unstructured-ingest 0.0.25__py3-none-any.whl → 0.1.1__py3-none-any.whl

{unstructured_ingest-0.0.25.dist-info → unstructured_ingest-0.1.1.dist-info}/RECORD

@@ -1,7 +1,51 @@
+test/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+test/integration/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+test/integration/utils.py,sha256=CWqzEGw6TA_ZoP9hRUkW64TWYssooBbufcTRmbJvod8,401
+test/integration/chunkers/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+test/integration/chunkers/test_chunkers.py,sha256=pqn1Rqh36jZTJL4qpU0iuOMFAEQ-LrKAPOgWtQMAt_I,1482
+test/integration/connectors/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+test/integration/connectors/conftest.py,sha256=Q8ScDzrzO2o-8D_kYFt8LL7QAhoFTRRtKJKMc2hLMcI,345
+test/integration/connectors/test_s3.py,sha256=fK0soCTkNxp-4hm4O2LPrhlZXvYmaeTmeEgeNh1b0k8,5839
+test/integration/connectors/databricks_tests/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+test/integration/connectors/databricks_tests/test_volumes_native.py,sha256=kS45mnNu9_U4qV3cxByEFXCYLEBWRy-fxxhzR3r93cs,5685
+test/integration/connectors/sql/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+test/integration/connectors/sql/test_postgres.py,sha256=A9vWj5pBdoEyL2m6d3e2Ep8ZZcnLhdXkaHPPlkTStbg,6581
+test/integration/connectors/sql/test_sqlite.py,sha256=F6Ljb6npmFZlq_5pvJj-0Hkk2mC3T-pMAGyhDm1UtM4,5702
+test/integration/connectors/utils/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+test/integration/connectors/utils/constants.py,sha256=0zSPnsZVqJuNhXduXvdXFQLZTRIQa5Fo_1qjBYVCfb8,209
+test/integration/connectors/utils/docker_compose.py,sha256=6XeYOKQFZCBRLEmcgH2mmBAaVs6R6jCWAhJLjq6p-aM,1771
+test/integration/connectors/utils/validation.py,sha256=Sf0ELATWG5K3E3d5S_ArtZeFFYdzoI5jN86U4DiqNyw,8422
+test/integration/embedders/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+test/integration/embedders/conftest.py,sha256=B2W771RbijR7G_GybsCzRyIvOzXqzbKZdRIlNDd5AGY,334
+test/integration/embedders/test_bedrock.py,sha256=0oBRNS_DtFDGQ22Z1T3t6VOJ31PrItgvnJpqcLe9Fg4,1903
+test/integration/embedders/test_huggingface.py,sha256=0mMTOO-Nh7KB70AGs_7LLQIxMYrnSPqyihriUeqACbM,1007
+test/integration/embedders/test_mixedbread.py,sha256=RrLv8SByMNXsgrlh94RbaT-VyxZ4-DILO-OPpmOwvSI,1441
+test/integration/embedders/test_octoai.py,sha256=LnR0BLttamW5PGid6jFxATDAi0x7hq5iWMXurbHP6TI,1328
+test/integration/embedders/test_openai.py,sha256=0jlFqEeeCneIWX9tGyC3TXeUNqsMXR7u5n7uEIaAQKo,1328
+test/integration/embedders/test_vertexai.py,sha256=OtoFzmrWWhGIO5Bbl5zt_4sp6qRHZxtaDQKpGcfzNLM,1345
+test/integration/embedders/test_voyageai.py,sha256=Zqf7nn1AxfBDBr5A9Jr-5pxes4QNvfKiyeGexCCm4nY,1346
+test/integration/embedders/togetherai.py,sha256=0W1ScD5yb1D9hPC2ewUsuCHLUOpCuM083YMBhqAI9fw,1395
+test/integration/embedders/utils.py,sha256=3AMKMBpgBep_0jFqrqMHH8BJo6w60kpouSZ5JPJTwIA,1850
+test/integration/partitioners/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+test/integration/partitioners/test_partitioner.py,sha256=KEpnhsz2YNAoQ2UZGOTsi1_uk1h4Vg-gGTsy5Fe9OCw,2846
