From d8cf02dbdc6848be4de2710e54f4463d702b6e7c Mon Sep 17 00:00:00 2001 From: Bryan Newbold Date: Fri, 22 May 2020 16:10:49 -0700 Subject: importers: clarify handling of ApiException One of these (in ingest importer pipeline) is an actual bug, the others are just changing the syntax to be more explicit/conservative. The ingest importer bug seems to have resulted in some bad file match imports; scale of impact is unknown. --- python/fatcat_tools/importers/common.py | 12 ++++++++---- python/fatcat_tools/importers/ingest.py | 1 + python/fatcat_tools/importers/pubmed.py | 1 + 3 files changed, 10 insertions(+), 4 deletions(-) (limited to 'python/fatcat_tools/importers') diff --git a/python/fatcat_tools/importers/common.py b/python/fatcat_tools/importers/common.py index 99c330a6..eafc6546 100644 --- a/python/fatcat_tools/importers/common.py +++ b/python/fatcat_tools/importers/common.py @@ -458,7 +458,8 @@ class EntityImporter: creator_id = rv.ident except ApiException as ae: # If anything other than a 404 (not found), something is wrong - assert ae.status == 404 + if ae.status != 404: + raise ae self._orcid_id_map[orcid] = creator_id # might be None return creator_id @@ -479,7 +480,8 @@ class EntityImporter: release_id = rv.ident except ApiException as ae: # If anything other than a 404 (not found), something is wrong - assert ae.status == 404 + if ae.status != 404: + raise ae self._doi_id_map[doi] = release_id # might be None return release_id @@ -495,7 +497,8 @@ class EntityImporter: release_id = rv.ident except ApiException as ae: # If anything other than a 404 (not found), something is wrong - assert ae.status == 404 + if ae.status != 404: + raise ae self._pmid_id_map[pmid] = release_id # might be None return release_id @@ -512,7 +515,8 @@ class EntityImporter: container_id = rv.ident except ApiException as ae: # If anything other than a 404 (not found), something is wrong - assert ae.status == 404 + if ae.status != 404: + raise ae self._issnl_id_map[issnl] = container_id # might be None return container_id diff --git a/python/fatcat_tools/importers/ingest.py b/python/fatcat_tools/importers/ingest.py index 7d5211fc..6cf1604b 100644 --- a/python/fatcat_tools/importers/ingest.py +++ b/python/fatcat_tools/importers/ingest.py @@ -108,6 +108,7 @@ class IngestFileResultImporter(EntityImporter): elif err.status == 400: self.counts['warn-extid-invalid'] += 1 continue + raise err release_ident = release.ident break if self.use_glutton_match and not release_ident and row.get('grobid'): diff --git a/python/fatcat_tools/importers/pubmed.py b/python/fatcat_tools/importers/pubmed.py index abcb21d9..3d3e3a8c 100644 --- a/python/fatcat_tools/importers/pubmed.py +++ b/python/fatcat_tools/importers/pubmed.py @@ -782,6 +782,7 @@ class PubmedImporter(EntityImporter): # NOTE: API behavior might change in the future? if "release_edit_editgroup_id_ident_id_key" in err.body: self.counts['skip-update-conflict'] += 1 + return False else: raise err finally: -- cgit v1.2.3