Jean-Luc Makiola
9643011e64
app/mcp_server.py: FastMCP (mcp SDK), streamable-http auf /mcp, statischer Bearer-Token (constant-time ASGI-Middleware), Fail-Fast ohne RAG_MCP_TOKEN. Tools rag_search (mit semester/fach/typ-Filter) + get_file_chunks. Läuft aus demselben Image wie der Ingestor und reused den Embed-Pfad → Vektoren sind garantiert kompatibel zum Ingest (der offizielle qdrant-MCP-Server kann nur fastembed → Dimension-/Schema-Mismatch). app/qdrant_store.py: search_chunks (query_points + optionaler Payload-Filter) und get_chunks_by_path (scroll, nach chunk_index sortiert). app/bulk.py: Amplification-Guard — /bulk-import lehnt mit 409 ab solange ein vorheriger Bulk noch BackgroundTasks abarbeitet. docker-compose.coolify.yml: rag-mcp-Service (nicht public, externes metamcp-net statt Stack-Coupling) + Traefik-Rate-Limit-Middleware am ingestor. tests/conftest.py: Settings-env_file in Tests neutralisieren (Dev-.env darf die Suite nicht kontaminieren). 68 passed, ruff clean. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
39 lines
708 B
TOML
39 lines
708 B
TOML
[project]
|
|
name = "rag-ingestor"
|
|
version = "0.1.0"
|
|
description = "Nextcloud → Qdrant RAG ingestion service"
|
|
requires-python = ">=3.12"
|
|
dependencies = [
|
|
"fastapi>=0.115",
|
|
"uvicorn[standard]>=0.32",
|
|
"httpx>=0.28",
|
|
"pydantic>=2.10",
|
|
"pydantic-settings>=2.7",
|
|
"pymupdf>=1.25",
|
|
"python-docx>=1.1",
|
|
"ollama>=0.4",
|
|
"qdrant-client>=1.12",
|
|
"mcp>=1.9",
|
|
]
|
|
|
|
[dependency-groups]
|
|
dev = [
|
|
"pytest>=8.3",
|
|
"pytest-asyncio>=0.25",
|
|
"respx>=0.22",
|
|
]
|
|
|
|
[tool.pytest.ini_options]
|
|
asyncio_mode = "auto"
|
|
testpaths = ["tests"]
|
|
|
|
[tool.uv]
|
|
package = true
|
|
|
|
[build-system]
|
|
requires = ["hatchling"]
|
|
build-backend = "hatchling.build"
|
|
|
|
[tool.hatch.build.targets.wheel]
|
|
packages = ["app"]
|