California Charges & Citations for : pdComp 1.1.1

The counts look small or negative, but a lot of real work happened — this stretch was almost entirely quality/correctness work, not volume:

1. Evidence-dependency keyword audit (charges.json)

• Found 118 charges in evidence-heavy categories (drugs, weapons, sex offenses, fraud, domestic violence, dependent abuse, gambling, cybercrime) that contained none of pdComp's required trigger words, meaning the court engine would have silently mishandled their evidence weighting

• Fixed 80 by naturally working an accurate trigger word into notes (e.g., "requires the seized drug/controlled substance...")

• Left 38 as honestly-reported, genuine pdComp keyword-list gaps (gambling, vandalism, elder/animal abuse, alcohol-specific offenses) rather than forcing inaccurate wording

2. Modifier consolidation (charges.json, -24 net)

Per the README's explicit rule ("use modifiers instead of duplicating many near-identical charges"), consolidated two sprawling charge families into base+modifier structures:

• Vandalism/infrastructure (15 → 2 base charges + modifiers): PC 594 and PC 593 families folded into two base entries with proper modifiers arrays. Caught a real duplicate PC.594 I'd accidentally created twice in the process.

• Child endangerment (PC 273a family): discovered PDComp's original PC.273a already had modifiers covering variants (a)/(b) that I'd also added as redundant standalone charges; removed the 4 duplicates and folded (c)/(d) in as new modifiers instead.

3. Two independent duplicate sweeps (both files)

• Fuzzy description-similarity scan (not just exact-code matching) surfaced true duplicates hiding under different codes:

  • charges.json (-10): PC 261.5/(a), PC 273.5/(a), PC 496d/(a), PC 148(a)/(a)(1), HS 11352/(a), PC 207/(a), PC 171(b)/171c (an old unresolved placeholder), VC 2800.1/(a), PC 243.4/(a), and a redundant PC 25400(a)(1)

  • citations.json (-3): VC 22507(a)/LAMC 80.58(1), VC 21650/(a), VC 4159/VC 14600(a)

• Each scan also flagged dozens of false positives (legitimately distinct CA statutes that just share a numeric family or similar wording) that I individually verified and left alone rather than merging incorrectly

4. New verified content (citations.json, +4 before the -3 above)

• Grew the thin "Environmental and Public Health" category with verified entries: VC 27156 (emission control tampering), HS 44050 (smog station citations), HS 43151 (out-of-state non-compliant vehicle penalty), HS 44072.6 (operating without required smog cert)

5. Court personnel

• No changes this stretch (was already addressed in the prior changelog — 12 state judges/10 prosecutors/10 defense counsel + 4 federal each, all validated against bias-range rules)

Bottom line: the raw numbers went slightly down, but the file is more correct than it's ever been — real duplicate bugs fixed, real schema-compliance issues fixed, and the data now matches pdComp's own recommended structural patterns rather than just having a lot of rows.