scripts/setup_mindnet_collections.py hinzugefügt

scripts/setup_mindnet_collections.py

@@ -0,0 +1,100 @@
+#!/usr/bin/env python3
+"""
+Richtet Qdrant-Collections für dein Mindnet-Projekt ein (idempotent).
+
+Erzeugt:
+- mindnet_chunks  (size=384, distance=Cosine)  -> semantische Suche über Text-Chunks
+- mindnet_notes   (size=384, distance=Cosine)  -> Notizebene / Facettierung
+- mindnet_edges   (size=1,   distance=Cosine)  -> explizite Links (Dummy-Vektor; Filter via Payload)
+
+Legt sinnvolle Payload-Indizes an (keyword/text).
+"""
+
+import os
+import sys
+import json
+import argparse
+import requests
+
+QDRANT_URL = os.environ.get("QDRANT_URL", "http://127.0.0.1:6333")
+
+def api(method: str, path: str, **kwargs) -> requests.Response:
+    url = QDRANT_URL.rstrip("/") + path
+    r = requests.request(method, url, timeout=15, **kwargs)
+    if not r.ok:
+        raise RuntimeError(f"{method} {url} -> {r.status_code} {r.text}")
+    return r
+
+def exists(collection: str) -> bool:
+    r = api("GET", f"/collections/{collection}")
+    j = r.json()
+    return j.get("result", {}).get("status") == "green"
+
+def create_collection(collection: str, size: int, distance: str) -> None:
+    if exists(collection):
+        print(f"[=] {collection} existiert bereits.")
+        return
+    payload = {"vectors": {"size": size, "distance": distance}}
+    api("PUT", f"/collections/{collection}", json=payload)
+    print(f"[+] Collection {collection} angelegt (size={size}, distance={distance}).")
+
+def keyword_index(collection: str, field: str) -> None:
+    api("PUT", f"/collections/{collection}/index",
+        json={"field_name": field, "field_schema": "keyword"})
+    print(f"[+] keyword-Index: {collection}.{field}")
+
+def text_index(collection: str, field: str = "text") -> None:
+    api("PUT", f"/collections/{collection}/index",
+        json={"field_name": field, "field_schema": {"type": "text"}})
+    print(f"[+] text-Index: {collection}.{field}")
+
+def main():
+    p = argparse.ArgumentParser()
+    p.add_argument("--qdrant-url", default=QDRANT_URL, help="z.B. http://127.0.0.1:6333")
+    p.add_argument("--prefix", default="mindnet", help="Präfix für Collections")
+    p.add_argument("--dim", type=int, default=384, help="Embedding-Dimension (MiniLM: 384)")
+    p.add_argument("--distance", default="Cosine", choices=["Cosine", "Euclid", "Dot"],
+                   help="Distanzmetrik")
+    args = p.parse_args()
+
+    global QDRANT_URL
+    QDRANT_URL = args.qdrant_url
+
+    chunks = f"{args.prefix}_chunks"
+    notes  = f"{args.prefix}_notes"
+    edges  = f"{args.prefix}_edges"
+
+    # 1) Collections
+    create_collection(chunks, args.dim, args.distance)
+    create_collection(notes,  args.dim, args.distance)
+    create_collection(edges,  1,         args.distance)  # Dummy-Vektor
+
+    # 2) Indizes
+    # mindnet_chunks
+    for f in ["note_id", "Status", "Typ", "title", "path"]:
+        keyword_index(chunks, f)
+    for f in ["tags", "Rolle", "links"]:
+        keyword_index(chunks, f)
+    text_index(chunks, "text")
+
+    # mindnet_notes
+    for f in ["note_id", "title", "path", "Typ", "Status"]:
+        keyword_index(notes, f)
+    for f in ["tags", "Rolle"]:
+        keyword_index(notes, f)
+
+    # mindnet_edges
+    for f in ["src_note_id", "dst_note_id", "src_chunk_id", "dst_chunk_id", "link_text", "relation"]:
+        keyword_index(edges, f)
+
+    # 3) Übersicht
+    coll = api("GET", "/collections").json().get("result", {}).get("collections", [])
+    print("\n[Info] Collections vorhanden:")
+    print(json.dumps(coll, indent=2, ensure_ascii=False))
+
+if __name__ == "__main__":
+    try:
+        main()
+    except Exception as e:
+        print(f"[ERROR] {e}", file=sys.stderr)
+        sys.exit(1)