datachain 0.7.11__py3-none-any.whl → 0.8.1__py3-none-any.whl

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.

Potentially problematic release.

This version of datachain might be problematic. Click here for more details.

Files changed (27) hide show
  1. datachain/catalog/catalog.py +56 -45
  2. datachain/cli.py +25 -3
  3. datachain/client/gcs.py +9 -0
  4. datachain/data_storage/sqlite.py +20 -6
  5. datachain/data_storage/warehouse.py +0 -1
  6. datachain/lib/arrow.py +82 -58
  7. datachain/lib/dc.py +167 -166
  8. datachain/lib/diff.py +197 -0
  9. datachain/lib/file.py +3 -1
  10. datachain/lib/listing.py +44 -0
  11. datachain/lib/meta_formats.py +38 -42
  12. datachain/lib/udf.py +0 -1
  13. datachain/query/batch.py +32 -6
  14. datachain/query/dataset.py +18 -17
  15. datachain/query/dispatch.py +125 -125
  16. datachain/query/session.py +8 -5
  17. datachain/query/udf.py +20 -0
  18. datachain/query/utils.py +42 -0
  19. datachain/remote/studio.py +53 -1
  20. datachain/studio.py +47 -2
  21. datachain/utils.py +1 -1
  22. {datachain-0.7.11.dist-info → datachain-0.8.1.dist-info}/METADATA +4 -3
  23. {datachain-0.7.11.dist-info → datachain-0.8.1.dist-info}/RECORD +27 -24
  24. {datachain-0.7.11.dist-info → datachain-0.8.1.dist-info}/LICENSE +0 -0
  25. {datachain-0.7.11.dist-info → datachain-0.8.1.dist-info}/WHEEL +0 -0
  26. {datachain-0.7.11.dist-info → datachain-0.8.1.dist-info}/entry_points.txt +0 -0
  27. {datachain-0.7.11.dist-info → datachain-0.8.1.dist-info}/top_level.txt +0 -0
{datachain-0.7.11.dist-info → datachain-0.8.1.dist-info}/METADATA RENAMED
@@ -1,6 +1,6 @@
1
1
  Metadata-Version: 2.1
2
2
  Name: datachain
3
- Version: 0.7.11
3
+ Version: 0.8.1
4
4
  Summary: Wrangle unstructured AI data at scale
5
5
  Author-email: Dmitry Petrov <support@dvc.org>
6
6
  License: Apache-2.0
@@ -46,6 +46,7 @@ Requires-Dist: iterative-telemetry>=0.0.9
46
46
  Requires-Dist: platformdirs
47
47
  Requires-Dist: dvc-studio-client<1,>=0.21
48
48
  Requires-Dist: tabulate
49
+ Requires-Dist: websockets
49
50
  Provides-Extra: docs
50
51
  Requires-Dist: mkdocs>=1.5.2; extra == "docs"
51
52
  Requires-Dist: mkdocs-gen-files>=0.5.0; extra == "docs"
@@ -83,7 +84,7 @@ Requires-Dist: requests-mock; extra == "tests"
83
84
  Requires-Dist: scipy; extra == "tests"
84
85
  Provides-Extra: dev
85
86
  Requires-Dist: datachain[docs,tests]; extra == "dev"
86
- Requires-Dist: mypy==1.13.0; extra == "dev"
87
+ Requires-Dist: mypy==1.14.0; extra == "dev"
87
88
  Requires-Dist: types-python-dateutil; extra == "dev"
88
89
  Requires-Dist: types-pytz; extra == "dev"
89
90
  Requires-Dist: types-PyYAML; extra == "dev"
