Minor GHP updates

[ppinchuk]

Minor updates to GHP collection and parsing before validation

[Copilot]

Pull request overview

Minor improvements to the GHP (geothermal heat pump) pipeline around document collection/parsing and pre-validation filtering, including tighter “substantive text” gating and optional limits on how many documents get parsed per jurisdiction.

Changes:

• Add max_num_docs_to_parse_per_jurisdiction request setting and plumb it through extraction to cap the number of docs passed into plugin parsing.
• Improve document-type validation prompts/graph to require substantive legal text (not just ToC/headings/citations) before treating content as legally binding.
• Normalize Docling “missing” confidence values to None (instead of NaN/pd.NA) and add a unit test for that behavior.

Reviewed changes

Copilot reviewed 18 out of 18 changed files in this pull request and generated 3 comments.

Show a summary per file

File	Description
tests/python/unit/services/test_services_cpu.py	Adds coverage for _read_docling confidence normalization behavior.
tests/python/unit/pipeline/test_pipeline_orchestration.py	Updates pipeline orchestration tests to align with the new filter_docs signature.
tests/python/integration/test_integrated_pipeline_orchestration.py	Adjusts integration test plugin to match updated plugin interface (filter_docs).
docs/source/dev/advanced_plugin_development.rst	Updates plugin contract documentation for filter_docs(..., max_num_docs=None).
compass/validation/graphs.py	Adds a “substantive legal text” gate to the legal-text/document-type decision tree.
compass/services/cpu.py	Normalizes Docling confidence values via _none_if_missing helper.
compass/plugin/one_shot/components.py	Tightens collection prompt to avoid treating headings/ToC/citations as relevant by themselves.
compass/plugin/interface.py	Extends filter_docs to accept max_num_docs and implements slicing before parsing.
compass/plugin/base.py	Updates the abstract filter_docs contract to include max_num_docs.
compass/pipeline/jurisdiction.py	Adds progress-bar status update while loading pre-parsed documents.
compass/pipeline/extraction.py	Passes the per-jurisdiction parse cap into extractor.filter_docs(...).
compass/pipeline/data_classes.py	Introduces DocParsingParams and adds max_num_docs_to_parse_per_jurisdiction to requests.
compass/pipeline/coordinator.py	Adds manifest-loading logs for extraction mode.
compass/pipeline/collection/steps.py	Minor formatting change around local file loader kwargs.
compass/extraction/water/plugin.py	Updates filter_docs signature for the new interface (but currently uses __).
compass/extraction/ghp/plugin_config.yaml	Tweaks GHP query templates and heuristic keywords toward “private heat exchange wells”.
compass/extraction/ghp/geothermal_heat_pump_schema.json5	Refines GHP schema scope/evidence rules to reject non-substantive headings/ToC-only evidence.

[codecov-commenter]

Codecov Report

❌ Patch coverage is 74.80315% with 32 lines in your changes missing coverage. Please review.
✅ Project coverage is 62.59%. Comparing base (8290f5e) to head (09b6077).

Files with missing lines	Patch %	Lines
compass/services/cpu.py	72.72%	9 Missing and 6 partials ⚠️
compass/pipeline/collection/steps.py	16.66%	10 Missing ⚠️
compass/plugin/interface.py	25.00%	3 Missing ⚠️
compass/validation/graphs.py	0.00%	3 Missing ⚠️
compass/plugin/one_shot/components.py	94.73%	1 Missing ⚠️

❌ Your patch status has failed because the patch coverage (74.80%) is below the target coverage (80.00%). You can increase the patch coverage or adjust the target coverage.

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #476      +/-   ##
==========================================
+ Coverage   61.72%   62.59%   +0.86%     
==========================================
  Files          77       77              
  Lines        6937     7042     +105     
  Branches      690      704      +14     
==========================================
+ Hits         4282     4408     +126     
+ Misses       2535     2501      -34     
- Partials      120      133      +13     

Flag	Coverage Δ
unittests	62.59% <74.80%> (+0.86%)	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Harness.
📢 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.