EP1907950A1

EP1907950A1 - Method and tool for searching in several data sources for a selected community of users

Info

Publication number: EP1907950A1
Application number: EP06778775A
Authority: EP
Inventors: Marina Riou; Lionel Dejean; Julien Manet
Original assignee: Airbus SAS
Current assignee: Airbus SAS
Priority date: 2005-07-11
Filing date: 2006-07-05
Publication date: 2008-04-09
Also published as: FR2888358B1; FR2888358A1; CN101218585B; JP2009500760A; BRPI0615495A2; WO2007006913A1; RU2433468C2; CA2614713A1; US20080270385A1; RU2008104915A; CN101218585A; US8463763B2

Abstract

The invention concerns a search method including a mode for defining at least one reading table (V) of contents of documents specific to a selected community of users (6) and comprising at least one category (T) of contents of documents, based on the structure of the metadata (M) of each data source (2) and a mode for using such a specific reading table (V) thus predefined as search filter and/or display format in a search for contents of documents conducted in the plurality of data sources (12), on behalf of said community of users.

Description

Method and tool for searching multiple data sources for a selected user community

The present invention relates to searching information in multiple data sources for a selected user community. More particularly, it relates to searching heterogeneous data sources that do not have a common data structure for said community of users.

It finds a general application in the field of research and knowledge management, for the benefit of a community of users sharing a profession, a center of interest or expertise and wishing to look for information in several sources of data.

We already know search tools by category of documents and / or keywords. Generally, categories are predefined manually or built on the fly during the search.

The relevance of the results is relatively insufficient, insofar as the categories used are usually too general and irrelevant, particularly for a community of users with specialized expertise.

The present invention overcomes this disadvantage. It aims to provide a search tool using pre-established reading grids specific to one or more communities of selected users. It thus offers, for a given community of users, help and assistance in finding information relevant to their profession or interests. It also aims to optimize the display of results, to facilitate reading according to specific and pre-established reading grids.

Finally, it aims to present the documents in a more satisfactory ranking for the community of users and improving the search time as well as the reading time of said results compared to the search tools of the prior art.

Thus, the present invention relates to a search method for searching in a plurality of data sources, each having structured metadata according to a chosen structure, information contents intended for at least one selected user community.

According to a general definition of the invention, the method comprises a mode of definition of at least one reading grid specific to a community of users and comprising at least one category of results, a function of the structure of the metadata of each source of data. data and a mode of using the specific reading grid thus predefined as a search filter and / or display format in a search conducted in the plurality of data sources for said community of users.

Such a method has the advantage of making it possible to quickly find relevant documents in a large and heterogeneous mass of information. This results in a saving of research and reading time, as well as an improvement of the relevance of the results.

In practice, the reading grid comprises a plurality of categories arranged according to a chosen hierarchical relation.

According to one embodiment, the definition mode of the reading grid comprises the following steps: a) defining at least one community of users according to at least one element shared by said users of said community; b) for the community thus defined, defining at least one data source, according to at least one parameter relating to said data source and / or said community; c) for the data source thus defined, searching for at least one metadata for defining at least one category; (d) in the case of a metadata positive search, define at least one rule of correspondence between the category and the metadata thus obtained; e) applying the correspondence rule thus defined to at least a part of the information contents of the data source and associating with said category the contents of documents verifying said correspondence rule; and f) repeating steps a) to e) for at least some metadata of each data source and obtaining at least one category, depending on the metadata structure of each data source and forming a predefined and specific reading grid. community of users for said plurality of data sources.

In practice, data sources are heterogeneous. For example, data sources belong to the group consisting of internal databases, external databases, internal knowledge management applications, document management applications, supplier content.

In one embodiment, the common element of a community of users belongs to the group formed by a center of interest, a project, a tool, a database, a method, a method, a job, an expertise.

According to another aspect of the invention, the mode of defining the preset and specific reading grids is implemented through a single access module placed between the chosen user community and said plurality of data sources.

According to yet another aspect of the invention, the parameter relating to the data source derives from the business analysis of the user community.

The present invention also relates to a search tool comprising a search server.

According to another aspect of the invention, the search server comprises: an access module for receiving requests from a community of users and able to transmit results to said users of the community;

a plurality of data sources each having metadata structured according to a chosen structure and each connected to the access module;

a module for preliminary definition of at least one reading grid specific to the community of users and comprising at least one category, a function of the metadata structure of each data source; and

a search module able to use said specific reading grid thus predefined as a search filter and / or display format in a search conducted in the plurality of data sources for the benefit of the user community. In practice, the definition module includes, for at least one community of users:

means for searching for at least one data source, at least one metadata of the source making it possible to define at least one category; suitable processing means, in case of positive search with regard to the metadata:

Defining at least one rule for correspondence between the category and the metadata thus obtained and applying the correspondence rule thus defined to at least a part of the results of the data source and associating the results verifying said correspondence rule with said category ; and

. repeating the processing for at least some metadata of each data source to obtain a specific reading grid applicable to the plurality of data sources and comprising at least one category, a function of the metadata structure of said data sources.

