Trainer_LLM/scripts/bootstrap_qdrant_plans.py

80 lines
3.2 KiB
Python
Raw Permalink Normal View History

#!/usr/bin/env python3
# -*- coding: utf-8 -*-
"""
Bootstrap (idempotent) für Qdrant-Collections rund um Pläne – v1.3.0
- Fügt fehlende Payload-Indizes hinzu (KEYWORD/FLOAT), idempotent.
- NEU: FLOAT-Index `plans.created_at_ts` für serverseitige Zeitfensterfilter.
"""
import os
from qdrant_client import QdrantClient
from qdrant_client.models import PayloadSchemaType
QDRANT_HOST = os.getenv("QDRANT_HOST", "localhost")
QDRANT_PORT = int(os.getenv("QDRANT_PORT", "6333"))
PLANS = os.getenv("PLAN_COLLECTION") or os.getenv("QDRANT_COLLECTION_PLANS", "plans")
TEMPLATES = os.getenv("PLAN_TEMPLATE_COLLECTION", "plan_templates")
SESSIONS = os.getenv("PLAN_SESSION_COLLECTION", "plan_sessions")
INDEX_SPECS = {
TEMPLATES: [
("discipline", PayloadSchemaType.KEYWORD),
("age_group", PayloadSchemaType.KEYWORD),
("target_group", PayloadSchemaType.KEYWORD),
("section_names", PayloadSchemaType.KEYWORD),
("section_must_keywords", PayloadSchemaType.KEYWORD),
("section_ideal_keywords", PayloadSchemaType.KEYWORD),
("section_supplement_keywords", PayloadSchemaType.KEYWORD),
("section_forbid_keywords", PayloadSchemaType.KEYWORD),
("goals", PayloadSchemaType.KEYWORD),
],
PLANS: [
("discipline", PayloadSchemaType.KEYWORD),
("age_group", PayloadSchemaType.KEYWORD),
("target_group", PayloadSchemaType.KEYWORD),
("sections.name", PayloadSchemaType.KEYWORD), # legacy, belassen
("plan_section_names", PayloadSchemaType.KEYWORD),
("goals", PayloadSchemaType.KEYWORD),
("created_by", PayloadSchemaType.KEYWORD),
("created_at", PayloadSchemaType.KEYWORD),
("created_at_ts", PayloadSchemaType.FLOAT), # NEU
("fingerprint", PayloadSchemaType.KEYWORD),
("title", PayloadSchemaType.KEYWORD),
],
SESSIONS: [
("plan_id", PayloadSchemaType.KEYWORD),
("executed_at", PayloadSchemaType.KEYWORD),
("coach", PayloadSchemaType.KEYWORD),
("group_label", PayloadSchemaType.KEYWORD),
],
}
def _create_indexes(client: QdrantClient, collection: str, specs):
try:
client.get_collection(collection)
print(f"[Bootstrap v1.3.0] Collection '{collection}' ok.")
except Exception as e:
print(f"[Bootstrap v1.3.0] WARN: Collection '{collection}' nicht gefunden (wird beim ersten Upsert erstellt). Details: {e}")
return
for field, schema in specs:
try:
client.create_payload_index(collection_name=collection, field_name=field, field_schema=schema)
print(f"[Bootstrap v1.3.0] Index created: {collection}.{field} ({schema})")
except Exception as e:
print(f"[Bootstrap v1.3.0] Index exists or skipped: {collection}.{field} -> {e}")
def main():
print(f"[Bootstrap v1.3.0] Qdrant @ {QDRANT_HOST}:{QDRANT_PORT}")
print(f"[Bootstrap v1.3.0] Collections: TEMPLATES={TEMPLATES} PLANS={PLANS} SESSIONS={SESSIONS}")
client = QdrantClient(host=QDRANT_HOST, port=QDRANT_PORT)
for coll, specs in INDEX_SPECS.items():
_create_indexes(client, coll, specs)
print("[Bootstrap v1.3.0] done.")
if __name__ == "__main__":
main()