+test/unit/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+test/unit/test_chunking_utils.py,sha256=0iPwfnMPpyTm-yOE0BXMnEQQP4iguS6NhOqgMQU5nhk,1390
+test/unit/test_error.py,sha256=RflmngCdFNKOLXVfLnUdNfY3Mfg3k7DTEzfIl0B-syU,840
+test/unit/test_interfaces.py,sha256=XNj8qasc1ltaeUv-2y31rv7R9xquo0rgRrMvBZoNZLw,9623
+test/unit/test_interfaces_v2.py,sha256=nyxUsRX1M6Mfhux7SqEhal85PIaWO5xhm6ZTcqpPpHI,790
+test/unit/test_logger.py,sha256=0SKndXE_VRd8XmUHkrj7zuBQHZscXx3ZQllMEOvtF9Y,2380
+test/unit/test_utils.py,sha256=xJ9WGpHBihWpQWvIzd6z99UIdZJba8U7c31h3q6C9To,4800
+test/unit/test_utils_v2.py,sha256=TWVAeE0OrcHgPyzGPtEnQakICsVrDeVhIKPMRQPX554,2638
+test/unit/embed/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+test/unit/embed/test_mixedbreadai.py,sha256=XFNJDP5pIgF3eQYwBiuEWmH3zZWx72Wpwyv-Q4m0DJg,1332
+test/unit/embed/test_octoai.py,sha256=Ha9EgAW64Q45hFj51tToe8RyKXWXwqAkdDqSFDMu37Q,831
+test/unit/embed/test_openai.py,sha256=0O1yshDcE0BMKv1yJqrNuiNLSdPhLpKqJ-D_wmnidsM,831
+test/unit/embed/test_vertexai.py,sha256=Pl7COc9E3tf_yGidkTEmTizNGyZF1F5zuL2TgPTMnfI,1048
+test/unit/embed/test_voyageai.py,sha256=DviCOJFhe5H4e26-kNyX3JNe8h3qB5Yl0KOe8rQEMrc,981
 unstructured_ingest/__init__.py,sha256=U4S_2y3zgLZVfMenHRaJFBW8yqh2mUBuI291LGQVOJ8,35
-unstructured_ingest/__version__.py,sha256=WG3ykkrrofptunFgyMVyh_5Uyla9d5aYDfBtMqyZ_lE,43
+unstructured_ingest/__version__.py,sha256=ch9Ch304-rlC6iFyomBT7OHb9bvtQNzaejmd5QwbzKE,42
 unstructured_ingest/error.py,sha256=qDncnJgbf5ils956RcO2CGlAKYDT5OaEM9Clv1JVTNc,1448
-unstructured_ingest/interfaces.py,sha256=0r0gQoHJQ4DVSQEVbUPBA3N6WyvGMkR1u6U2SwUvoAQ,31361
+unstructured_ingest/interfaces.py,sha256=m03BgenxSA34HbW157L7V9TGxK_dTG7N2AnAhF31W-U,31364
 unstructured_ingest/logger.py,sha256=S5nSqGcABoQyeicgRnBQFjDScCaTvFVivOCvbo-laL0,4479
 unstructured_ingest/main.py,sha256=82G_7eG4PNhc_xIqj4Y_sFbDV9VI-nwSfsfJQMzovMk,169
 unstructured_ingest/processor.py,sha256=XKKrvbxsb--5cDzz4hB3-GfWZYyIjJ2ah8FpzQKF_DM,2760
@@ -9,7 +53,7 @@ unstructured_ingest/cli/__init__.py,sha256=9kNcBOHuXON5lB1MJU9QewEhwPmId56vXqB29
 unstructured_ingest/cli/cli.py,sha256=LutBTBYMqboKw8cputHVszpenyfnySzcUC15ifwuYyg,1049
 unstructured_ingest/cli/cmd_factory.py,sha256=UdHm1KacTombpF6DxyTSwTCuApsKHUYw_kVu5Nhcy3Y,364
 unstructured_ingest/cli/common.py,sha256=I0El08FHz5kxw7iz0VWOWPrvcJD1rBgXJSwVIpVmmwU,204
-unstructured_ingest/cli/interfaces.py,sha256=nWZVXAoLEP08eDPj10c2nwHNbd-HXOHFa4YvEdUJ8y8,24084
+unstructured_ingest/cli/interfaces.py,sha256=lpaaOdAQ4NMsawVaHSk5lXCcZ0Mw85kRzfElu1ODCB0,24090
 unstructured_ingest/cli/utils.py,sha256=KNhkFNKOeEihc8HlvMz_MTbYVQNFklrBKbC8xg9h1xE,7982
 unstructured_ingest/cli/base/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 unstructured_ingest/cli/base/cmd.py,sha256=BbfjA2v203Jh-7DL6bzxQ7fOeNixd5BsBMuzXz6h5IQ,583
@@ -106,11 +150,11 @@ unstructured_ingest/connector/notion/connector.py,sha256=8A9d-Pej-uXzjEy85zUloxI
