How duckartem-arts/improving-scoring-models-using-public-data works
The repository combines two workflows: KAD case collection and enrichment in enricher, followed by tax-case classification and evidence extraction in judical_search. The first incrementally gathers case listings, details, and decision text; the second normalizes and filters case data, extracts structured artifacts, and produces results or optional reports. The README documents several modes and capabilities not visible in the sampled code; those are shown as documented relationships rather than claims of source inspection.
Main entry points: src/main.py Judicial search runner · enricher/main.py Enricher runner · evidence_extractor.py Evidence extraction · cases.py Case search
Components 17
Case collection
- Enricher runner
enricher/main.py - Cookie handling
enricher/services/cookies.py - Case search
enricher/services/cases.py - Case details
enricher/services/case_details.py - Decision PDF text
enricher/services/case_details.py
Tax analysis
- Judicial search runner
judical_search/src/main.py - Text normalization
judical_search/src/utils/normalization.py - Term synonyms
judical_search/src/utils/synonyms.py - Tax relevance filter
judical_search/src/core/tax_filter.py
Evidence processing
- Evidence extraction
judical_search/src/extractors/evidence_extractor.py - Artifact model
judical_search/src/core/artifact.py - Data quality
judical_search/src/processors/quality_enhancer.py
Results and reports
- Tax results
judical_search/output/result.json - Quality metrics
judical_search/src/metrics.py - Result visualization
judical_search/src/visualization.py
Other components
- Operator
- KAD court search
Connections 26
- Operator → Enricher runner: runs collection
README.md - Enricher runner → Cookie handling: updates cookies
enricher/main.py - Cookie handling → KAD court search: opens site
- Enricher runner → Case search: fetches cases
enricher/main.py - Case search → Cookie handling: loads cookies
enricher/services/cases.py - Case search → KAD court search: searches cases
README.md - Enricher runner → Case details: fetches details
enricher/main.py - Case details → KAD court search: requests case page
README.md - Enricher runner → Decision PDF text: extracts decisions
enricher/main.py - Decision PDF text → KAD court search: fetches PDFs
README.md - Operator → Judicial search runner: runs analysis
README.md - Judicial search runner → Text normalization: normalizes text
README.md - Tax relevance filter → Text normalization: normalizes for filter
judical_search/src/core/tax_filter.py - Judicial search runner → Tax relevance filter: classifies cases
judical_search/src/main.py - Judicial search runner → Evidence extraction: extracts evidence
judical_search/src/main.py - Evidence extraction → Text normalization: normalizes text
judical_search/src/extractors/evidence_extractor.py - Evidence extraction → Term synonyms: uses terms
judical_search/src/extractors/evidence_extractor.py - Evidence extraction → Artifact model: creates artifacts
judical_search/src/extractors/evidence_extractor.py - Judicial search runner → Data quality: enhances data
judical_search/src/main.py - Data quality → Term synonyms: uses terms
judical_search/src/processors/quality_enhancer.py - Judicial search runner → Tax results: writes results
README.md - Operator → Quality metrics: requests metrics
README.md - Quality metrics → Tax results: evaluates predictions
- Operator → Result visualization: requests visualization
README.md - Judicial search runner → Result visualization: creates demo view
judical_search/src/main.py - Result visualization → Tax results: reads results