72 lines
2.1 KiB
Python
72 lines
2.1 KiB
Python
|
"""
|
||
|
|
检查 Qdrant 中课程知识库的数据结构
|
||
|
|
"""
|
||
|
|
|
||
|
|
import asyncio
|
||
|
|
import sys
|
||
|
|
from pathlib import Path
|
||
|
|
|
||
|
|
sys.path.insert(0, str(Path(__file__).parent.parent))
|
||
|
|
|
||
|
|
from qdrant_client import AsyncQdrantClient
|
||
|
|
|
||
|
|
from app.core.config import get_settings
|
||
|
|
|
||
|
|
|
||
|
|
from app.core.qdrant_client import QdrantClient
|
||
|
|
|
||
|
|
|
||
|
|
async def check_course_kb():
|
||
|
|
"""检查课程知识库"""
|
||
|
|
settings = get_settings()
|
||
|
|
|
||
|
|
client = QdrantClient()
|
||
|
|
qdrant = await client.get_client()
|
||
|
|
|
||
|
|
tenant_id = "szmp@ash@2026"
|
||
|
|
course_kb_id = "75c465fe-277d-455d-a30b-4b168adcc03b"
|
||
|
|
|
||
|
|
safe_tenant_id = tenant_id.replace('@', '_')
|
||
|
|
prefix = settings.qdrant_collection_prefix
|
||
|
|
|
||
|
|
expected_collection = f"{prefix}{safe_tenant_id}_{course_kb_id}"
|
||
|
|
|
||
|
|
print(f"\n{'='*80}")
|
||
|
|
print(f"检查课程知识库 Collection")
|
||
|
|
print(f"{'='*80}")
|
||
|
|
print(f"租户 ID: {tenant_id}")
|
||
|
|
print(f"课程知识库 ID: {course_kb_id}")
|
||
|
|
print(f"预期 Collection 名称: {expected_collection}")
|
||
|
|
|
||
|
|
collections = await qdrant.get_collections()
|
||
|
|
collection_names = [c.name for c in collections.collections]
|
||
|
|
|
||
|
|
print(f"\n租户的所有 Collections:")
|
||
|
|
for name in collection_names:
|
||
|
|
if safe_tenant_id in name:
|
||
|
|
print(f" - {name}")
|
||
|
|
|
||
|
|
if expected_collection in collection_names:
|
||
|
|
print(f"\n✅ 课程知识库 Collection 存在: {expected_collection}")
|
||
|
|
|
||
|
|
points, _ = qdrant.scroll(
|
||
|
|
collection_name=expected_collection,
|
||
|
|
limit=3,
|
||
|
|
with_vectors=False,
|
||
|
|
)
|
||
|
|
|
||
|
|
print(f"\n课程知识库数据 (共 {len(points)} 条):")
|
||
|
|
for i, point in enumerate(points, 1):
|
||
|
|
payload = point.get('payload', {})
|
||
|
|
print(f"\n [{i}] id: {point.get('id')}")
|
||
|
|
print(f" payload keys: {list(payload.keys())}")
|
||
|
|
if 'metadata' in payload:
|
||
|
|
print(f" metadata: {payload['metadata']}")
|
||
|
|
else:
|
||
|
|
print(f"\n❌ 课程知识库 Collection 不存在!")
|
||
|
|
print(f" 可用的 Collections: {collection_names}")
|
||
|
|
|
||
|
|
|
||
|
|
if __name__ == "__main__":
|
||
|
|
asyncio.run(check_course_kb())
|