Multi-analysis specification with TreatmentPatternsModule

[SivamshIndukuri]

This PR updates the TreatmentPatterns Module to enable multiple analyses to be run simultaneously.

Summary of Changes

1. Added createModuleSpecification:
   This function builds a TreatmentPatterns analysis spec from lists of target, event, and exit cohort definitions, converting them into the cohort data frame the package expects, and returns an analysis settings list.

2. Added createMultiAnalysisModuleSpecification:
   This function takes a list of analysis specifications made by createAnalysisSpecification

3. Modified createModuleSpecification:
   I modified the return specification to be wrapped by a list so it matches the createMultiAnalysisModuleSpecification returned results. Now it returns a list in the form of list(tpAnalysisList = list())

4. Modified execute:
   I modified this function so it can loop through the tpAnalysisList and run each treatment pattern. All the logs of passed and failed pathways are written to the treatment_pathway_runs.csv in the work folder. The execute method also makes a new table in the results folder called analysis_cohorts, which has the columns:
   event_cohort_name, event_cohort_id, target_cohort_id, target_event_id, and analysis_id

5. Modified data model:
   Modified the model to allow the analysis_cohorts table to be inserted

eg.

tp <- TreatmentPatternsModule$new()

modSpecs <- list(
tp$createAnalysisSpecification(...),
tp$createAnalysisSpecification(...),
)

tpModuleSpecifications <- tp$createMultiAnalysisModuleSpecification(tpAnalysis = modSpecs)

[SivamshIndukuri]

The script now checks that the results folder exists, reads completed analysis IDs from any passed_pathway_runs.csv files, and then runs only the treatment patterns that have not yet been executed. The work folder is updated with two files: passed_pathway_runs.csv and failed_pathway_runs.csv.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 99.03%. Comparing base (6187132) to head (f6a97e7).
⚠️ Report is 1 commits behind head on develop.

Additional details and impacted files

@@             Coverage Diff             @@
##           develop     #273      +/-   ##
===========================================
+ Coverage    97.82%   99.03%   +1.20%     
===========================================
  Files           16       16              
  Lines         4193     4748     +555     
===========================================
+ Hits          4102     4702     +600     
+ Misses          91       46      -45     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[anthonysena]

Please remove the createAnalysisSpecification function - I think all that is needed here is the createModuleSpecification changes (that appear to be in createAnalysisSpecification) and the multi analysis function. Reach out with any questions or concerns.

[anthonysena]

@SivamshIndukuri a few additional comments based on a deeper review. Let me know if you have any questions or concerns. Thanks!

[anthonysena]

One more addition to the logic here: we should consider the value of the incremental flag of the executionSettings: https://ohdsi.github.io/Strategus/reference/createCdmExecutionSettings.html#arg-incremental. If executionSettings$incremental == FALSE, we should remove any info on prior runs (passed and failed) and re-run from scratch.

[SivamshIndukuri]

Adding executionSettings$incremental to the 'if condition' should resolve this. It will act like 0 pathways passed and resultAppend will be false.

[anthonysena]

cohortAnalysisTable provides the combination of target cohorts, event cohorts, exit cohorts and the current analysisId. cohorts already contains the target/event/exit cohorts, so would it be easiest to construct cohortAnalysisTable by appending the current analysisId? Also, we aim to use camelCase in R code per OHDSI code guidelines here: https://ohdsi.github.io/Hades/codeStyle.html#Interfacing_between_R_and_SQL. If you need to move between snake & camelCase, DatabaseConnector has some utility functions described in that link.

[SivamshIndukuri]

Yes that would be the easiest. I originally went with this to make it easier when making the queries for OhdsiReportGenerator but it should still work in this format.