The present invention also relates to an information carrier readable by a computer system, possibly removable, totally or partially, in particular a CD-ROM or magnetic medium, such as a hard disk or a diskette, or a transmissible medium, such as an electrical or optical signal, comprising instructions of a computer program making it possible to implement the search method mentioned above, when this program is loaded and executed by a computer system.

Finally, a subject of the present invention is a computer program stored on an information medium, said program comprising instructions for implementing the above-mentioned search method, when this program is loaded and executed by a computer system. .

Other features and advantages of the invention will emerge in the light of the detailed description below and the drawings in which:

FIG. 1 schematically represents an architecture of the installation making it possible to implement the search method according to the invention;

FIG. 2 schematically illustrates the essential functions of the search method according to the invention;

FIG. 3 is a flowchart illustrating the steps of the module for defining specific and pre-established reading gates according to the invention; and FIGS. 4A and 4B show an example of a reading grid in accordance with the invention and illustrate the use of such a grid as search criteria and display format in a search carried out in a data source. chosen for the benefit of a community of users chosen according to the invention. With reference to FIG. 1, the search method comprises a search server 2 comprising several modules. An access module 3 makes it possible to receive requests 4 originating from users of a community 6. The access module 3 transmits results 8 to the users of said community via computer tools of the microcomputer-like type 10 of conventional technology. .

The access module 3 is interposed between the microcomputers 10 of the user communities 6 and a plurality of data sources 12 each having metadata structured according to a chosen structure and each connected to the access module 3.

The data sources 12 may be contents of 12A providers, 12B knowledge management applications, 12C document management applications, 12D knowledge sources, 12E internal business databases.

As indicated above, the data sources 12 are heterogeneous here, i.e. they do not necessarily have a common data structure. For example, the data sources are of type "Oracle",

DB2, Lotus, Verity, Filnet, Windchill, Plumtree, SAP.

As will be seen in more detail below, the search tool makes it possible to search in the plurality of data sources 12 or in a subset of one of the databases 12.

As will be seen in more detail below, the search tool according to the invention constitutes a suitable access tool for a selected community of users (step E1, FIG. 2).

The search tool also provides a search wizard in addition to keywords by providing pre-defined and community-specific reading grids that can be used as filters in category and keyword research (step E2 , Figure 2).

The search tool also makes it possible to search for documents in selected databases belonging to heterogeneous sources (step E3, FIG. 2).

The search tool also makes it possible, thanks to the reading grids, to provide consolidated search results of significant relevance (step E4, FIG. 2).

Finally, thanks to pre-established and specific reading grids, the results are displayed in an organized manner with improved visualization thanks to a classification using said reading grids (step E5, FIG. 2). With reference to FIG. 3, the search according to the invention uses a mode of defining grids for reading document contents specific to a chosen community of users.

As will be seen in more detail below, this reading grid includes one or more categories of document contents arranged according to a chosen hierarchical relationship, defined according to the metadata structure of each data source in which the search is conducted. .

As seen above, the search tool will use these specific and predefined search grids in addition to keywords as search filters and / or display formats in a category-based document content search. keywords in the plurality of data sources.

With reference to FIG. 3, the method first defines a community of users (step E10). This definition of a community is based on a focus, a project, a tool, a database, a method, a method, a profession, an expertise, or a criterion of belonging common to all community users.

Thus, the community is defined as a group of people sharing interests, projects, using databases, tools, methods, or processes.

For a community thus defined at the end of step E10, it is intended to identify or define at least one data source (step E12). This identification is performed according to a parameter relating to said source and / or to the community, for example following a business analysis (step E11).

From the data source 12 thus identified, it is possible to deduce the corresponding database, step E13.

Depending on the centers of interest of the community, it is possible to deduce some already existing potential reading grids V (step E 14), still called points of view or viewpoint.

Reading grids or viewpoints V are composed of categories T classified hierarchically according to a tree or a tree structure deduced according to the metadata structure M from databases and / or data sources 12.

The method thus provides a list or directory of databases 12 potentially usable or already used by the community. Following the step E14, the method deduces the categories T belonging to the potential reading gates V arising from the interests of the community of users (step E15).