 unstructured_ingest/connector/notion/helpers.py,sha256=-eEB8eSqdD5bWX_QEA2hZz1siucC0FNEUEqCEJptiVk,20702
 unstructured_ingest/connector/notion/interfaces.py,sha256=SrTT-9c0nvk0fMqVgudYF647r04AdMKi6wkIkMy7Szw,563
 unstructured_ingest/connector/notion/types/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-unstructured_ingest/connector/notion/types/block.py,sha256=AKOY-o6CTFC-caWlkLfKskMuFemH4-Vdrhv7HnRkS8w,3009
-unstructured_ingest/connector/notion/types/database.py,sha256=UvrxuCd62wIYtgDKSkyGEBZHwfqvPKq1U3qr3w-zLAI,2551
+unstructured_ingest/connector/notion/types/block.py,sha256=w3j3F_z-50dpIpBt4Ib8_U4eINZRFMfGHdbE3hjkbu0,3028
+unstructured_ingest/connector/notion/types/database.py,sha256=1SrP5sxWhif8dxCteXnJAFX2HwoXI2EJy9IRIzM_nGM,2570
 unstructured_ingest/connector/notion/types/date.py,sha256=Ah0ekF18S_9xVDT2Ps1NGD1eOihtInGIYji_BDIalig,729
 unstructured_ingest/connector/notion/types/file.py,sha256=xc5UQ46qWvVd3SkKJctRBqMVERCNc_UVVc21pu66IME,1291
-unstructured_ingest/connector/notion/types/page.py,sha256=PR3xT7OdO31zHDpp3bhgc5GLdcFuk8F6jOqGlOu5xNg,1408
+unstructured_ingest/connector/notion/types/page.py,sha256=oZS5ausaiA68Ux-i6mOA0qYywP0X7YchXL8gWscMaxQ,1427
 unstructured_ingest/connector/notion/types/parent.py,sha256=VTNyL5JNVLb5AqR5P-c658DC9bUgkRKPA9fI2CFZWoU,1695
 unstructured_ingest/connector/notion/types/rich_text.py,sha256=V0fqXLAq7H5A6Av0IM8TqqhqW45VWD8K79sHdh1FyA8,5450
 unstructured_ingest/connector/notion/types/user.py,sha256=VK-XYFt2WdtEqm_LGnmE22ms7xw84ia3pSBhpmy2IKg,1800
@@ -167,14 +211,15 @@ unstructured_ingest/connector/notion/types/database_properties/unique_id.py,sha2
 unstructured_ingest/connector/notion/types/database_properties/url.py,sha256=iXQ2tVUm9UlKVtDA0NQiFIRJ5PHYW9wOaWt2vFfSVCg,862
 unstructured_ingest/connector/notion/types/database_properties/verification.py,sha256=J_DLjY-v2T6xDGMQ7FkI0YMKMA6SG6Y3yYW7qUD1hKA,2334
 unstructured_ingest/embed/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-unstructured_ingest/embed/bedrock.py,sha256=5-pKWwOEGHKOHa06wYuKOhvT8Xu72ke6nrpCnRtkAaU,3872
-unstructured_ingest/embed/huggingface.py,sha256=ku_JQr72KBG8n5b6KRkXIbeBGzdgLw_KKIEm1dFK3oM,2729
-unstructured_ingest/embed/interfaces.py,sha256=L5WimR69bmEvliIBlZ8wOCH_YDA9DWteCu6QEsKCV5I,1113
-unstructured_ingest/embed/mixedbreadai.py,sha256=NSrAt1_bjphTHLUnlzzWSBU25UBCZlpYaLdWSRSGyqs,5504
-unstructured_ingest/embed/octoai.py,sha256=0zxAUAMzodGkqMwqMkEvSfgWLNHtEnhdvUofvJDQD1A,2368
-unstructured_ingest/embed/openai.py,sha256=4Ee4A2rQ8OlSh_yiJSFmok_qqRDi1A3KyayB5YiPLFw,2058
-unstructured_ingest/embed/vertexai.py,sha256=cgyRyTm_dO_qyedwbIhOQIFvKjCqZBoDh606ykzTYHI,3598
-unstructured_ingest/embed/voyageai.py,sha256=6BWNJUZOqkHSMaO2XPVZVYAVRrAtpMWQZEKp0qgp20Q,2631
+unstructured_ingest/embed/bedrock.py,sha256=-PRdZsF44vwi6G4G75gdO31AJKfZWClOXkJQAk7rEO8,3096
+unstructured_ingest/embed/huggingface.py,sha256=2cBiQhOhfWHX3hS-eKjocysOkUaRlyRfUj9Kxjrp6cE,1934
