Skip to content

Commit 9e08994

Browse files
committed
fix(.extract-worker): bump extract-python=0.5.10
1 parent 53dc76b commit 9e08994

5 files changed

Lines changed: 44 additions & 27 deletions

File tree

workers/extract-worker/extract_worker/activities.py

Lines changed: 8 additions & 2 deletions
Original file line numberDiff line numberDiff line change
@@ -19,8 +19,14 @@
1919
to_raw_progress,
2020
write_artifact,
2121
)
22-
from extract_core import BasePipelineConfig, Pipeline, PipelineConfig
23-
from extract_core.objects import InputDoc, OutputFormat, SupportedExt
22+
from extract_core import (
23+
BasePipelineConfig,
24+
InputDoc,
25+
OutputFormat,
26+
Pipeline,
27+
PipelineConfig,
28+
SupportedExt,
29+
)
2430
from icij_common.es import (
2531
DOC_CONTENT_TYPE,
2632
DOC_LANGUAGE,

workers/extract-worker/extract_worker/config.py

Lines changed: 12 additions & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -1,16 +1,27 @@
1-
from datashare_python.config import WorkerConfig
1+
import datashare_python
2+
import extract_core
3+
import extract_python
4+
from datashare_python.config import LoggingConfig, WorkerConfig
25
from datashare_python.objects import DatashareModel
36
from pydantic import Field
47

58
from .constants import TorchDevice
69

10+
loggers = {
11+
datashare_python.__name__: "INFO",
12+
extract_python.__name__: "INFO",
13+
extract_core.__name__: "INFO",
14+
}
15+
_DEFAULT_LOGGING_CONFIG = LoggingConfig(loggers=loggers)
16+
717

818
class MarkdownExtractConfig(DatashareModel):
919
target_n_pages_per_batch: int = 100
1020

1121

1222
class ExtractWorkerConfig(WorkerConfig):
1323
device: TorchDevice = Field(default=TorchDevice.CPU, frozen=True)
24+
logging: LoggingConfig = _DEFAULT_LOGGING_CONFIG
1425

1526
markdown: MarkdownExtractConfig = Field(default_factory=MarkdownExtractConfig)
1627

workers/extract-worker/pyproject.toml

Lines changed: 4 additions & 4 deletions
Original file line numberDiff line numberDiff line change
@@ -10,13 +10,13 @@ readme = "README.md"
1010
requires-python = ">=3.11,<3.14"
1111
dependencies = [
1212
"datashare-python~=0.8.20",
13-
"extract-core==0.5.6",
13+
"extract-core==0.5.10",
1414
"temporalio==1.23.0",
1515
]
1616

1717
[project.optional-dependencies]
1818
base = [
19-
"extract-python[docling,marker]==0.5.9",
19+
"extract-python[docling,marker]==0.5.11",
2020
]
2121
cpu = [
2222
"torch==2.11.0",
@@ -28,7 +28,7 @@ gpu = [
2828
"torchvision==0.26.0+cu129; sys_platform == 'linux'",
2929
]
3030
mineru = [
31-
"extract-python[mineru]==0.5.9",
31+
"extract-python[mineru]==0.5.11",
3232
]
3333

3434
[project.entry-points."datashare.workflows"]
@@ -45,7 +45,7 @@ worker_config_cls = "extract_worker.config:WORKER_CONFIG_CLS"
4545

4646
[dependency-groups]
4747
dev = [
48-
"extract-python[docling]==0.5.9",
48+
"extract-python[docling]==0.5.11",
4949
"black>=26.1.0",
5050
"nest-asyncio>=1.6.0",
5151
"pre-commit>=4.5.1",

