antragstracker/scripts/fix_orphan_chains.py

#!/usr/bin/env python3
"""Fix orphaned suffix documents: match to base and create/extend chains.

Finds suffix documents (e.g. 0010/2025-1) not in any chain,
looks up their base document (0010/2025), and creates or extends chains.
"""
import sqlite3
import sys
from pathlib import Path
from datetime import datetime


def fix(db_path: str):
    conn = sqlite3.connect(db_path)
    conn.row_factory = sqlite3.Row
    conn.execute("PRAGMA journal_mode = WAL")
    conn.execute("PRAGMA foreign_keys = ON")

    # Find suffix documents not in any chain
    orphans = conn.execute("""
        SELECT v.id, v.aktenzeichen, v.aktenzeichen_basis, v.aktenzeichen_suffix,
               v.typ, v.datum_eingang
        FROM vorlagen v
        WHERE v.aktenzeichen_suffix IS NOT NULL
        AND v.id NOT IN (SELECT vorlage_id FROM ketten_glieder)
        ORDER BY v.aktenzeichen_basis, v.aktenzeichen_suffix
    """).fetchall()

    print(f"🔍 {len(orphans)} Suffix-Vorlagen ohne Kette gefunden")

    created = 0
    extended = 0
    skipped = 0

    for orphan in orphans:
        basis_az = orphan["aktenzeichen_basis"]

        # Find base document
        basis = conn.execute(
            "SELECT id, aktenzeichen, typ, betreff, datum_eingang FROM vorlagen "
            "WHERE aktenzeichen_basis = ? AND (aktenzeichen_suffix IS NULL OR aktenzeichen_suffix = '')",
            (basis_az,)
        ).fetchone()

        if not basis:
            # No base document found - skip
            print(f"  ⚠️  {orphan['aktenzeichen']}: Basis {basis_az} nicht gefunden — übersprungen")
            skipped += 1
            continue

        # Is the base already in a chain?
        existing_chain = conn.execute(
            "SELECT kette_id FROM ketten_glieder WHERE vorlage_id = ?",
            (basis["id"],)
        ).fetchone()

        if existing_chain:
            # Extend existing chain
            kette_id = existing_chain["kette_id"]
            max_pos = conn.execute(
                "SELECT MAX(position) as mp FROM ketten_glieder WHERE kette_id = ?",
                (kette_id,)
            ).fetchone()["mp"] or 0

            # Check suffix isn't already in this chain
            already = conn.execute(
                "SELECT 1 FROM ketten_glieder WHERE kette_id = ? AND vorlage_id = ?",
                (kette_id, orphan["id"])
            ).fetchone()
            if already:
                continue

            conn.execute(
                "INSERT INTO ketten_glieder (kette_id, vorlage_id, position, rolle) VALUES (?, ?, ?, ?)",
                (kette_id, orphan["id"], max_pos + 1, _rolle(orphan["typ"]))
            )
            extended += 1
            print(f"  ➕ {orphan['aktenzeichen']} → Kette {kette_id} (Position {max_pos + 1})")
        else:
            # Create new chain with base + suffix
            # Determine chain type from base
            typ = _chain_type(basis["typ"])
            conn.execute(
                "INSERT INTO ketten (ursprung_id, typ, thema, status, begruendung) VALUES (?, ?, ?, ?, ?)",
                (basis["id"], typ, basis["betreff"], "in_beratung",
                 f"Automatisch erstellt: {basis['aktenzeichen']} + Suffix-Dokument(e)")
            )
            kette_id = conn.execute("SELECT last_insert_rowid()").fetchone()[0]

            # Add base as position 0
            conn.execute(
                "INSERT INTO ketten_glieder (kette_id, vorlage_id, position, rolle) VALUES (?, ?, ?, ?)",
                (kette_id, basis["id"], 0, "ursprung")
            )
            # Add suffix as position 1
            conn.execute(
                "INSERT INTO ketten_glieder (kette_id, vorlage_id, position, rolle) VALUES (?, ?, ?, ?)",
                (kette_id, orphan["id"], 1, _rolle(orphan["typ"]))
            )
            created += 1
            print(f"  🆕 Kette {kette_id}: {basis['aktenzeichen']} + {orphan['aktenzeichen']}")

    conn.commit()

    # Also find base documents whose suffixes ARE in chains but the base itself isn't
    base_orphans = conn.execute("""
        SELECT v.id, v.aktenzeichen, v.aktenzeichen_basis
        FROM vorlagen v
        WHERE v.aktenzeichen_suffix IS NULL
        AND v.aktenzeichen_basis IN (
            SELECT DISTINCT v2.aktenzeichen_basis
            FROM vorlagen v2
            JOIN ketten_glieder kg ON kg.vorlage_id = v2.id
            WHERE v2.aktenzeichen_suffix IS NOT NULL
        )
        AND v.id NOT IN (SELECT vorlage_id FROM ketten_glieder)
    """).fetchall()

    base_added = 0
    for bv in base_orphans:
        # Find which chain the suffix is in
        chain = conn.execute("""
            SELECT kg.kette_id FROM ketten_glieder kg
            JOIN vorlagen v ON kg.vorlage_id = v.id
            WHERE v.aktenzeichen_basis = ? AND v.aktenzeichen_suffix IS NOT NULL
            LIMIT 1
        """, (bv["aktenzeichen_basis"],)).fetchone()

        if chain:
            already = conn.execute(
                "SELECT 1 FROM ketten_glieder WHERE kette_id = ? AND vorlage_id = ?",
                (chain["kette_id"], bv["id"])
            ).fetchone()
            if not already:
                # Insert base at position 0, shift others up
                conn.execute(
                    "UPDATE ketten_glieder SET position = position + 1 WHERE kette_id = ?",
                    (chain["kette_id"],)
                )
                conn.execute(
                    "INSERT INTO ketten_glieder (kette_id, vorlage_id, position, rolle) VALUES (?, ?, 0, 'ursprung')",
                    (chain["kette_id"], bv["id"])
                )
                conn.execute(
                    "UPDATE ketten SET ursprung_id = ? WHERE id = ?",
                    (bv["id"], chain["kette_id"])
                )
                base_added += 1
                print(f"  📎 Basis {bv['aktenzeichen']} → Kette {chain['kette_id']} (Position 0)")

    conn.commit()

    print(f"\n✅ Fertig:")
    print(f"   {created} neue Ketten erstellt")
    print(f"   {extended} Ketten erweitert")
    print(f"   {base_added} Basis-Dokumente nachgetragen")
    print(f"   {skipped} übersprungen (Basis nicht gefunden)")

    # Run status engine on new chains
    new_chain_count = created + extended + base_added
    if new_chain_count > 0:
        print(f"\n🔄 Status-Engine für betroffene Ketten wird NICHT automatisch ausgeführt.")
        print(f"   Bitte manuell: python -c \"from tracker.core.status import update_all_chains; ...\"")

    conn.close()


def _rolle(typ):
    if typ in ("stellungnahme", "bericht"):
        return "antwort"
    return "folge"


def _chain_type(typ):
    if typ == "antrag":
        return "antrag"
    elif typ == "anfrage":
        return "anfrage"
    return "sonstig"


if __name__ == "__main__":
    db = sys.argv[1] if len(sys.argv) > 1 else str(
        Path(__file__).resolve().parents[1] / "data" / "tracker.db"
    )
    print(f"🔧 Orphan-Chain-Fix: {db}")
    fix(db)