A collection or python parsers, validators, and serializers for adsabs ingest pipeline. For the time being all we have is simple stuff used to ingest data into ADS Classic, but this is meant to grow and change in due time as our metadata models become more sophisticated.
We are now starting to implement parsers that can be used to ingest directly into Solr (e.g. dubcore/arxiv). See the repository ADS_Direct_Import for an example of this application.
There are several different classes of parser that can be used for different content sources (JATS, Proceedings of Science website, Zenodo, etc.) The basic syntax for calling a parser is simple: create an instance of the parser, and obtain the parsed output using the "parse" method of the parser on the required input.
For example, the APS parser requires a file pointer to an APS-JATS format fulltext.xml file (see the subdirectories under /proj/ads/fulltext/sources/downloads/cache/APS). Such a file could be parsed with:
from pyingest.parsers.aps import APSJATSParser
infile = "data_directory/apsjats-doi-fulltext.xml"
parser = APSJATSParser()
with open(infile,'rU') as fp:
output = parser.parse(fp)
All parsers are designed to give output in a format that can be serialized by pyingest.serializers.classic.Tagged() to create an output file in ADS tagged format. Most parsers will return a single document; the exception is procsci.PoSParser, which returns a list of documents (which collectively make a TOC-formatted entry when passed to the classic serializer element by element).
Alberto Accomazzi, Director, NASA ADS
Matthew Templeton, NASA ADS