legal-ai/scripts/extract-citations.py

#!/usr/bin/env python3
"""Extract case law citations from block-yod and link to case_law table."""

import asyncio
import re
import sys
from pathlib import Path
from uuid import UUID

sys.path.insert(0, str(Path(__file__).parent.parent / "mcp-server" / "src"))

from legal_mcp.services.db import get_pool, init_schema, close_pool

# Patterns for Israeli case law citations
CITATION_PATTERNS = [
    # עע"מ, בג"ץ, ע"א, etc.
    re.compile(r'(עע"מ|בג"ץ|ע"א|בר"ם|עת"מ|עמ"נ|ע"ע|רע"א|דנ"א|בש"א)\s*(\d[\d/\-]+)'),
    # ערר with number
    re.compile(r'ערר\s*\(?\s*(?:מרכז|ירושלים|חי\'?|ת"א|דרום|צפון)?\s*\)?\s*(\d[\d/\-]+)'),
    # ערר without district
    re.compile(r'ערר\s+(\d{3,5}[\-/]\d{2,4})'),
]


def extract_citations_from_text(text: str) -> list[dict]:
    """Find all case law citations in text."""
    citations = []
    seen = set()

    for pattern in CITATION_PATTERNS:
        for match in pattern.finditer(text):
            full_match = match.group(0)
            if full_match in seen:
                continue
            seen.add(full_match)

            # Get surrounding context (50 chars before and after)
            start = max(0, match.start() - 50)
            end = min(len(text), match.end() + 100)
            context = text[start:end].replace("\n", " ")

            citations.append({
                "citation_text": full_match,
                "context": context,
            })

    return citations


async def main():
    await init_schema()
    pool = await get_pool()

    async with pool.acquire() as conn:
        # Get all block-yod content with decision info
        blocks = await conn.fetch(
            """SELECT db.content, d.id as decision_id, c.case_number
               FROM decision_blocks db
               JOIN decisions d ON d.id = db.decision_id
               JOIN cases c ON c.id = d.case_id
               WHERE db.block_id = 'block-yod' AND db.word_count > 0
               ORDER BY c.case_number"""
        )

        # Get existing case_law for matching
        case_laws = await conn.fetch("SELECT id, case_number, case_name FROM case_law")
        case_law_map = {}
        for cl in case_laws:
            # Index by various forms of the case number
            case_law_map[cl["case_number"]] = cl["id"]
            # Also index by short number (e.g., "3975/22" from "עע"מ 3975/22")
            parts = cl["case_number"].split()
            if len(parts) > 1:
                case_law_map[parts[-1]] = cl["id"]

    total_citations = 0
    total_linked = 0

    for block in blocks:
        case_number = block["case_number"]
        decision_id = block["decision_id"]
        text = block["content"]

        citations = extract_citations_from_text(text)

        if not citations:
            continue

        print(f"\n{case_number}: {len(citations)} ציטוטים נמצאו")

        async with pool.acquire() as conn:
            for cit in citations:
                total_citations += 1

                # Try to match to case_law table
                case_law_id = None
                for key, cl_id in case_law_map.items():
                    if key in cit["citation_text"] or cit["citation_text"] in key:
                        case_law_id = cl_id
                        break

                if case_law_id:
                    # Check if already exists
                    existing = await conn.fetchval(
                        """SELECT id FROM case_law_citations
                           WHERE case_law_id = $1 AND decision_id = $2""",
                        case_law_id, decision_id,
                    )
                    if not existing:
                        await conn.execute(
                            """INSERT INTO case_law_citations
                               (case_law_id, decision_id, citation_type, context_text)
                               VALUES ($1, $2, 'support', $3)""",
                            case_law_id, decision_id, cit["context"],
                        )
                        total_linked += 1
                        print(f"  ✅ {cit['citation_text'][:40]} → קושר לפסיקה")
                else:
                    print(f"  ⬜ {cit['citation_text'][:40]} — לא נמצא ב-DB")

    # Summary
    async with pool.acquire() as conn:
        total_in_db = await conn.fetchval("SELECT count(*) FROM case_law_citations")

    await close_pool()

    print(f"\n{'='*50}")
    print(f"סה\"כ ציטוטים שנמצאו: {total_citations}")
    print(f"סה\"כ קושרו לפסיקה ב-DB: {total_linked}")
    print(f"סה\"כ ב-case_law_citations: {total_in_db}")


if __name__ == "__main__":
    asyncio.run(main())