edrn / edrn.bmdb Goto Github PK

The RDF generation for organ-specific information for a biomarker doesn't include the study.

The FHCRC_ID column of the study data should be used to make a <bmdb:referencesStudy> predicate whose rdf:resource is http://edrn.nci.nih.gov/data/protocols/{PROTOCOL-ID}.

BMDB RDF says that for biomarker http://edrn.jpl.nasa.gov/bmdb/biomarkers/view/53 the AssociatedDataset predicate is https://edrn-labcas.jpl.nasa.gov/labcas-ui/c/index.html?collection_id=Barretts_Esophagus_Methylation_Profile_Dataset.

However, the actual subject URI is https://edrn-labcas.jpl.nasa.gov/labcas-ui/c/index.html?collection_id=Barrett's_Esophagus_Methylation_Profile_Dataset.

After a while, queries into the RDF service for the Focus Bio Marker Data Base (BMDB) fail and emit this in the logs:

focusbmdb-rdf    |   File "/app/src/edrn/bmdb/server.py", line 50, in __call__
focusbmdb-rdf    |     cursor.execute('SELECT COUNT(*) FROM `biomarkers`')
focusbmdb-rdf    |   File "/app/eggs/PyMySQL-0.10.0-py3.8.egg/pymysql/cursors.py", line 163, in execute
focusbmdb-rdf    |     result = self._query(query)
focusbmdb-rdf    |   File "/app/eggs/PyMySQL-0.10.0-py3.8.egg/pymysql/cursors.py", line 321, in _query
focusbmdb-rdf    |     conn.query(q)
focusbmdb-rdf    |   File "/app/eggs/PyMySQL-0.10.0-py3.8.egg/pymysql/connections.py", line 504, in query
focusbmdb-rdf    |     self._execute_command(COMMAND.COM_QUERY, sql)
focusbmdb-rdf    |   File "/app/eggs/PyMySQL-0.10.0-py3.8.egg/pymysql/connections.py", line 742, in _execute_command
focusbmdb-rdf    |     raise err.InterfaceError(0, '')
focusbmdb-rdf    | pymysql.err.InterfaceError: (0, '')

This discussion suggests that having a universal connection object is the culprit. We need to create one per thread.

The EDRN Portal needs to provide biomarker phase information at the organ→study level. To do so, it must get this information from the BMDB RDF.

See: EDRN/P5#140

This package is still using Python 2.7, which reached end of life on 2020-01-01.

Python 3 has been out 12 years. It's time to leave Python 2 behind.

Currently phases are of the form

http://edrn.nci.nih.gov/rdf/rdfs/bmdb-1.0.0#Phase = One

or

http://edrn.nci.nih.gov/rdf/rdfs/bmdb-1.0.0#Phase = Three

which is dumb for sorting purposes. But this is how ahart did it back in the day. We need to use actual numbers so we can sort properly, i.e.:

http://edrn.nci.nih.gov/rdf/rdfs/bmdb-1.0.0#Phase = 1

or

http://edrn.nci.nih.gov/rdf/rdfs/bmdb-1.0.0#Phase = 3