Publishing Medline Database using Greenstone Digital Library Software

1
Publishing Medline Database using Greenstone
Digital Library Software

• Major Project Presentation by
• Doncy Augustine
• Trainee (2001-2002)
• National Centre for Science Information
• Indian Institute of Science

2
Contents to be covered

• Introduction
• Need for the project
• Scope
• Medline, XML, GSDL Refer
• System Design
• Implementation
• Demo
• Observations
• Conclusion

3
Introduction

• Demand for Digital Libraries.
• New way of organizing information using GSDL.
• Checking the capability of GSDL to index
  documents at record level.

4
Objective

• To test the feasibility of using the Greenstone
  Digital Library Software (GSDL) for publishing
  bibliographic databases like Medline, for online
  access and CD-ROM distribution.

5
Need for the Project

• Economy
• Medline
• GSDL
• Publish and disseminate information for
  developing countries at a lower cost.

6
Scope of the Project

• Medline bibliographic database records are
  available in XML format under freely license.
• 2000 bibliographic records were downloaded from
  Medline.

7
Medline

• Medline
• National Library of Medicine
• Collection of 11 million references from 4300
  biomedical journals
• Print equivalents
• Index Medicus
• Index to Dental Literature
• International Nursing Index
• PubMed
• Available in XML tagged format

8
eXtensible Markup Language(XML)

• Enables data exchange
• Subset of SGML
• Describing data
• Syntax Declaration (lt?xml version"1.0"?gt) ,
  Elements Attributes
• Valid with a DTD

9
GSDL

• Open Source
• Developed by Computer Science Dept., University
  of Waikato,
• New Zealand.
• File level indexing and searching
• Plugins- ReferPlug
• Publishing Collection on CD-ROM

10
Refer Format

• For bibliographic documents
• .bib extension
• Single character tags preceded with sign
• Eg A Seagull TA
• Records separated by single line

11
System Design
Mapping Algorithm
Working with Refer
Understanding XML DTD
Publishing Collection on CD-ROM
Conversion XML to Refer
Building Collection
12
Working with Refer

• Checking GSDLs potential to access bibliographic
  records.
• Sample file downloaded from Internet (122
  records, 195 Kb)
• - ReferPlug
• ReferPlug ? RefPlug

13
Sample XML records

• lt!DOCTYPE MedlineCitationSet (View Source for
  full doctype...)gt
• ltMedlineCitationSetgt
• ltMedlineCitation CitationOwner"NLM"gt
•   ltMedlineIDgt99201558lt/MedlineIDgt
•   ltPMIDgt10101342lt/PMIDgt
• ltDateCreatedgt
• ltYeargt1999lt/Yeargt
•   ltMonthgt04lt/Monthgt
•   ltDaygt15lt/Daygt
•   lt/DateCreatedgt ..
• lt/MedlineCitationSetgt

14
DTD

• NLM Medline DTD
• Medline Citation DTD
• NLM Common DTD

15
DTD

• lt!-- This is the top level element for
  MedlineCitation --gt
• lt!ELEMENT MedlineCitation (MedlineID.Ref,
• PMID.Ref,
• DateCreated.Ref,
• DateCompleted?,
• DateRevised?,
• Article,
• MedlineJournalInfo,
• ChemicalList?,
• CitationSubset,
• MeshHeadingList?,
• SpaceFlightMission,
• InvestigatorList?,
• GeneralNote)gt

16
Mapping Algorithm
17
XML to Refer- Perl Program

• Input in XML format
• Check for Medline Fields
• Identification of necessary fields
• Extraction of values
• Concatenation to refer tags
• Variable Assignment
• Print the output (Refer.bib)

18
Building Collection

• Collection Information
• Source Data
• Configuring the Collection
• Building the Collection
• Export into CD-ROM

19
DEMO
20
Observations

• Suitable for Publishing Databases
• Non English characters
• Lack of corresponding Refer tags
• Eg Language

21
Conclusion

• Field Based indexing and Searching
• Worth Trying for processing bibliographic records
  in XML format directly

22
Questions?Suggestions?
23
Thank You