Update the way ReCiter handles books

[paulalbert1]

Scope

Approximately 0.1% of records in PubMed are for books although this has increased in the past year.

Data model

Books have a different data model. Key differences include:

Description	Book XML Attribute	Journal Article XML Attribute
Publication Type	<PublicationType>Book [Chapter]</PublicationType>	<PublicationType>Journal Article</PublicationType>
Source Title	<BookTitle>	<JournalTitle>
Identifier (ISBN/ISSN)	<ISBN>	<ISSN>
Publisher	<Publisher>	N/A
Publication Place	<PlaceOfPublication>	N/A
Authors	<AuthorList><Author>...</Author></AuthorList>	same
Editors (for books)	<EditorList><Editor>...</Editor></EditorList>	N/A
Pagination	<PageRange> (especially for chapters)	<MedlinePgn>
Publication Frequency	N/A	Could be inferred from <JournalIssue><PubFrequency>
DOI	<ELocationID EIdType="doi">...</ELocationID>	same
Abstract	<AbstractText>...</AbstractText> (sometimes omitted)	same

Effect

The inconsistent data model causes chaos. For example, for personIdentifier = tme2002 and PMID = 34818336 (see also API), the wrong authors are listed. What probably is occurring is that the author list if shifting by one.

Another example: mtoth and 21204454: https://eutils.ncbi.nlm.nih.gov/entrez/eutils/efetch.fcgi?db=pubmed&id=21204454&retmode=xml

Options

1. Update ReCiter PubMed Retrieval Tool not to return books. We could do this like so: cole c[au] NOT (booksdocs[Filter])

2. Update data model across the projects to handle books:

   • ReCiter PubMed Retrieval Tool
   • ReCiter
   • ReCiterDB
   • ReCiter Publication Manager

3. Exclude books from ReCiter Feature Generator and Article Retrieval output.

   • Approach 1: Exclude cases where PublicationType = Book [Chapter], or
   • Approach 2: Require JournalTitle attribute
   • Include a flag in application.properties to exclude books

[paulalbert1]

I'm not sure this is still an issue.