AutoClass migration is playing havoc with 'find latest files by edit date'
The more robust process would be to log files in Metamist, retaining only one active file
Contributes to solving #272
Proposed Changes
After completing the HTML generation process:
query for all 'web' files in the same dataset
find any which fit the same subtypes as the current analysis (genome/exome, familial/singleton)
If any active files are found, mark them as inactive (@illusional I remember an issue where either querying for active=True didn't really limit the search, or updating active to inactive wasn't persisted. Is this still an issue?)
Register the current output file
This change would be a predicate for shifting the report index generator to using metamist.
The downstream plan would be to find the most recently created metamist entry, not necessarily the only active one - the 'deactivating' behaviour would be a nice to have for tidiness but not essential.
Fixes
Proposed Changes
After completing the HTML generation process:
active=Truedidn't really limit the search, or updating active to inactive wasn't persisted. Is this still an issue?)This change would be a predicate for shifting the report index generator to using metamist.
The downstream plan would be to find the most recently created metamist entry, not necessarily the only active one - the 'deactivating' behaviour would be a nice to have for tidiness but not essential.
Checklist