workers/extract-worker/uv.dist.lock

Lines changed: 10 additions & 10 deletions
Original file line numberDiff line numberDiff line change
@@ -625,9 +625,9 @@ dev = [
625625
requires-dist = [
626626
{ name = "cuda-bindings", marker = "sys_platform == 'linux' and extra == 'gpu'", specifier = "==12.9.4" },
627627
{ name = "datashare-python", specifier = "~=0.8.20" },
628-
{ name = "extract-core", specifier = "==0.5.6" },
629-
{ name = "extract-python", extras = ["docling", "marker"], marker = "extra == 'base'", specifier = "==0.5.9" },
630-
{ name = "extract-python", extras = ["mineru"], marker = "extra == 'mineru'", specifier = "==0.5.9" },
628+
{ name = "extract-core", specifier = "==0.5.10" },
629+
{ name = "extract-python", extras = ["docling", "marker"], marker = "extra == 'base'", specifier = "==0.5.11" },
630+
{ name = "extract-python", extras = ["mineru"], marker = "extra == 'mineru'", specifier = "==0.5.11" },
631631
{ name = "temporalio", specifier = "==1.23.0" },
632632
{ name = "torch", marker = "sys_platform == 'linux' and extra == 'gpu'", specifier = "==2.11.0+cu129", index = "https://download.pytorch.org/whl/cu129", conflict = { package = "datashare-extract-worker", extra = "gpu" } },
633633
{ name = "torch", marker = "extra == 'cpu'", specifier = "==2.11.0", index = "https://download.pytorch.org/whl/cpu", conflict = { package = "datashare-extract-worker", extra = "cpu" } },
@@ -639,7 +639,7 @@ provides-extras = ["base", "cpu", "gpu", "mineru"]
639639
[package.metadata.requires-dev]
640640
dev = [
641641
{ name = "black", specifier = ">=26.1.0" },
642-
{ name = "extract-python", extras = ["docling"], specifier = "==0.5.9" },
642+
{ name = "extract-python", extras = ["docling"], specifier = "==0.5.11" },
643643
{ name = "nest-asyncio", specifier = ">=1.6.0" },
644644
{ name = "pre-commit", specifier = ">=4.5.1" },
645645
{ name = "psutil", specifier = ">=6.1.0" },
@@ -928,7 +928,7 @@ wheels = [
928928

929929
[[package]]
930930
name = "extract-core"
931-
version = "0.5.6"
931+
version = "0.5.10"
932932
source = { registry = "https://pypi.org/simple" }
933933
dependencies = [
934934
{ name = "docling-slim", marker = "(platform_machine == 'x86_64' and sys_platform == 'linux') or (platform_machine != 'x86_64' and extra == 'extra-24-datashare-extract-worker-base' and extra == 'extra-24-datashare-extract-worker-mineru') or (platform_machine != 'x86_64' and extra == 'extra-24-datashare-extract-worker-cpu' and extra == 'extra-24-datashare-extract-worker-gpu') or sys_platform == 'darwin' or (sys_platform != 'linux' and extra == 'extra-24-datashare-extract-worker-base' and extra == 'extra-24-datashare-extract-worker-mineru') or (sys_platform != 'linux' and extra == 'extra-24-datashare-extract-worker-cpu' and extra == 'extra-24-datashare-extract-worker-gpu')" },
@@ -938,22 +938,22 @@ dependencies = [
938938
{ name = "pydantic", marker = "(platform_machine == 'x86_64' and sys_platform == 'linux') or (platform_machine != 'x86_64' and extra == 'extra-24-datashare-extract-worker-base' and extra == 'extra-24-datashare-extract-worker-mineru') or (platform_machine != 'x86_64' and extra == 'extra-24-datashare-extract-worker-cpu' and extra == 'extra-24-datashare-extract-worker-gpu') or sys_platform == 'darwin' or (sys_platform != 'linux' and extra == 'extra-24-datashare-extract-worker-base' and extra == 'extra-24-datashare-extract-worker-mineru') or (sys_platform != 'linux' and extra == 'extra-24-datashare-extract-worker-cpu' and extra == 'extra-24-datashare-extract-worker-gpu')" },
939939
{ name = "pydantic-extra-types", extra = ["pycountry"], marker = "(platform_machine == 'x86_64' and sys_platform == 'linux') or (platform_machine != 'x86_64' and extra == 'extra-24-datashare-extract-worker-base' and extra == 'extra-24-datashare-extract-worker-mineru') or (platform_machine != 'x86_64' and extra == 'extra-24-datashare-extract-worker-cpu' and extra == 'extra-24-datashare-extract-worker-gpu') or sys_platform == 'darwin' or (sys_platform != 'linux' and extra == 'extra-24-datashare-extract-worker-base' and extra == 'extra-24-datashare-extract-worker-mineru') or (sys_platform != 'linux' and extra == 'extra-24-datashare-extract-worker-cpu' and extra == 'extra-24-datashare-extract-worker-gpu')" },
940940
]
941-
sdist = { url = "https://files.pythonhosted.org/packages/70/ce/f479d3474e66e998762b7410d7dbc9108aec9990b182839db7a88fe79d68/extract_core-0.5.6.tar.gz", hash = "sha256:6efce5cff7638e6bf148a9888cafe9fad11478d4de01d7be33ccc1607c416f9a", size = 138009, upload-time = "2026-06-10T16:38:24.396Z" }
941+
sdist = { url = "https://files.pythonhosted.org/packages/5c/9a/490c923bffe01c650ff09513d9f1779c9020309c0ee0f7662bb46bccd3bc/extract_core-0.5.10.tar.gz", hash = "sha256:58c5f1106b86000379d9926639c5e751997a3ccdd4dc754283e7a2cd73acb3f0", size = 138212, upload-time = "2026-06-18T14:06:20.025Z" }
942942
wheels = [
943-
{ url = "https://files.pythonhosted.org/packages/17/64/b471f49c52feb61b9420a0d84fc3a3e6dfa8933a58f2aa99e14705e57840/extract_core-0.5.6-py3-none-any.whl", hash = "sha256:0d1ff4be6e31737443a36a8eb4096f722f0bc3fbdd0364f2d2a5f24b0ead23e8", size = 9977, upload-time = "2026-06-10T16:38:23.523Z" },
943+
{ url = "https://files.pythonhosted.org/packages/14/53/f5c96c7735162b9397dd298b140208e129ebe7cdc23f8dfabb2bc662b6cd/extract_core-0.5.10-py3-none-any.whl", hash = "sha256:c736bbb4168246e4116f954241467c83ec03bce3428088e7fcd3ef9f63c09342", size = 10213, upload-time = "2026-06-18T14:06:20.833Z" },
944944
]
945945

