Improve metadata extraction, especially for PDFs

[anjackson]

We need to check we're doing a good enough job with what we have, and we should look at exploiting additional tools in order to improve metadata extraction from PDFs.

• GROBID (or Grobid) means GeneRation Of BIbliographic Data. (Java+JNI?)
• pdfextract from CrossRef (ruby)
• PDF Structure and Syntactic Analysis for Metadata Extraction and Tagging (python)
• Or even using the Mendeley API

[anjackson]

More on GROBiD

• https://github.com/kermitt2/grobid-example/blob/master/src/main/java/org/grobidExample/ExampleBibTex.java
• Integrated under Tika:
  • https://t.co/DolGYP7BMS
  • https://wiki.apache.org/tika/GrobidJournalParser