#!/usr/bin/env python3
# -*- coding: utf-8 -*-
"""
scripts/diag_payload_indexes.py
Listet die Payload-Indizes (payload_schema) für alle Mindnet-Collections.

Nutzt client.get_collection(name) und zeigt:
- payload_schema (Feld -> Schema/Typ)
- Vector-Konfiguration (size, distance, named vectors)
- Segment-Infos (optional komprimiert)

Aufruf:
  python3 -m scripts.diag_payload_indexes
  python3 -m scripts.diag_payload_indexes --raw   # vollständiges JSON

Hinweis: In Qdrant werden die Payload-Indizes im Collection-Info-Endpunkt unter
`result.payload_schema` geliefert.
"""
from __future__ import annotations
import argparse, json
from qdrant_client.http import models as rest
from app.core.qdrant import QdrantConfig, get_client, collection_names

def compact_schema(ps: dict) -> dict:
    out = {}
    for k, v in (ps or {}).items():
        # v ist z.B. {"type":"keyword"} oder {"type":"text"} etc.
        if isinstance(v, dict):
            out[k] = v.get("type") or v
        else:
            out[k] = v
    return out

def get_info(client, name: str) -> dict:
    info = client.get_collection(collection_name=name)
    # Das SDK liefert .dict(). Wir extrahieren das Wesentliche.
    d = info.dict() if hasattr(info, "dict") else info
    result = d.get("result", {})
    vectors = result.get("vectors") or {}
    if isinstance(vectors, dict) and "config" in vectors:
        vectors = vectors.get("config")
    payload_schema = result.get("payload_schema") or {}
    segments = result.get("segments_count")
    return {
        "name": name,
        "vectors": vectors,
        "payload_schema": compact_schema(payload_schema),
        "segments_count": segments,
    }

def main():
    ap = argparse.ArgumentParser()
    ap.add_argument("--raw", action="store_true", help="Gibt das rohe get_collection-Result zurück")
    args = ap.parse_args()

    cfg = QdrantConfig.from_env()
    client = get_client(cfg)
    notes, chunks, edges = collection_names(cfg.prefix)

    cols = [notes, chunks, edges]
    if args.raw:
        for name in cols:
            info = client.get_collection(collection_name=name)
            d = info.dict() if hasattr(info, "dict") else info
            print(json.dumps({"collection": name, "raw": d}, ensure_ascii=False))
    else:
        out = [ get_info(client, name) for name in cols ]
        print(json.dumps({"prefix": cfg.prefix, "collections": out}, ensure_ascii=False, indent=2))

if __name__ == "__main__":
    main()