device for execution, or distributed via a network.

  Further, the constructed program can be stored in a portable storage medium such as a hard disk, a flexible disk, or a CD-ROM, and can be installed or distributed in a computer.

  The present invention is not limited to the above-described embodiment, and various modifications and applications can be made within the scope of the claims.

  The present invention can be applied to search of a file system of a general-purpose file server.

It is a figure for demonstrating the principle of this invention. It is a principle block diagram of this invention. It is a block diagram of the tense search apparatus in one embodiment of this invention. It is a flowchart of the transposition index creation process in one embodiment of this invention. It is an example of the structure of the temporal displacement index storage part in one embodiment of this invention. It is an example of the structure of the document list part in one embodiment of this invention. It is a flowchart of the search process in one embodiment of this invention. It is a flowchart of the search syntax analysis process in one embodiment of the present invention. It is an example of the extracted search attribute and search condition in one embodiment of this invention. It is a flowchart of the search process in one embodiment of this invention. It is a flowchart of the search result combining process in an embodiment of the present invention.

Explanation of symbols

110 User terminal 115 Search result 120 Search syntax 125 User interface unit 130 Search request analysis unit, search request analysis unit 135 Search result combination unit, search result combination unit 140 Search execution unit, search execution unit 150 Document list storage unit, document list unit 155 Temporary transposed index storage means, temporal transposed index storage section 160 Indexing means, indexing section 165 Morphological analysis section 170 Duplicate document storage means, duplicate document section 175 Document collection means, document collection interface section 180 Shared document (file) server

Claims (8)

A document search method in a search device connected to a shared document server and searching for a document,
A document collection unit receives a document operation notification and document for creating, updating, or deleting a new document from the shared document server, duplicates the document, assigns a document identifier to each document, and stores it in a duplicate document storage unit. A document collection step to store;
An indexing unit acquires the document from the duplicate document storage unit, assigns the document identifier to each word of the document, stores the document identifier in the time-replaced index storage unit, and stores the document identifier and operation date An indexing step of storing information on the transition and a document list including the document in a document list storage unit;
A search request analyzing means for analyzing the input search request and extracting a search attribute and a search condition;
Search execution means searches for a word from the temporal transposed index storage means based on the search condition, obtains a document identifier (transposed index) corresponding to the word, and uses the document list storage means based on the transposed index. A search step for obtaining a search result document by searching;
When the search result combining means has a plurality of document identifiers of the document acquired in the search step, a search result combining step for deleting and outputting the old operation date and time document from the search result;
A document search method characterized by: In the search step,
When a date, time of creation, update, or deletion of a document is specified as a search condition corresponding to the search attribute, a document identifier matching the date is acquired from the document list storage unit, and the document identifier The document search method according to claim 1, wherein a document corresponding to is acquired. In the search step,
When a document name is specified as a search condition corresponding to the search attribute, a document identifier matching the document name is acquired from the document list storage unit, and a document corresponding to the document identifier is acquired. The document search method according to claim 1. A document search device connected to a shared document server and searching for a document,
Document collection means for receiving a document operation notification and document for creating, updating, or deleting a new document from the shared document server, copying the document, assigning a document identifier for each document, and storing the document in the duplicate document storage means When,
Obtaining the document from the duplicate document storage means, assigning the document identifier for each word of the document and storing it in the temporally-transposed index storage means, and information on the transition of the document including the document identifier and operation date and time, and Indexing means for storing a document list including the document in a document list storage means;
A search request analysis means for analyzing the input search request and extracting a search attribute and a search condition;
By searching for a word from the temporal transposed index storage unit based on the search condition, obtaining a document identifier (transposed index) corresponding to the word, and searching the document list storage unit based on the transposed index, A search execution means for obtaining a search result document;
When there are a plurality of document identifiers of the documents acquired by the search execution means, search result combining means for deleting and outputting a document with an old operation date and time from the search results;
A document search apparatus characterized by comprising: The search execution means includes
When a date, time of creation, update, or deletion of a document is specified as a search condition corresponding to the search attribute, a document identifier matching the date is acquired from the document list storage unit, and the document identifier The document search apparatus according to claim 4, further comprising means for acquiring a document corresponding to. The search execution means includes
When a document name is specified as a search condition corresponding to the search attribute, a document identifier matching the document name is acquired from the document list storage unit, and a document corresponding to the document identifier is acquired. 5. The document retrieval apparatus according to claim 4, further comprising means.   7. A document search program for causing a computer to function as each means constituting the document search apparatus according to claim 4.   A computer-readable recording medium storing the document search program according to claim 7.

Images