QueryLake/pyproject.toml at main · kmccleary3301/QueryLake · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
[build-system]
requires = ["hatchling>=1.25.0"]
build-backend = "hatchling.build"

[project]
name = "querylake-backend"
version = "0.0.0"
description = "QueryLake Backend (FastAPI + Ray Serve)"
readme = "README.md"
requires-python = ">=3.10,<3.14"
license = { text = "Apache-2.0" }
authors = [
  { name = "QueryLake Team" }
]
keywords = ["querylake", "rag", "retrieval", "ray-serve", "fastapi", "llm"]
classifiers = [
  "Development Status :: 4 - Beta",
  "Intended Audience :: Developers",
  "Intended Audience :: Science/Research",
  "Programming Language :: Python :: 3",
  "Programming Language :: Python :: 3.10",
  "Programming Language :: Python :: 3.11",
  "Programming Language :: Python :: 3.12",
  "License :: OSI Approved :: Apache Software License",
]

# NOTE:
# - Base dependencies target an "API-only" environment (no local inference engines).
# - Heavier stacks (HF/torch, OCR, vLLM) live in optional extras.
dependencies = [
  # Core API
  "fastapi",
  "uvicorn",
  "sse-starlette",
  "python-multipart>=0.0.22",

  # HTTP clients / streaming
  "httpx",
  "httpx-sse",

  # Auth / crypto
  "PyJWT>=2.10.1",
  "python-jose",
  "passlib",
  "fastapi-login",
  "cryptography",
  "pycryptodome",
  "eciespy",
  "eth-keys",

  # DB / storage
  "sqlalchemy==2.0.35",
  "sqlmodel==0.0.22",
  "psycopg2-binary==2.9.9",
  "pgvector",
  "redis",

  # Utilities
  "tiktoken",
  "openai",
  "jsonpatch",
  "jsonpath-ng",

  # Document + web tooling (non-ML)
  "pypdf",
  "pypdfium2",
  "py7zr",
  "markdownify",
  "readability-lxml",
  "lxml-html-clean",
  "selenium",

  # Web search provider
  "google-search-results",

  # Ray Serve (core runtime) — keep pinned to match the cluster Ray version.
  "ray[serve]==2.53.0",

  # GPU inspection (VRAM-aware scheduling, diagnostics)
  "nvidia-ml-py",
]

[project.optional-dependencies]
# CLI / setup helpers (interactive config + HF downloads).
cli = [
  "colorama",
  "huggingface_hub[cli]",
]

# Local HF / torch inference (embeddings, rerank, etc).
inference-hf = [
  "torch",
  "transformers",
  "sentence-transformers",
  "FlagEmbedding",
  "huggingface_hub[cli]",
  "safetensors",
  "sentencepiece>=0.1.99",
  "lm-format-enforcer",
  "peft",
]

# OCR stack (Marker/Surya + optional OCRmyPDF).
ocr = [
  "torch",
  "pillow",
  "ocrmypdf",
  "marker-pdf==0.3.9",
  "surya-ocr==0.6.11",
]

# vLLM runtime is intentionally separate in production (run as upstream HTTP service),
# but is provided here as an optional extra for experiments.
vllm = [
  "vllm>=0.14.1",
]

dev = [
  "pytest",
  "pytest-asyncio",
]

[project.urls]
Homepage = "https://github.com/kmccleary3301/QueryLake"
Documentation = "https://github.com/kmccleary3301/QueryLake/tree/main/docs"
Source = "https://github.com/kmccleary3301/QueryLake"
Issues = "https://github.com/kmccleary3301/QueryLake/issues"

[project.scripts]
querylake-backend = "start_querylake:main"

[tool.hatch.build.targets.wheel]
packages = ["QueryLake"]