diff options
| author | Bryan Newbold <bnewbold@archive.org> | 2020-04-02 17:36:41 -0700 |
|---|---|---|
| committer | Bryan Newbold <bnewbold@archive.org> | 2020-04-02 17:36:41 -0700 |
| commit | 7c5d51190cb8e58dea6231b48dc942fb35ef54a9 (patch) | |
| tree | 23267feb6eedfead299eba850a08f91a53412057 | |
| parent | cbfa0d9e1ed1dca6fa34f134cf81d4dd8a0a99f8 (diff) | |
| download | fatcat-covid19-7c5d51190cb8e58dea6231b48dc942fb35ef54a9.tar.gz fatcat-covid19-7c5d51190cb8e58dea6231b48dc942fb35ef54a9.zip | |
transform: include missing external idents
| -rwxr-xr-x | elastic_transform.py | 4 |
1 files changed, 4 insertions, 0 deletions
diff --git a/elastic_transform.py b/elastic_transform.py index 38c9ecf..587aabd 100755 --- a/elastic_transform.py +++ b/elastic_transform.py @@ -50,6 +50,8 @@ def fulltext_to_elasticsearch(row, force_bool=True): 'pages', 'number', 'license', + ] + EXT_IDS = [ 'doi', 'pmid', 'pmcid', @@ -61,6 +63,8 @@ def fulltext_to_elasticsearch(row, force_bool=True): ] for key in BIBLIO_KEYS: t[key] = release.get(key) or None + for key in EXT_IDS: + t[key] = release['ext_ids'].get(key) or None abstracts = [] abstract_langs = [] |