fix(extract-fields): Sometimes, abstract item is not a string ...

But can be a number.

fix(harvest-from-doi): Prevent Request Header Fields Too Large

refactor(Makefile): Add stage numbers on output data files

chore(data): Use doi-vol2.txt as default input

chore(data): Remove data extracts (-50) ...

And ignore big .gz files

docs: Find lines containing non-ASCII characters

perf(harvest-from-doi): Set query size to 90 (from 10)

fix(harvest-from-doi): Replace middle point with normal point ...

And
- replace en dash with normal hyphen-minus
- remove soft hyphen (discretionary hyphen)
- remove "http" part of a DOI
- remove "doi:" part of a DOI
- remove URL following a DOI
- remove duplicates of a DOI
- remove PMID following DOI
- remove part preceding a DOI

fix(harvest-from-doi): Remove BOM automatically

chore(data): Remove BOM from doi-vol3.txt

docs: Remove BOM from dois.txt

fix(data): Remove BOM from dois.txt

fix(harvest-from-doi): More cleaning of DOIs

chore(data): CRLF to LF (for doi-vol3.txt)

fix(harvest-from-doi): Clean up dois

chore(data): CRLF to LF end of lines

fix(data): Remove Doi header

fix(Makefile): typo

feat(harvest-from-doi): Make it work from dois.txt

chore(data): Add volume 3's DOIs

chore(data): Ignore corpus-simple-cnrs.json

perf(enrich-rnsr): Remove one level of expand ...

Making real batch of 100 elements, instead of one per document.

refactor(detect-cnrs): Remove two ezs statements

feat(detect-cnrs): Add the CNRS detecting ...

oui or non in ws.cnrs

chore(git): Ignore last full file (label-pascal)

docs(packages): Add packages to config

chore(data): Update sample data ...

Using last .ini scripts

feat(enrich-institutes): Deduplicates ws.etab

feat(enrich-institutes): Deduplicate RNSR ids

feat(extract-fields): Add doi field