@@ -98,7 +99,7 @@ Requires-Dist: unstructured[pdf]; extra == "examples"
98
99
  Requires-Dist: pdfplumber==0.11.4; extra == "examples"
99
100
  Requires-Dist: huggingface_hub[hf_transfer]; extra == "examples"
100
101
  Requires-Dist: onnx==1.16.1; extra == "examples"
101
- Requires-Dist: ultralytics==8.3.48; extra == "examples"
102
+ Requires-Dist: ultralytics==8.3.53; extra == "examples"
102
103
 
103
104
  ================
104
105
  |logo| DataChain
{datachain-0.7.11.dist-info → datachain-0.8.1.dist-info}/RECORD RENAMED
@@ -2,7 +2,7 @@ datachain/__init__.py,sha256=ofPJ6B-d-ybSDRrE7J6wqF_ZRAB2W9U8l-eeuBtqPLg,865
2
2
  datachain/__main__.py,sha256=hG3Y4ARGEqe1AWwNMd259rBlqtphx1Wk39YbueQ0yV8,91
3
3
  datachain/asyn.py,sha256=5aKrjnUxk0mtnZeFKNJd1DCE0MsnSoyJBZkr0y9H_a0,9313
4
4
  datachain/cache.py,sha256=s0YHN7qurmQv-eC265TjeureK84TebWWAnL07cxchZQ,2997
5
- datachain/cli.py,sha256=wQiYQ_qSVCGvS06pkknT9_FIBdFRzBdeRusW9uXE3vQ,42505
5
+ datachain/cli.py,sha256=gNXVoMfKINUhKjOpYN48tpyNBK13M0hkQWqra4jNSJQ,43137
6
6
  datachain/cli_utils.py,sha256=jrn9ejGXjybeO1ur3fjdSiAyCHZrX0qsLLbJzN9ErPM,2418
7
7
  datachain/config.py,sha256=g8qbNV0vW2VEKpX-dGZ9pAn0DAz6G2ZFcr7SAV3PoSM,4272
8
8
  datachain/dataset.py,sha256=P-pDBgvPqJGDhq_I7fwCfb6hY8E8mIAO8Q0NT7SNlNE,19128
@@ -14,18 +14,18 @@ datachain/nodes_fetcher.py,sha256=ILMzUW5o4_6lUOVrLDC9gJPCXfcgKnMG68plrc7dAOA,11
14
14
  datachain/nodes_thread_pool.py,sha256=uPo-xl8zG5m9YgODjPFBpbcqqHjI-dcxH87yAbj_qco,3192
15
15
  datachain/progress.py,sha256=5KotcvvzAUL_RF0GEj4JY0IB1lyImnmHxe89YkT1XO4,4330
16
16
  datachain/py.typed,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
17
- datachain/studio.py,sha256=Hr0Ha0kou0so4i8i-gWiXC1AYlJ2arI1D55cc7mi3tg,7253
17
+ datachain/studio.py,sha256=BegIXunW1n-sZtHSe3a30Mw2MXexVGRn_GU-OzjRRKM,8725
18
18
  datachain/telemetry.py,sha256=0A4IOPPp9VlP5pyW9eBfaTK3YhHGzHl7dQudQjUAx9A,994
19
- datachain/utils.py,sha256=-mSFowjIidJ4_sMXInvNHLn4rK_QnHuIlLuH1_lMGmI,13897
19
+ datachain/utils.py,sha256=9iP8RGm3MHemj13qt1VxRGYAsA6v-627M22o0fr76_M,13906
20
20
  datachain/catalog/__init__.py,sha256=g2iAAFx_gEIrqshXlhSEbrc8qDaEH11cjU40n3CHDz4,409
21
- datachain/catalog/catalog.py,sha256=s4fat0jjP3JPq0RGQ9zfzRkX1JavxxCrcB1tJKMgsks,57686
21
+ datachain/catalog/catalog.py,sha256=CacK-tfgMM-ZpE0cW7Rfosx1aqXV0shyUy0TfHZnBOQ,58385
22
22
  datachain/catalog/datasource.py,sha256=D-VWIVDCM10A8sQavLhRXdYSCG7F4o4ifswEF80_NAQ,1412
23
23
  datachain/catalog/loader.py,sha256=HA_mBC7q_My8j2WnSvIjUGuJpl6SIdg5vvy_lagxJlA,5733
24
24
  datachain/client/__init__.py,sha256=1kDpCPoibMXi1gExR4lTLc5pi-k6M5TANiwtXkPoLhU,49
25
25
  datachain/client/azure.py,sha256=ffxs26zm6KLAL1aUWJm-vtzuZP3LSNha7UDGXynMBKo,2234
26
26
