diff options
-rwxr-xr-x | python/fatcat_import.py | 2 | ||||
-rw-r--r-- | python/fatcat_tools/importers/shadow.py | 6 |
2 files changed, 7 insertions, 1 deletions
diff --git a/python/fatcat_import.py b/python/fatcat_import.py index 1f026edc..843685aa 100755 --- a/python/fatcat_import.py +++ b/python/fatcat_import.py @@ -482,7 +482,7 @@ def main(): help="create release and file entities based on GROBID PDF metadata extraction") sub_shadow_lib.set_defaults( func=run_shadow_lib, - auth_var="FATCAT_API_AUTH_TOKEN", + auth_var="FATCAT_AUTH_WORKER_SHADOW", ) sub_shadow_lib.add_argument('json_file', help="JSON file to import from (or stdin)", diff --git a/python/fatcat_tools/importers/shadow.py b/python/fatcat_tools/importers/shadow.py index cfe1b1cf..261cf888 100644 --- a/python/fatcat_tools/importers/shadow.py +++ b/python/fatcat_tools/importers/shadow.py @@ -130,6 +130,9 @@ class ShadowLibraryImporter(EntityImporter): if not existing: return True + if not existing.extra: + existing.extra = {} + if existing.extra.get('shadows') and list(fe.extra['shadows'].keys())[0] in existing.extra['shadows']: # already imported from this shadow library; skip self.counts['exists'] += 1 @@ -172,6 +175,9 @@ class ShadowLibraryImporter(EntityImporter): existing.sha1 = existing.sha1 or fe.sha1 existing.sha256 = existing.sha256 or fe.sha256 edit = self.api.update_file(self.get_editgroup_id(), existing.ident, existing) + # add sha1 to non-entity edit row, so we can do more aggressive + # group-level de-dupe + edit.sha1 = existing.sha1 self._edits_inflight.append(edit) self.counts['update'] += 1 return False |