adybbroe / activefires-pp / 40192466936

75%

main: 75%

push

github

Refactor active-fire processing into shared processing API

Introduce a shared object-oriented processing layer for active-fire
filtering and use it from both the operational post-processing runner
and the standalone system-test workflow.

Add `ProcessingOptions`, `ActiveFiresProcessor`, and
`ActiveFiresProcessingResult` to separate core fire processing from
Posttroll orchestration and output handling. The processor now owns the
common sequence for reading detections, applying national boundary
filtering, removing spurious detections, and applying the exclusion
mask.

Update `ActiveFiresPostprocessing` to delegate filtering to the shared
processor while retaining the existing `fires_filtering()` interface as
a thin compatibility wrapper. Existing processing-related imports are
also kept available from `post_processing.py` to avoid unnecessary API
breakage.

Update the standalone `systemtest_read_and_filter.py` workflow to use
the same processing implementation as the live runner. This removes the
previous divergence where the standalone example did not apply the
spurious-detection filtering step.

Preserve the distinction between input files containing no detections
and files where detections are present initially but all removed during
filtering by recording the original detection count in the processing
result.

Add tests for the new processing API and update existing mocks to patch
the canonical processing module after moving the implementation out of
`post_processing.py`.

This refactoring keeps the operational processing behaviour unchanged
while establishing a single reusable processing path for future
development and testing.

Signed-off-by: Adam.Dybbroe <a000680@c22526.ad.smhi.se>

201 of 338 branches covered (59.47%)

993 of 1331 relevant lines covered (74.61%)

2.97 hits per line