mirror of
https://github.com/NixOS/nixpkgs.git
synced 2026-11-04 11:50:04 +00:00
177 lines
3.8 KiB
Nix
177 lines
3.8 KiB
Nix
{
|
|
lib,
|
|
stdenv,
|
|
buildPythonPackage,
|
|
fetchFromGitHub,
|
|
|
|
# build-system
|
|
hatchling,
|
|
|
|
# dependencies
|
|
boto3,
|
|
docling-core,
|
|
docling-slim,
|
|
httpx,
|
|
pandas,
|
|
pydantic-settings,
|
|
|
|
# optional-dependencies
|
|
# rq:
|
|
rq,
|
|
msgpack,
|
|
# ray:
|
|
redis,
|
|
psutil,
|
|
ray,
|
|
# gdrive:
|
|
google-api-python-client,
|
|
google-auth-oauthlib,
|
|
# gcloudstorage:
|
|
google-cloud-storage,
|
|
# azure:
|
|
azure-storage-blob,
|
|
# opensearch:
|
|
opensearch-py,
|
|
# kafka:
|
|
confluent-kafka,
|
|
|
|
# tests
|
|
aiohttp,
|
|
pytest-asyncio,
|
|
pytestCheckHook,
|
|
starlette,
|
|
redisTestHook,
|
|
writableTmpDirAsHomeHook,
|
|
}:
|
|
|
|
buildPythonPackage (finalAttrs: {
|
|
pname = "docling-jobkit";
|
|
version = "3.9.1";
|
|
pyproject = true;
|
|
__structuredAttrs = true;
|
|
|
|
src = fetchFromGitHub {
|
|
owner = "docling-project";
|
|
repo = "docling-jobkit";
|
|
tag = "v${finalAttrs.version}";
|
|
hash = "sha256-tRws9cRy8zEuy16f/ZTAT99SiAfpVKKtwjnVIa6OLFk=";
|
|
};
|
|
|
|
build-system = [
|
|
hatchling
|
|
];
|
|
|
|
pythonRelaxDeps = [
|
|
"pandas"
|
|
];
|
|
dependencies = [
|
|
boto3
|
|
docling-core
|
|
docling-slim
|
|
httpx
|
|
pandas
|
|
pydantic-settings
|
|
]
|
|
++ docling-slim.optional-dependencies.standard;
|
|
|
|
optional-dependencies = {
|
|
vlm = docling-slim.optional-dependencies.models-vlm-inline;
|
|
rq = [
|
|
rq
|
|
msgpack
|
|
];
|
|
# codeflare-sdk is not packaged in nixpkgs
|
|
ray = [
|
|
msgpack
|
|
psutil
|
|
ray
|
|
]
|
|
++ ray.optional-dependencies.serve
|
|
++ redis.optional-dependencies.hiredis;
|
|
gdrive = [
|
|
google-api-python-client
|
|
google-auth-oauthlib
|
|
];
|
|
gcloudstorage = [
|
|
google-cloud-storage
|
|
];
|
|
azure = [
|
|
azure-storage-blob
|
|
];
|
|
opensearch = [
|
|
opensearch-py
|
|
];
|
|
kafka = [
|
|
confluent-kafka
|
|
];
|
|
# astrapy (astradb), office365-rest-python-client (sharepoint) and boxsdk (box)
|
|
# are not packaged in nixpkgs
|
|
};
|
|
|
|
pythonImportsCheck = [
|
|
"docling"
|
|
"docling_jobkit"
|
|
];
|
|
|
|
nativeCheckInputs = [
|
|
aiohttp
|
|
pytest-asyncio
|
|
pytestCheckHook
|
|
ray
|
|
redisTestHook
|
|
starlette
|
|
writableTmpDirAsHomeHook
|
|
]
|
|
++ lib.concatAttrValues finalAttrs.passthru.optional-dependencies;
|
|
|
|
pytestFlags = [
|
|
# Upstream's addopts stop the run after 5 failures
|
|
"--maxfail=0"
|
|
];
|
|
|
|
disabledTests = [
|
|
# requires network access
|
|
"test_chunk_file"
|
|
"test_chunker_manager_shared_across_workers"
|
|
"test_clear_converters_clears_caches"
|
|
"test_convert_file"
|
|
"test_convert_url"
|
|
"test_convert_warmup"
|
|
"test_convert_with_callbacks"
|
|
"test_replicated_convert"
|
|
"test_threaded_request_reaches_conversion"
|
|
|
|
# Expect the astradb, box and sharepoint connectors to be registered, but their
|
|
# dependencies are not packaged in nixpkgs
|
|
"test_builtin_source_connectors_registered"
|
|
"test_builtin_target_connectors_registered"
|
|
]
|
|
++ lib.optionals stdenv.hostPlatform.isDarwin [
|
|
# Flaky due to comparison with magic object
|
|
# https://github.com/docling-project/docling-jobkit/issues/45
|
|
"test_options_validator"
|
|
];
|
|
|
|
disabledTestPaths = [
|
|
# astrapy, boxsdk and office365-rest-python-client are not packaged in nixpkgs
|
|
"tests/test_astradb_target_processor.py"
|
|
"tests/test_box_helper.py"
|
|
"tests/test_box_source_processor.py"
|
|
"tests/test_box_target_processor.py"
|
|
"tests/test_sharepoint_coords.py"
|
|
"tests/test_sharepoint_helper.py"
|
|
"tests/test_sharepoint_source_processor.py"
|
|
"tests/test_sharepoint_target_processor.py"
|
|
];
|
|
|
|
__darwinAllowLocalNetworking = true;
|
|
|
|
meta = {
|
|
description = "Running a distributed job processing documents with Docling";
|
|
homepage = "https://github.com/docling-project/docling-jobkit";
|
|
changelog = "https://github.com/docling-project/docling-jobkit/blob/${finalAttrs.src.tag}/CHANGELOG.md";
|
|
license = lib.licenses.mit;
|
|
maintainers = with lib.maintainers; [ codgician ];
|
|
};
|
|
})
|