  datachain/client/fileslice.py,sha256=bT7TYco1Qe3bqoc8aUkUZcPdPofJDHlryL5BsTn9xsY,3021
27
27
  datachain/client/fsspec.py,sha256=kf1blSGNcEXJ0tra3y5i35jc1aAy-67wMHXkqjlRMXg,12736
28
- datachain/client/gcs.py,sha256=cnTIr5GS6dbYOEYfqehhyQu3dr6XNjPHSg5U3FkivUk,4124
28
+ datachain/client/gcs.py,sha256=41Q9hxE9CYt-31B9m-ogqrEv4NbF2SWF-j8N_AC3VQ0,4580
29
29
  datachain/client/hf.py,sha256=XeVJVbiNViZCpn3sfb90Fr8SYO3BdLmfE3hOWMoqInE,951
30
30
  datachain/client/local.py,sha256=f2HBqWH8SQM5CyiJ0ljfePVROg2FszWaAn6E2c8RiLE,4596
31
31
  datachain/client/s3.py,sha256=CVHBUZ1Ic2Q3370nl-Bbe69phuWjFlrVv9dTJKBpRT0,6019
@@ -35,8 +35,8 @@ datachain/data_storage/job.py,sha256=w-7spowjkOa1P5fUVtJou3OltT0L48P0RYWZ9rSJ9-s
35
35
  datachain/data_storage/metastore.py,sha256=hfTITcesE9XlUTxcCcdDyWGGep-QSjJL9DUxko5QCeI,37524
36
36
  datachain/data_storage/schema.py,sha256=-QVlRvD0dfu-ZFUxylEoSnLJLnleMEjVlcAb2OGu-AY,9895
37
37
  datachain/data_storage/serializer.py,sha256=6G2YtOFqqDzJf1KbvZraKGXl2XHZyVml2krunWUum5o,927
38
- datachain/data_storage/sqlite.py,sha256=D_ZQ0PHmZzHO2dinv4naVJocUDIZUwV4WAz692C1cyk,22521
39
- datachain/data_storage/warehouse.py,sha256=tjIkU-5JywBR0apCyqTcwSyaRtGxhu2L7IVjrz-55uc,30802
38
+ datachain/data_storage/sqlite.py,sha256=iJv1QxwVifOowtYhIDqYVoea21dvkQIdxklGNIend3c,22961
39
+ datachain/data_storage/warehouse.py,sha256=CMW36ZNrAzoZ003xGkdnIxC7S0PnQxC2588iUMKI9SM,30779
40
40
  datachain/func/__init__.py,sha256=TG6JHFKtLi06Nd5iLszXIflEq-VKZcKMdgo_KiQ8SGQ,1055
41
41
  datachain/func/aggregate.py,sha256=7_IPrIwb2XSs3zG4iOr1eTvzn6kNVe2mkzvNzjusDHk,10942
42
42
  datachain/func/array.py,sha256=zHDNWuWLA7HVa9FEvQeHhVi00_xqenyleTqcLwkXWBI,5477
@@ -49,24 +49,25 @@ datachain/func/random.py,sha256=pENOLj9rSmWfGCnOsUIaCsVC5486zQb66qfQvXaz9Z4,452
49
49
  datachain/func/string.py,sha256=8az3BTeezlaZt6NW-54GWX7WSosAOVMbTr6bXIYyJq4,5958
50
50
  datachain/func/window.py,sha256=0MB1yjpVbwOrl_WNLZ8V3jkJz3o0XlYinpAcZQJuxiA,1688
51
51
  datachain/lib/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
52
- datachain/lib/arrow.py,sha256=b5efxAUaNNYVwtXVJqj07D3zf5KC-BPlLCxKEZbEG6w,9429
52
+ datachain/lib/arrow.py,sha256=33Od_XECCfWR9PUDBdevSooXS4mpMdPx_hoMLjpaELU,9734
53
53
  datachain/lib/clip.py,sha256=lm5CzVi4Cj1jVLEKvERKArb-egb9j1Ls-fwTItT6vlI,6150
54
54
  datachain/lib/data_model.py,sha256=zS4lmXHVBXc9ntcyea2a1CRLXGSAN_0glXcF88CohgY,2685
55
55
  datachain/lib/dataset_info.py,sha256=IjdF1E0TQNOq9YyynfWiCFTeZpbyGfyJvxgJY4YN810,2493
56
- datachain/lib/dc.py,sha256=qMhpVPdWeuXBDhmKKoq3fkq12Cx_ZPxDdsl_juu482o,89595
57
- datachain/lib/file.py,sha256=4dDWXVCHHP2uELDPHP_LheyTyyr01jwp5wp3HaOIeFI,15028
56
+ datachain/lib/dc.py,sha256=dLBvM_Fr26AQBD3t_-oLSymLyoJQxmkJQs3cC-mZCu8,90492
