diff options
| author | Bryan Newbold <bnewbold@robocracy.org> | 2020-01-15 14:03:50 -0800 |
|---|---|---|
| committer | Bryan Newbold <bnewbold@robocracy.org> | 2020-01-15 14:03:50 -0800 |
| commit | 234dfa11d4e552d4ef784d8c7b13bfdaf42c597c (patch) | |
| tree | 91e8ac4c9e60f54b0fc4480cbaf5a452d0797b77 | |
| parent | ec445720e768d3f0cc56c03052833b86d5068547 (diff) | |
| download | fatcat-234dfa11d4e552d4ef784d8c7b13bfdaf42c597c.tar.gz fatcat-234dfa11d4e552d4ef784d8c7b13bfdaf42c597c.zip | |
ingest: allow more sources to auto-import
| -rw-r--r-- | python/fatcat_tools/importers/ingest.py | 3 |
1 files changed, 2 insertions, 1 deletions
diff --git a/python/fatcat_tools/importers/ingest.py b/python/fatcat_tools/importers/ingest.py index 16643eb5..e53dcae5 100644 --- a/python/fatcat_tools/importers/ingest.py +++ b/python/fatcat_tools/importers/ingest.py @@ -29,6 +29,7 @@ class IngestFileResultImporter(EntityImporter): self.ingest_request_source_whitelist = [ 'fatcat-changelog', 'fatcat-ingest-container', + 'fatcat-ingest', 'arabesque', ] if kwargs.get('skip_source_whitelist', False): @@ -57,7 +58,7 @@ class IngestFileResultImporter(EntityImporter): self.counts['skip-ingest_request_source'] += 1 return False if source.startswith('arabesque'): - if row['reqeust'].get('link_source') not in ('arxiv', 'pmc'): + if row['reqeust'].get('link_source') not in ('arxiv', 'pmc', 'unpaywall', 'doi'): self.counts['skip-arabesque-source'] += 1 return False if source.startswith('savepapernow'): |