Resolve discrepancies between initial run dates and embargo dates of SMILE vs. those of Hera

[qu8n]

See Quan's thread with Anne Marie in #smile-pm for the latest.

I spot checked a few of these samples and confirmed that their initial run dates are correct, but only based on the Cohort Complete data in the SMILE database.

The initial run dates from HERA in the join tables above are mostly dated year 2020, but the earliest Cohort Complete event date in the SMILE database is dated year 2022. Is it possible that SMILE is missing some Cohort Complete data?

Neo4j query to spot check initial run date of a sample primary ID:

MATCH (sm:SampleMetadata{primaryId:"06208_R_1"})<-[:HAS_METADATA]-(s:Sample)<-[:HAS_COHORT_SAMPLE]-(c:Cohort)-[:HAS_COHORT_COMPLETE]->(cc:CohortComplete)
RETURN MIN(cc.date)

Notes for discussion:

• Since initialRunDates are correct based on data CURRENTLY in SMILE, we will need to look into what data is missing from SMILE.
• To assist with this effort, we will need sample cohort data specifically for the samples that have discrepant initial run dates in TEMPO vs. SMILE
• If cohort data is missing in SMILE then we need to recover it and/or figure out why it isn't in SMILE

Resolution:

• Anne Marie provided a mapping of sample-initial run dates that SMILE can use to override the calculated run dates.
• for the samples that are providing in the mapping above, recalculate their embargo dates and access level values
• add logic in smile-server during import/update of cohort complete events. If the sample's calculated initial run date is after the initial run date currently stored in the db for that sample then leave the data as is (don't update)

[ao508]

Please add your planning poker estimate with Zenhub @qu8n