946946
[[package]]
947947
name = "extract-python"
948-
version = "0.5.9"
948+
version = "0.5.11"
949949
source = { registry = "https://pypi.org/simple" }
950950
dependencies = [
951951
{ name = "extract-core", marker = "(platform_machine == 'x86_64' and sys_platform == 'linux') or (platform_machine != 'x86_64' and extra == 'extra-24-datashare-extract-worker-base' and extra == 'extra-24-datashare-extract-worker-mineru') or (platform_machine != 'x86_64' and extra == 'extra-24-datashare-extract-worker-cpu' and extra == 'extra-24-datashare-extract-worker-gpu') or sys_platform == 'darwin' or (sys_platform != 'linux' and extra == 'extra-24-datashare-extract-worker-base' and extra == 'extra-24-datashare-extract-worker-mineru') or (sys_platform != 'linux' and extra == 'extra-24-datashare-extract-worker-cpu' and extra == 'extra-24-datashare-extract-worker-gpu')" },
952952
{ name = "icij-common", marker = "(platform_machine == 'x86_64' and sys_platform == 'linux') or (platform_machine != 'x86_64' and extra == 'extra-24-datashare-extract-worker-base' and extra == 'extra-24-datashare-extract-worker-mineru') or (platform_machine != 'x86_64' and extra == 'extra-24-datashare-extract-worker-cpu' and extra == 'extra-24-datashare-extract-worker-gpu') or sys_platform == 'darwin' or (sys_platform != 'linux' and extra == 'extra-24-datashare-extract-worker-base' and extra == 'extra-24-datashare-extract-worker-mineru') or (sys_platform != 'linux' and extra == 'extra-24-datashare-extract-worker-cpu' and extra == 'extra-24-datashare-extract-worker-gpu')" },
953953
]
954-
sdist = { url = "https://files.pythonhosted.org/packages/21/21/7dab87894d407a6cbc378bcf528dfd7b0b5646aa1dc5b9d6039e72c88d81/extract_python-0.5.9.tar.gz", hash = "sha256:4a9a94872003541682f6ffb63bf8199c9574b1f35ac5e238da40536e7397db8a", size = 251636, upload-time = "2026-06-17T15:42:34.775Z" }
954+
sdist = { url = "https://files.pythonhosted.org/packages/2e/f0/188d48b603ce9bf5c1f3cbebd667ece1a7658301a648fa9e8477a885e3c5/extract_python-0.5.11.tar.gz", hash = "sha256:caaf1540bd864d4f5a6215587f533d1a53cd85b79e103f316d308ce748c1655b", size = 251694, upload-time = "2026-06-18T14:06:17.417Z" }
955955
wheels = [
956-
{ url = "https://files.pythonhosted.org/packages/0b/61/b7c36a2eca1fb1895f4e124873c3745b73c66412e002833505c7b1f419d2/extract_python-0.5.9-py3-none-any.whl", hash = "sha256:cc32eb31863dda95944d65dab89a8bb71c9f037bd890f28f192380dcc9de1828", size = 9681, upload-time = "2026-06-17T15:42:33.919Z" },
956+
{ url = "https://files.pythonhosted.org/packages/90/a8/e549f12d25a2f272ccf8e75a36dd81aaf2c68fd65dd5c9db29047049405b/extract_python-0.5.11-py3-none-any.whl", hash = "sha256:f3713d3173f7d7a59d8f064379a5da0eae70bd51eab394d86c322e851ab00f95", size = 9755, upload-time = "2026-06-18T14:06:18.631Z" },
957957
]
958958

959959
[package.optional-dependencies]

workers/extract-worker/uv.lock

Lines changed: 10 additions & 10 deletions
Some generated files are not rendered by default. Learn more about customizing how changed files appear on GitHub.

0 commit comments

Comments
 (0)