climatehealth development
notes
Unreleased
1.0.1
- CRAN resubmission focused on Fedora Linux check reliability and test
stability.
- Fixed INLA model execution under check-core constraints:
- limits INLA thread usage when
_R_CHECK_LIMIT_CORES_ is
set
- keeps compatibility across INLA versions by conditionally passing
thread-related arguments only when supported
- Hardened CRAN behavior for fragile INLA end-to-end disease tests:
- retained CRAN-safe skips for diarrhea and malaria integration
paths
- avoids Fedora-specific parallel worker failures seen in prior
checks
- Improved integration-test governance in CI:
- heavy integration tests are opt-in via
RUN_INTEGRATION=true
- standard PR/push CI remains fast and high-signal by default
- Build hygiene improvement:
- excludes
.covrignore from package build input via
.Rbuildignore
1.0.0
- Added new descriptive statistics public APIs:
run_descriptive_stats()
run_descriptive_stats_api()
- Added run-scoped output folders for descriptive stats:
<output_path>/descriptive_stats/<run_id>/All/
<output_path>/descriptive_stats/<run_id>/<Region>/
- Kept backward-compatible aliases with runtime deprecation warnings:
common_descriptive_stats_core()
common_descriptive_stats()
common_descriptive_stats_api()
- Improved descriptive stats robustness:
- correlation now validates and filters numeric columns
explicitly
- centralized preflight column and parameter validation for enabled
features
- typed summary behavior for non-numeric columns
- safer plot device cleanup with managed PDF helper
- Expanded tests for wrapper behavior, run-folder outputs, API payload
validation, endpoint-style contract responses, and deprecated alias
compatibility.
- Moved API runtime assets previously held in
climatehealth_pipelines into the package:
- plumber routes and startup scripts under
inst/plumber/
- throttling modules and tests
- package-scoped API and developer config templates under
inst/extdata/config_templates/
- operational helper scripts under
inst/scripts/
- Improved package/API compatibility for runtime usage:
- enabled API mode in plumber runtime
- added direct endpoint bindings for descriptive statistics, malaria,
and diarrhea
- enforced API-safe return payloads for malaria and diarrhea in API
mode
- extended internals to accept API-style record payloads and map
records
- updated throttle estimation to size JSON-record payloads instead of
file paths
- Cleaned up the air pollution module:
- standardized argument and variable naming
- refreshed documentation
- resolved R CMD check notes for global variable bindings
- added backward-compatible support for legacy
air_pollution_do_analysis() argument names
- expanded air pollution regression and integration tests