summaryrefslogtreecommitdiffstats
path: root/python/fatcat_tools/workers/changelog.py
diff options
context:
space:
mode:
Diffstat (limited to 'python/fatcat_tools/workers/changelog.py')
-rw-r--r--python/fatcat_tools/workers/changelog.py52
1 files changed, 50 insertions, 2 deletions
diff --git a/python/fatcat_tools/workers/changelog.py b/python/fatcat_tools/workers/changelog.py
index c8584ccf..b84d5e70 100644
--- a/python/fatcat_tools/workers/changelog.py
+++ b/python/fatcat_tools/workers/changelog.py
@@ -102,6 +102,12 @@ class EntityUpdatesWorker(FatcatWorker):
# ccdc.cam.ac.uk: crystal structures
"10.5517/",
]
+ self.live_pdf_ingest_doi_prefix_acceptlist = [
+ # biorxiv and medrxiv
+ "10.1101/",
+ # researchgate
+ "10.13140/",
+ ]
def want_live_ingest(self, release, ingest_request):
"""
@@ -115,13 +121,55 @@ class EntityUpdatesWorker(FatcatWorker):
link_source = ingest_request.get('ingest_request')
ingest_type = ingest_request.get('ingest_type')
+ doi = ingest_request.get('ext_ids', {}).get('doi')
+
+ is_document = release.release_type in (
+ 'article-journal',
+ 'paper-conference',
+ 'article',
+ 'report',
+ 'chapter',
+ 'manuscript',
+ 'review',
+ 'thesis',
+ 'letter',
+ 'editorial',
+ 'abstract',
+ 'entry',
+ 'patent',
+ 'post',
+ 'review-book',
+ )
+ is_not_pdf = release.release_type in (
+ 'dataset',
+ 'stub',
+ 'software',
+ 'figure',
+ 'graphic',
+ )
+
+ # accept list sets a default "crawl it" despite OA metadata for
+ # known-OA DOI prefixes
+ in_acceptlist = False
+ if doi:
+ for prefix in self.live_pdf_ingest_doi_prefix_acceptlist:
+ if doi.startswith(prefix):
+ in_acceptlist = True
if self.ingest_oa_only and link_source not in ('arxiv', 'pmc'):
es = release_to_elasticsearch(release)
- if not es['is_oa']:
+ # most datacite documents are in IRs and should be crawled
+ is_datacite_doc = False
+ if release.extra and ('datacite' in release.extra) and is_document:
+ is_datacite_doc = True
+ if not (es['is_oa'] or in_acceptlist or is_datacite_doc):
return False
- doi = ingest_request.get('ext_ids', {}).get('doi')
+ # if ingest_type is pdf but release_type is almost certainly not a PDF,
+ # skip it. This is mostly a datacite thing.
+ if ingest_type == "pdf" and is_not_pdf:
+ return False
+
if ingest_type == "pdf" and doi:
for prefix in self.ingest_pdf_doi_prefix_blocklist:
if doi.startswith(prefix):