Temporal annual distribution functionality

[alex-odysseus]

Supporting OHDSI/WebAPI#2331 to enable temporal annual distribution extending the existing temporal functionality

Temporal annual distribution will be effective for Feature Analyses specified in newly added PrespecTemporalAnnualAnalysis.csv

The idea of the functionality is to extract an underlying year for a Covariate of interest so that it can be easily grouped by in Cohort Characterization result visualization in WebAPI / ATLAS

Minor refactoring using Stream and StreamSupport classes

Standard Java formatting applied in a few places (hopefully acceptable)

Referencing SqlRender 1.18.1 in pom.xml

Targeting feature for 3.7.0 (for an unknown reason there is still 3.5.1-SNAPSHOT in 'develop' pom.xml even though 3.6.0 has been officially released

[alex-odysseus]

@anthonysena Anthony, I can't assign reviewers, please do so

[anthonysena]

@alex-odysseus - I've provided some feedback on this PR for your review. I have not reviewed the other associated PRs so I'm unsure of the impact this proposed change would have on this feature. That said, here is my feedback:

1. We need to make sure any enhancements to FeatureExtraction are done so that users from the R side can access this functionality. As it currently stands, I don't see a way that an R user could invoke this functionality.
2. I think we can simplify the way in which this feature is invoked by simply passing in an option to perform the annual distribution as noted in the comments.
3. All R command checks and unit tests must pass for this to be considered in a state for merging into the develop branch.

Thanks!

[anthonysena]

The temporal_annual appears to be the way in which a user would indicate to FeatureExtraction to compute the annual distribution yet this parameter is never exposed to the user best I can tell. In reviewing the code, the approach appears to be: specify the analysis IDs in the PrespecTemporalAnnualAnalysis.csv and then the Java layer will then inject the temporal_annual parameter value.

[anthonysena]

@alex-odysseus thanks for reviewing this with me on Atlas/WebAPI WG call. Here are some specific tasks from our discussion:

• Remove inst/csv/PrespecTemporalAnnualAnalysis.csv as it is redundant with the analyses already defined in the package.
• Add an annualPrevalence parameter to the getDbCovariateData function to expose the ability to compute the annual prevalence. Make sure this is passed to the Java layer so it can properly pass this to the SQL layer...
• Tidy up the Java class based on the feedback above. For the createQuerySql function of that class, expose an annualPrevalence parameter to enable passing in the parameter from R and the SkeletonCohortCharacterization layers respectively.