For the data source thus defined (step E16), it is planned to search (step E17) at least one metadata M for defining at least one category T of document contents.

In the case of a positive search with regard to the metadata (positive test, step E18), at least one correspondence rule is defined between the category T and the metadata M thus obtained.

Thanks to the knowledge of the metadata structure of the databases (table, structure), the method identifies the usable fields.

The method here consists in identifying the metadata M that may correspond to one or more categories T. If the metadata M is not available or is not reliable (negative test, step E19), other metadata M are identified for deduce the corresponding metadata using matching rules E.

These correspondence rules E can be of simple or trivial type (step E21), in which the field is used which is correctly valued.

The correspondence rule E may be more complex (step E22), in which the correspondence rule E is deduced using other fields or other metadata M.

Thus, if the metadata M exists, the value of the field or a part of this value is used according to a simple or complex correspondence rule E.

As will be seen in more detail below, these matching rules E are constructed using a management tool called query builder. After validation of the matching rules E by the community

C (optional step E23), one obtains a set of categories T called reading grids V, points of view or viewpoints, arranged relative to each other according to a chosen hierarchical relation, deduced according to the structure of the metadata.

This operation is repeated for each category T, step E24, in each database 12, step E25.

After the exploration of all the databases 12, the reading grid V is validated, step E26, for the community of users 6. It may happen that, during the step E20, no reading grid

V is defined for the current database 12 (step E27).

Each database 12 is thus reviewed to identify the fields M making it possible to define each category T. For this one uses the knowledge of the structure of the databases with, if necessary, the help of the business or IT representative of the database. this base.

The model of the data processed by the method according to the invention is of the type b 'relational data base. For example, with reference to FIG. 1, a community of users 6 has several data sources 12 on which are connected adapters (not shown) specific to each base 12. Each community of users 6 has reading grids. V or viewpoints that serve as search filters and / or grouping criteria on the display. Equations E of correspondence and matching rules are defined and generated by a management tool called query builder. User requests 4 will therefore be used as a display or filter. The queries or queries are therefore processed according to equations established in the base 12.

A portion of the database 12 may be used. This portion is called still report.

The E equations still called mapping rules are the trivial or complex correspondence rule sets that define the categories T. As seen above, a reading grid V is a set of T categories with a tree structure. These equations make it easier to maintenance in case of evolution or change of structure of the databases.

In practice, the matching rules are calculated according to the data relating to the community of users 6, data relating to the database 12, data relating to the reading grid V and, where appropriate, data relating to to the subset of the database.

Each category T belonging to a reading grid V is defined (FIG. 3) using one or more metadata M, according to conditions to be verified, and operators. Operators can be Boolean operators AND, OR, "equal" or "not equal".

With reference to FIGS. 4A and 4B, an example of reading grids is illustrated and the use of such a grid as a search and display format criterion in a search conducted in a data source selected according to FIG. 'invention. With reference to FIG. 4A, there is illustrated an example of a search using a reading grid V (motors) defined according to the invention with reference to FIG.

First, an MC keyword is defined. Here the keyword MC is "fuel flow". The user then chooses a reading grid, here viewpoint V equals "engines". This reading grid V here comprises two main categories "Rolls Royce" and "CFMI". The "Rolls Royce" category includes subcategories "Trent 500 - A340", "Trent 700 - A330" and "Trent 900 - A380".

The community of users 6 is here the community called "fadec" and the data sources 12 are selected from the databases "RTN", "FTA", "fadecDB" and "ISAIM".

The community user here selects the category T = Trent 500 - A340.

The user query here comprises the keyword MC, the reading grid V associated with its correspondence rule E which will point to the database 12 selected by itself, here the database 12 = RTN. By virtue of the correspondence rule E which defines a part of the reading grid V, the RTN documents and the fields M1 = ATA and M2 = AC_model will be used.

The search is conducted for the community 6 = fadec in the database 12 = RTN with the reading grid V = engines in the category T = Trent 500 - A340.

The rule of correspondence allows to associate according to the equations established by the query builder all the documents answering the condition M1 =

ATA starting with 7 or equal to 80, as well as the fields M2 = AC_model equal to 340-541 or ACjnodel equal 340-642. This equation is implemented in XML text format in the search server.

With reference to FIG. 4B, the database 12 (RTN) contains several documents, in particular the six documents D, individualized in D1 to D6. The reading gate V thus makes it possible to achieve a search criterion with the type V = motors and the value T = Trent 500 - A340. It is known that this value T = Trent 500 - A340 corresponds to the metadata M1 = ATA beginning with 7 or M1 = ATA equal to 80, or M2 = AC_model starting with 340-541 or M2 = ACjnodel equal to 340-642.

