# TreatmentPatterns 3.1.2

• Fixes run-to-run stability of TreatmentPatterns.
• Fixes issue for andromeda >= 1.0.0
• Fixed issue where doubles casted to numeric were reported in scientific notation.
• Fixed issue when parallelizing multiple computePathways(), caused by database-side conflicts.
• Fixed issue where the same subject_id was assigned a different pseudo-subject_id when using multiple cohort tables.
• Fixed issue when multiple target cohorts and multiple cohort tables were specified.
• Added export of problematic records when combination check fails.
• Deprecated SQL implementation for fetching cohort information.
• Internal change to always use CDMConnector internally (either using ODBC or JDBC).
• Require CDMConnector >= 2.0.0.
• Gracefully exit when there is no internet connection when running internal helper functions for testing.

# TreatmentPatterns 3.1.1

• Fixed issue when setting filterTreatments = "Changes" and made it ambiguous regardless of the Andromeda version.
• Fixed issue where combination treatments were ordered differently when using Andromeda >= 1.0.0.
• Fixed compatibility with Spark when using CDMConnector with ODBC.
• Updated the way data is copied from the cdm_reference to andromeda.

# TreatmentPatterns 3.1.0

• Added exportPatientLevel() function to export patient level data.
• Added startAnchor parameter to dictate the window of interest.
• Added windowStart parameter to dictate the window of interest.
• Added endAnchor parameter to dictate the window of interest.
• Added windowEnd parameter to dictate the window of interest.
• Added concatTargets parameter to specify if multiple target cohorts should be treated separately, or continuously.
• Added overlapMethod parameter to pick a method how to deal with non-combination overlap between events.
• Removed indexOffset parameter.
• Removed includeTreatments parameter.
• Disabled failing test on CRAN environment
• Significantly reduced tests that run on CRAN.

# TreatmentPatterns 3.0.3

• Added compatibility with Andromeda 0.6.7 and 1.0.0

# TreatmentPatterns 3.0.2

• Added unit-testing for CDMInterface on Postgres, Redshift, SQL Server, Oracle, BigQuery, Spark, Snowflake
• Fixed some SQL issues for specific databases
• Fixed issue where a crash occurred on R 4.2.x, when saving the input arguments.
• Uncapped pathway length from 5.
• Fixed edge case on long repeating events to collapse.

# TreatmentPatterns 3.0.1

• Deprecated shiny app
• Removed shiny-related dependencies
• Added ggplot2 version of sunburst plot
• Fixed failing examples

# TreatmentPatterns 3.0.0

• Save parameters from computePathways() and export()
• A TreatmentPatternsResults (TPR) object is now returned from export()
• Updated output file and column names to be snake_case rather than camelCase
  • The TPR object houses all the different tables, and has methods for plotting.
  • Writing CSV-files on export() is now optional (export(outputPath = "some/path/"))
  • The TPR object has the following methods: saveAsCsv(), saveAsZip(), uploadResultsToDb(), load(), plotSunburst(), plotSankey(), plotEventDuration()
  • The TPR object has the following fields:, attrition, metadata, treatment_pathways, summary_event_duration, counts_age, counts_sex, counts_year, cdm_source_info, analyses, arguments
• Added ResultModelManager support
• Multiple target cohorts are now supported
• Allow target, event, and exit cohorts to be in different tables
• Stratification is now optional export(stratify = TRUE)
• When data is pulled from the database, subjects are assigned a new subject ID. This prevents rounding subject ID’s, and unintentional grouping of subjects.

# TreatmentPatterns 2.7.0

• Converted subject_id’s to pseudo_subject_id’s when pulling in data from the database. Actual subject_id’s are stored as character().
• Added copyright statements in R-files.
• Renamed periodPriorToIndex to indexDateOffset.
• Added life cycle badge to ReadMe
• Added more informative error / warning messaging when no data is available to compute pathways.
• Extended output for event duration in pathways.
• Renamed summaryStatsTherapyDuration.csv to summaryEventDuration.csv.
• Updated status messaging in the console.
• Added attrition.csv that capture status messaging and time stamps.

• # TreatmentPatterns 2.6.9

• Disabled tests when required suggested packages are not available.
• Updated examples to only run when required suggested packages are available.

# TreatmentPatterns 2.6.8

• Updated some tests to work with later versions of omopgenerics.
• Fixed issue with where combinations sometimes got miss-classified.
• Fixed issue when event starts and ends on end-date of target.
• Fixed issue when collapsing events when there is also a combination, when filterTreatments = "All".
• Added check in tests to only run if packages are availible. (noSuggests, M1).
• createSankeyDiagram() now supports pathways over 3 levels long.

