"""sources/klasifikacija.csv no partijām: tools/batch1.py (1. partija) + sources/batches/*.csv (point,type,function,rationale).
Visas partijas — MI priekšlikums (status ai_proposed), līdz cilvēks tās pārskata.
python3 tools/merge_batches.py  →  sources/klasifikacija.csv; pēc tam python3 tools/generate.py"""
import csv
import glob
import os
import sys

sys.path.insert(0, os.path.dirname(__file__))
import batch1  # noqa: E402

BY, DATE = "claude-opus-5-5", "2026-10-24"
TYPES = {"function", "procedural", "residual"}


def main(out="sources/klasifikacija.csv"):
    rows, seen = [], {}
    for point, t, fn, why in batch1.ROWS:
        rows.append((point, t, fn, why, "batch1"))
    for path in sorted(glob.glob("sources/batches/*.csv")):
        for r in csv.DictReader(open(path, encoding="utf-8")):
            rows.append((r["point"].strip(), r["type"].strip(), (r["function"] or "").strip(), (r["rationale"] or "").strip(),
                         os.path.basename(path)))
    # sources/remap.csv: point (exact id, or institution prefix ending with ".") from → to, after a catalogue split
    if os.path.exists("sources/remap.csv"):
        rules = list(csv.DictReader(open("sources/remap.csv", encoding="utf-8")))
        used = {i: 0 for i in range(len(rules))}
        remapped = []
        for point, t, fn, why, src in rows:
            for i, r in enumerate(rules):
                k = r["point"]
                if t == "function" and fn == r["from"] and (point == k or (k.endswith(".") and point.startswith(k))):
                    fn, used[i] = r["to"], used[i] + 1
                    break
            remapped.append((point, t, fn, why, src))
        rows = remapped
        idle = [rules[i]["point"] for i, n in used.items() if not n]
        if idle:
            sys.exit("remap.csv: neizmantoti noteikumi " + ", ".join(idle))
        print("remap:", sum(used.values()), "punkti pārvietoti")
    bad = []
    for point, t, fn, why, src in rows:
        if point in seen:
            bad.append(f"{point}: divreiz ({seen[point]}, {src})")
        seen[point] = src
        if t not in TYPES:
            bad.append(f"{point}: type {t!r}")
        if (t == "function") != bool(fn):
            bad.append(f"{point}: type {t} un function {fn!r} nesader")
        if not why:
            bad.append(f"{point}: nav pamatojuma")
    if bad:
        sys.exit("\n".join(bad[:40]))
    with open(out, "w", encoding="utf-8", newline="") as f:
        w = csv.writer(f)
        w.writerow(["point", "type", "function", "status", "by", "date", "rationale"])
        for point, t, fn, why, _src in sorted(rows):
            w.writerow([point, t, fn, "ai_proposed", BY, DATE, why])
    print(out, len(rows), "punkti no", len({r[4] for r in rows}), "partijām")


if __name__ == "__main__":
    main()