The result of the search shows four documents. It can be seen that the document D3 is deleted insofar as the key word "fuel flow" is not present in the document, as well as the document D1 where the field M1 (ATA) does not start with 7 or is not equal to 80, but here equal to 24.

Of course, the reading grids V according to the invention can be used for several communities of users. Similarly, a user can combine multiple grids as a search criterion.

As regards the display of the results, the reading grids in accordance with the invention also offer the user a display format deriving from said reading grids. For example, with reference to FIG. 4B, the results can be classified according to the reading grids "function," ATA "," motors ". This results in a ranking of the results that is more relevant and specific to the user community.

Claims

A search method for searching in a plurality of data sources (12), each having metadata (M) structured according to a selected structure, information contents for at least one selected user community (6). ), characterized in that it comprises a mode of definition of at least one reading grid (V) specific to said community of users (6) and comprising at least one category (T) of results, function of the structure metadata (M) of each data source (12) and a mode of use of said specific reading grid (V) thus predefined as a search filter and / or display format in a search for content of documents conducted in the plurality of data sources (12), for said community of users (6).

2. Method according to claim 1, wherein the reading gate (V) comprises a plurality of categories (T) arranged in a chosen hierarchical relationship.

The method according to claim 1 or claim 2, wherein said definition mode of the read pattern (V) comprises the following steps: a) defining at least one user community (6) according to at least one shared item by said users of said community; b) for the community thus defined, defining at least one data source (12) according to at least one parameter relating to said source and / or the community; c) for the data source (12) thus defined, searching for at least one metadata (M) for defining at least one category (T) of document contents; d) in the case of a positive search for the metadata (M), defining at least one correspondence rule (E) between the category (T) and the metadata (M) thus obtained; e) applying the correspondence rule (E) thus defined to at least a part of the information contents of the data source (12) and associating with said category (T) the contents of documents verifying said correspondence rule (E) ; f) repeating steps a) to e) for at least some metadata (M) of each data source and obtaining at least one category (T), depending on the metadata structure of each data source and forming a reading grid (V) specific to the user community (6) for said plurality of data sources (12).

The method of claim 1, wherein the data sources (12) are heterogeneous.

The method of claim 4, wherein the data sources (12) belong to the group formed by internal databases (12E), internal knowledge management applications (12B), document management applications (12C). ), contents of external suppliers (12A), sources of internal knowledge (12D).

The method of claim 1, wherein the common element of a community of users (6) belongs to the group formed by a center of interest, a project, a tool, a database, a method, a method, a profession, an expertise.

The method according to claim 1, wherein the mode of defining the specific reading grids (V) is set up through a single access model (3) placed between the community of users (6) and the plurality of data sources (12).

The method of claim 1, wherein the parameter relating to the data source is derived from the business analysis of the user community. •

9. A search tool comprising a search server comprising:

an access module (3) for receiving requests from users belonging to a community (6) and capable of transmitting results to said users of the community; a plurality of data sources (12) each having metadata (M) structured according to a chosen structure and each connected to the access module (3);

a module for preliminary definition of at least one specific reading grid (V) and comprising at least one category, a function of the structure of the metadata (M) of each data source (12); and

a search module able to use said specific predefined reading grid (V) as a search filter and / or a result display criterion in a search carried out in the plurality of data sources, for the benefit of the community of users.

The search tool according to claim 9, wherein the definition module comprises:

means for searching for at least one data source (12), at least one metadata (M) of the source making it possible to define at least one category;

suitable processing means, in case of positive search with regard to the metadata:

. defining at least one correspondence rule (E) between the category (T) and the metadata (M) thus obtained and applying the correspondence rule (E) thus defined to at least a part of the results of the data source ( 12) and associating the results verifying said matching rule (E) with said category (T); and

. repeating the processing for at least some metadata (M) of each data source (12) to obtain a specific search and / or display grid (V) applicable to the plurality of data sources

(12) and comprising at least one category (T) according to the structure of the metadata of said data sources.

11. Information medium readable by a computer system, possibly removable, totally or partially, in particular a CD-ROM or magnetic medium, such as a hard disk or a diskette, or a transmissible medium, such as an electrical or optical signal, including instructions for a computer program for carrying out the search method according to one of claims 1 to 8, when the program is loaded and executed by a computer system.

12. Computer program stored on an information medium, said program comprising instructions for implementing the search method according to one of claims 1 to 8, when the program is loaded and executed by a computer system.