57
+ datachain/lib/diff.py,sha256=Yurzyi7PzZzY80HOnVTpwtbWzSJ1LqN8NgZWwZOh_UU,6732
58
+ datachain/lib/file.py,sha256=KeccxOulTQCLitdHZoTaq96xpE-5kmWZCrT9X9bRkD0,15049
58
59
  datachain/lib/hf.py,sha256=a-zFpDmZIR4r8dlNNTjfpAKSnuJ9xyRXlgcdENiXt3E,5864
59
60
  datachain/lib/image.py,sha256=AMXYwQsmarZjRbPCZY3M1jDsM2WAB_b3cTY4uOIuXNU,2675
60
- datachain/lib/listing.py,sha256=cVkCp7TRVpcZKSx-Bbk9t51bQI9Mw0o86W6ZPhAsuzM,3667
61
+ datachain/lib/listing.py,sha256=C-JmWMJcErbjNJ7ygwaXjP7Ak3nS-MjYTgSn7vvkICg,5536
61
62
  datachain/lib/listing_info.py,sha256=9ua40Hw0aiQByUw3oAEeNzMavJYfW0Uhe8YdCTK-m_g,1110
62
- datachain/lib/meta_formats.py,sha256=6_gB23fWlvd-edOO3UvDHvj6dBXVL61T7x8RX51FW84,6685
63
+ datachain/lib/meta_formats.py,sha256=hDPfEkcmiLZOjhBBXuareMdnq65Wj8vZvxjmum6cROM,6377
63
64
  datachain/lib/model_store.py,sha256=DNIv8Y6Jtk1_idNLzIpsThOsdW2BMAudyUCbPUcgcxk,2515
64
65
  datachain/lib/pytorch.py,sha256=dA3r1JY0wqV_907a1D0lFaEN-7v3fMRpc1ePFE9CnvA,6168
65
66
  datachain/lib/settings.py,sha256=ZELRCTLbi5vzRPiDX6cQ9LLg9TefJ_A05gIGni0lll8,2535
66
67
  datachain/lib/signal_schema.py,sha256=ziRTctom0-wAqURZfkfG6dc_3P2FcYxKjYsKC49NQ1Q,25415
67
68
  datachain/lib/tar.py,sha256=3WIzao6yD5fbLqXLTt9GhPGNonbFIs_fDRu-9vgLgsA,1038
68
69
  datachain/lib/text.py,sha256=UNHm8fhidk7wdrWqacEWaA6I9ykfYqarQ2URby7jc7M,1261
69
- datachain/lib/udf.py,sha256=-j0krjNAELTqRI0dB1N65AmawtcIY5vN---AuUcW8Us,13637
70
+ datachain/lib/udf.py,sha256=Rsf_6mN6qCQVknl99yvi1guta1AMnG9MsoPn14ff5Fc,13609
70
71
  datachain/lib/udf_signature.py,sha256=GXw24A-Olna6DWCdgy2bC-gZh_gLGPQ-KvjuI6pUjC0,7281
71
72
  datachain/lib/utils.py,sha256=QrjVs_oLRXEotOPUYurBJypBFi_ReTJmxcnJeH4j2Uk,1596
72
73
  datachain/lib/vfile.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
@@ -87,16 +88,18 @@ datachain/model/ultralytics/bbox.py,sha256=OZ9XBdyMOYc401P-RhfSN9QaYvMpnx2Phu9pt
87
88
  datachain/model/ultralytics/pose.py,sha256=71KBTcoST2wcEtsyGXqLVpvUtqbp9gwZGA15pEPtX5A,2959
88
89
  datachain/model/ultralytics/segment.py,sha256=Z1ab0tZRJubSYNH4KkFlzhYeGNTfAyC71KmkQcToHDQ,2760
89
90
  datachain/query/__init__.py,sha256=7DhEIjAA8uZJfejruAVMZVcGFmvUpffuZJwgRqNwe-c,263
90
- datachain/query/batch.py,sha256=5fEhORFe7li12SdYddaSK3LyqksMfCHhwN1_A6TfsA4,3485
91
- datachain/query/dataset.py,sha256=JrImhguXj2ZDwJpfuyhcgxSIlqSPy5NmLDLc3muFQJs,54610
92
- datachain/query/dispatch.py,sha256=fZ0TgGFRcsrYh1iXQoZVjkUl4Xetom9PSHoeDes3IRs,11606
91
+ datachain/query/batch.py,sha256=6w8gzLTmLeylststu-gT5jIqEfi4-djS7_yTYyeo-fw,4190
92
+ datachain/query/dataset.py,sha256=1wJuiFgXgtYarJAgLmgQESrvp0ayIQbJUiCZyAe0NkU,54774
93
+ datachain/query/dispatch.py,sha256=Uw4highKfs70ioEddSK2zknjpvz_q59OHc8s43nXa_I,12004