# TreatmentPatterns 2.6.7

• Updated URLs in description
• Updated examples to fix compatibility with CDMConnector >1.4.0
• Updated so more tests run on CRAN

# TreatmentPatterns 2.6.6

• Internal update to shinyApp.
• Update to createSankeyDiagram() to properly handle combinations consisting of >2.
• Added article to package website about best practices.

# TreatmentPatterns 2.6.5

• Removed stringi, rjson, and googleVis as dependencies.
• Fixed unit tests that had dummy data outside of observation data.
• Fix for complex edge-case paths with re-occuring treatments.
• Added unit tests for edge-cases.
• Added warning when minEraDuration > minPostCombinationWindow or combinationWindow.
• Fixed issue when filterTreatments was set to "changes", age and sex columns were dropped from output.
• datatypes of cohort_table are now checked.

# TreatmentPatterns 2.6.4

• Fixed issue with minPostCombinationWindow broken by the fix for re-occurring treatments.
• Added unit tests for minPostCombinationWindow.

# TreatmentPatterns 2.6.3

• Fixed issue where there were duplicate rows in the exported treatmentPathways.csv file.
• Added more comprehensive description for parameter ageWindow for export().
• Added additional tests validating the counts per pathway being exported by export().
• Fixed issue with re-occurring treatments in pathways.
• Added dedicated logical tests for various pathways.
• Added dedicated tests for target-event cohort overlaps.
• Fixed spelling error in outputs, and interaction with them in shiny module.

# TreatmentPatterns 2.6.2

• Fixed spelling mistake ‘summaryStatsTherapyDuraion’ to ‘summaryStatsTherapyDuration’.
• Added error when only one level of data is available in the data when computing a sankey diagram.
• Removed createSankeyDiagram
• Renamed createSankeyDiagram2 to createSankeyDiagram
• Removed createSunburstPlot
• Renamed createSunburstPlot2 to createSunburstPlot
• Fixed issue where target cohorts had to be at least 1 day long to be included when using DatabaseConnector.
• Fixed CRAN error when running examples.
• Renamed cellCount to minCellCount in export documentation.
• Fixed issue where counts of pathways were not computed correctly.
• Fixed spelling error summaryStatsTherapyDuraion to summaryStatsTherapyDuration.
• Fixed issue in export where merging two tables could fail, if one was empty.

# TreatmentPatterns 2.6.1

• Bumped R version to 4.2.1
• Added additional unit-tests for createSunburstPlot2() and createSankeyDiagram2()
• Updated to use DATEDIFF to compute difference between dates when fetching data from the database.
• Added functionality to choose how groups below the minimum cell count are censored.
• Made some parameters ambiguous for createSankeyDiagram2 and createSunburstPlot2() for more control over the figures.
• Removed addNoPath parameter.
• Renamed minFreq parameter to minCellCount.
• Added Shiny app with exported module classes.
• Added censoring options of treatment pathways falling below the minCellCount.
• Simplified part of computePathways().
• Significantly improved performance of export().

# TreatmentPatterns 2.6.0

• Added createSunburstPlot2() as a replacement of createSunburstPlot(). createSunburstPlot2() will fully replace createSunburstPlot() in a future version.
• Moved DatabaseConnector and SqlRender to Suggests
• Removed obsolete dependencies: data.table, fs, and glue.
• Internal performance updates.
• Internal code style updates.
• Moved basic filtering when fetching cohort table.
• Updated ReadMe with functionalities
• Internal updates to LRFS paths, to allow paths of identical duration.

# TreatmentPatterns 2.5.2

• Resolved issue with finalize method of CDMInterface class.
• Resolved issue with schema references.
• Unified case style.
• Refactored code for createSunburstPlot().

# TreatmentPatterns 2.5.1

• Added checks for user input: cohorts and cohortTableName.
• Added option to directly return HTML when using createSankeyDiagram.
• Added option to directly return HTML when using createSunburstPlot.
• Added option for ageWindow to be a vector.
• Added input checking for export().
• Added additional check for frequency for createSunburstPlot.
• Resolved issue in createSunburstPlot, when converting from data.table to nested JSON.

# TreatmentPatterns 2.5.0

• Updated interface
• Some internal OO usage
• CDM & DatabaseConnector
• Uses Andromeda to be able to handle bigger than RAM data sets.
  • Shift from data.table to dplyr.
• Updated vignettes using new interface
• General code clean up
• Intermediate files are cached and accessable through Andromeda for review.
• Outputted CSV-files re-imagined to be more flexible for use post TreatmentPatterns.
• Sunburst and Sankey plots are now directly usable with treatmentPathways.csv.