about summary refs log tree commit diff
path: root/pkgs/development/python-modules/txtai/default.nix
blob: 4bf59621b5ca774f196cb34b34176db17c9041c9 (plain) (blame)
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
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
{
  lib
, buildPythonPackage
, pythonOlder
, fetchFromGitHub
, pythonRelaxDepsHook
# propagated build input
, faiss
, torch
, transformers
, huggingface-hub
, numpy
, pyyaml
, regex
# optional-dependencies
, aiohttp
, fastapi
, uvicorn
# TODO add apache-libcloud
# , apache-libcloud
, rich
, duckdb
, pillow
, networkx
, python-louvain
, onnx
, onnxruntime
, soundfile
, scipy
, ttstokenizer
, beautifulsoup4
, nltk
, pandas
, tika
, imagehash
, timm
, fasttext
, sentencepiece
, accelerate
, onnxmltools
, annoy
, hnswlib
# TODO add pymagnitude-lite
#, pymagnitude-lite
, scikit-learn
, sentence-transformers
, croniter
, openpyxl
, requests
, xmltodict
# native check inputs
, unittestCheckHook
}:
let
  version = "7.0.0";
  api = [ aiohttp fastapi uvicorn ];
  # cloud = [ apache-libcloud ];
  console = [ rich ];

  database = [ duckdb pillow ];

  graph = [ networkx python-louvain ];

  model = [ onnx onnxruntime ];

  pipeline-audio = [ onnx onnxruntime soundfile scipy ttstokenizer ];
  pipeline-data = [ beautifulsoup4 nltk pandas tika ];
  pipeline-image = [ imagehash pillow timm ];
  pipeline-text = [ fasttext sentencepiece ];
  pipeline-train = [ accelerate onnx onnxmltools onnxruntime ];
  pipeline = pipeline-audio ++ pipeline-data ++ pipeline-image ++ pipeline-text ++ pipeline-train;

  similarity = [
    annoy
    fasttext
    hnswlib
    # pymagnitude-lite
    scikit-learn
    sentence-transformers
  ];
  workflow = [
    # apache-libcloud
    croniter
    openpyxl
    pandas
    pillow
    requests
    xmltodict
  ];
  all = api ++ console ++ database ++ graph ++ model ++ pipeline ++ similarity ++ workflow;

  optional-dependencies = {
    inherit api console database graph model pipeline-audio pipeline-image
      pipeline-text pipeline-train pipeline similarity workflow all;
  };
in
buildPythonPackage {
  pname = "txtai";
  inherit version;
  format = "setuptools";

  disabled = pythonOlder "3.8";

  src = fetchFromGitHub {
    owner = "neuml";
    repo = "txtai";
    rev = "refs/tags/v${version}";
    hash = "sha256-dYAzbdGEu9gAUzptuSfej6T5LPyTneRWigdAMlGgeqM=";
  };

  nativeBuildInputs = [
    pythonRelaxDepsHook
  ];

  pythonRemoveDeps = [
    # We call it faiss, not faiss-cpu.
    "faiss-cpu"
  ];

  propagatedBuildInputs = [
    faiss
    torch
    transformers
    huggingface-hub
    numpy
    pyyaml
    regex
  ];

  passthru.optional-dependencies = optional-dependencies;

  pythonImportsCheck = [ "txtai" ];

  # some tests hang forever
  doCheck = false;

  preCheck = ''
    export TRANSFORMERS_CACHE=$(mktemp -d)
  '';

  nativeCheckInputs = [
    unittestCheckHook
  ] ++ optional-dependencies.api ++ optional-dependencies.similarity;

  unittestFlagsArray = [
    "-s" "test/python" "-v"
  ];

  meta = with lib; {
    description = "Semantic search and workflows powered by language models";
    changelog = "https://github.com/neuml/txtai/releases/tag/v${version}";
    homepage = "https://github.com/neuml/txtai";
    license = licenses.asl20;
    maintainers = with maintainers; [ happysalada ];
  };
}