93
94
  datachain/query/metrics.py,sha256=r5b0ygYhokbXp8Mg3kCH8iFSRw0jxzyeBe-C-J_bKFc,938
94
95
  datachain/query/params.py,sha256=O_j89mjYRLOwWNhYZl-z7mi-rkdP7WyFmaDufsdTryE,863
95
96
  datachain/query/queue.py,sha256=waqM_KzavU8C-G95-4211Nd4GXna_u2747Chgwtgz2w,3839
96
97
  datachain/query/schema.py,sha256=b_KnVy6B26Ol4nYG0LqNNpeQ1QYPk95YRGUjXfdaQWs,6606
97
- datachain/query/session.py,sha256=vvLIJ5b8eElovHLAWq_CZJXmN5t7C7iAZA7x9wPPOms,5905
98
+ datachain/query/session.py,sha256=09FtSS3cvfRv2iBQZcbCBMIiRywv7Guhy5nmLEiltq4,5998
99
+ datachain/query/udf.py,sha256=GY8E9pnzPE7ZKl_jvetZpn9R2rlUtMlhoYj4UmrzFzw,594
100
+ datachain/query/utils.py,sha256=u0A_BwG9PNs0DxoDcvSWgWLpj3ByTUv8CqH13CIuGag,1293
98
101
  datachain/remote/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
99
- datachain/remote/studio.py,sha256=WiK6fpRAw0a6Dth4XXI0YInEHH4gDU7AUHHDNd3wJzg,11616
102
+ datachain/remote/studio.py,sha256=3DlgESETzxm3dgb6zzjjGxsddSkacT2dARnteLAfMxQ,13366
100
103
  datachain/sql/__init__.py,sha256=6SQRdbljO3d2hx3EAVXEZrHQKv5jth0Jh98PogT59No,262
101
104
  datachain/sql/selectable.py,sha256=cTc60qVoAwqqss0Vop8Lt5Z-ROnM1XrQmL_GLjRxhXs,1765
102
105
  datachain/sql/types.py,sha256=ASSPkmM5EzdRindqj2O7WHLXq8VHAgFYedG8lYfGvVI,14045
@@ -118,9 +121,9 @@ datachain/sql/sqlite/vector.py,sha256=ncW4eu2FlJhrP_CIpsvtkUabZlQdl2D5Lgwy_cbfqR
118
121
  datachain/toolkit/__init__.py,sha256=eQ58Q5Yf_Fgv1ZG0IO5dpB4jmP90rk8YxUWmPc1M2Bo,68
119
122
  datachain/toolkit/split.py,sha256=z3zRJNzjWrpPuRw-zgFbCOBKInyYxJew8ygrYQRQLNc,2930
120
123
  datachain/torch/__init__.py,sha256=gIS74PoEPy4TB3X6vx9nLO0Y3sLJzsA8ckn8pRWihJM,579
121
- datachain-0.7.11.dist-info/LICENSE,sha256=8DnqK5yoPI_E50bEg_zsHKZHY2HqPy4rYN338BHQaRA,11344
122
- datachain-0.7.11.dist-info/METADATA,sha256=ADTTf0_eJImM-tIPR-jQydM3N9Iis-ECRxWgkwLM8lU,8412
123
- datachain-0.7.11.dist-info/WHEEL,sha256=PZUExdf71Ui_so67QXpySuHtCi3-J3wvF4ORK6k_S8U,91
124
- datachain-0.7.11.dist-info/entry_points.txt,sha256=0GMJS6B_KWq0m3VT98vQI2YZodAMkn4uReZ_okga9R4,49
125
- datachain-0.7.11.dist-info/top_level.txt,sha256=lZPpdU_2jJABLNIg2kvEOBi8PtsYikbN1OdMLHk8bTg,10
126
- datachain-0.7.11.dist-info/RECORD,,
124
+ datachain-0.8.1.dist-info/LICENSE,sha256=8DnqK5yoPI_E50bEg_zsHKZHY2HqPy4rYN338BHQaRA,11344
125
+ datachain-0.8.1.dist-info/METADATA,sha256=iZV5_BcFNOY82HkPBDJ6h2Y895wYy5UQfo_VJaDZN6Q,8437
126
+ datachain-0.8.1.dist-info/WHEEL,sha256=PZUExdf71Ui_so67QXpySuHtCi3-J3wvF4ORK6k_S8U,91
127
+ datachain-0.8.1.dist-info/entry_points.txt,sha256=0GMJS6B_KWq0m3VT98vQI2YZodAMkn4uReZ_okga9R4,49
128
+ datachain-0.8.1.dist-info/top_level.txt,sha256=lZPpdU_2jJABLNIg2kvEOBi8PtsYikbN1OdMLHk8bTg,10
129
+ datachain-0.8.1.dist-info/RECORD,,