Calls to mark_processed do not use the solr/metrics/links field

[marblestation]

There are several calls to mark_processed without type here:

ADSMasterPipeline/adsmp/app.py

Line 643 in 5e057e4

self.mark_processed(failed_bibcodes, type=None, status='solr-failed')

which does not mark the right processed field in the database (e.g., solr, metrics or links processed fields). Is this the intended behavior?

[romanchyla]

if the type is empty, it will update 'processed' timestamp, otherwise it updates specific timetamps

the specific column timestamps are used for tracking, the main timestamp (the one that is used by the code to decide whether something was already tried) is the processed

so it makes sense if mark_processed is called with type=None from several places

it would make sense if it was called twice (or multiple times) for a bibcode, e.g.

1. type=solr
2. type=metrics
3. type=None

what would not make sense, if a particular bibcode was mark_processed several times with the same type set (i.e. the bibcode be present in multiple sets)