+unstructured_ingest/embed/interfaces.py,sha256=au4Xp8ciDvo4bidlUbazFW2aC7NZW5-UDLKXBFVzAX4,2025
+unstructured_ingest/embed/mixedbreadai.py,sha256=OwFWWukvkQaXhjgs6b6N6D4w7sYrtcHNhsHAj-Bocj4,4268
+unstructured_ingest/embed/octoai.py,sha256=jHytDfQgup0v1PBcmlMv1nIh9Obg8WGO5qtLmN-Ot5g,1473
+unstructured_ingest/embed/openai.py,sha256=JXo4boivNoo2lBzHuS4Z0FZ1zlgUGAPVt0X3HY540ZU,1282
+unstructured_ingest/embed/togetherai.py,sha256=BL7NzExSE-laQqrp4ybUgoZ9JG_eop4hk-s2yCO_d5c,1451
+unstructured_ingest/embed/vertexai.py,sha256=X5bGJdXyR5nAFH_ocAVgEowmd60nOBykyfclYo3VfBM,2808
+unstructured_ingest/embed/voyageai.py,sha256=bjom9QqWmH1Mv08ewg8ZG7gO3rQPMVS0_ztm2KBAOjI,1821
 unstructured_ingest/enhanced_dataclass/__init__.py,sha256=gDZOUsv5eo-8jm4Yu7DdDwi101aGbfG7JctTdOYnTOM,151
 unstructured_ingest/enhanced_dataclass/core.py,sha256=d6aUkDynuKX87cHx9_N5UDUWrvISR4jYRFRTvd_avlI,3038
 unstructured_ingest/enhanced_dataclass/dataclasses.py,sha256=aZMsoCzAGRb8Rmh3BTSBFtNr6FmFTY93KYGLk3gYJKQ,1949
@@ -276,11 +321,11 @@ unstructured_ingest/v2/cli/base/dest.py,sha256=zDjqek7anr0JQ2ptEl8KIAsUXuCuHRnBQ
 unstructured_ingest/v2/cli/base/importer.py,sha256=nRt0QQ3qpi264-n_mR0l55C2ddM8nowTNzT1jsWaam8,1128
 unstructured_ingest/v2/cli/base/src.py,sha256=cpQ43qQju4e5s_YSaPxUtA70BaisRkTBdjtlPhqn5Mg,2872
 unstructured_ingest/v2/cli/utils/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-unstructured_ingest/v2/cli/utils/click.py,sha256=Wn2s3PuvBCKB0lsK-W7X_Y0eYyWnS6Y9wWo1OhVBOzY,6344
+unstructured_ingest/v2/cli/utils/click.py,sha256=HCEcdHf8Lck0zcx3kidKjLbHDHXIBxPRL2MGgtKtDlg,6967
 unstructured_ingest/v2/cli/utils/model_conversion.py,sha256=uJQKpbTC5ysOdVaRq2SWEjG8btBimVZYzX9NVL7xnzs,7500
 unstructured_ingest/v2/interfaces/__init__.py,sha256=Rfa8crx6De7WNOK-EjsWWwFVpsUfCc6gY8B8tQ3ae9I,899
-unstructured_ingest/v2/interfaces/connector.py,sha256=KG0pHdAcpuO5h72xrAkJzADmjxbav31TZ2Wo3PBvwT0,765
-unstructured_ingest/v2/interfaces/downloader.py,sha256=PKT1kr79Mz1urW_8xCyq9sBuK93gDvyTXg5e4ma4htU,2871
+unstructured_ingest/v2/interfaces/connector.py,sha256=qUFFJ3qgDMenTCZMtVRjq1DIwsVak6pxNjQOH2eVkMw,1623
+unstructured_ingest/v2/interfaces/downloader.py,sha256=Lj3nTY1hPA71GfNeedFVCdHdZsHLle8qrx5RtXAy9GY,2940
 unstructured_ingest/v2/interfaces/file_data.py,sha256=ieJK-hqHCEOmoYNGoFbCHziSaZyMtRS9VpSoYbwoKCE,1944
 unstructured_ingest/v2/interfaces/indexer.py,sha256=Bd1S-gTLsxhJBLEh1lYm_gXqwQLaEZMoqPq9yGxtN_E,713
 unstructured_ingest/v2/interfaces/process.py,sha256=BgglTu5K93FnDDopZKKr_rkK2LTZOguR6kcQjKHjF40,392
@@ -302,13 +347,13 @@ unstructured_ingest/v2/pipeline/steps/stage.py,sha256=cphKgHScLz2rNLZRI5Olsb6dAH
 unstructured_ingest/v2/pipeline/steps/uncompress.py,sha256=CFSy4tGp6BAvF0oIwWFN8v4zFzh5pRDeESjEn5iP9hE,1756
 unstructured_ingest/v2/pipeline/steps/upload.py,sha256=zlgXgwReX9TBOdfTpS9hETah4SeSmzPB2g8dAGfLIvM,1987
 unstructured_ingest/v2/processes/__init__.py,sha256=FaHWSCGyc7GWVnAsNEUUj7L8hT8gCVY3_hUE2VzWtUg,462
-unstructured_ingest/v2/processes/chunker.py,sha256=76PrpCSd8k3DpfdZcl8I10u7vciKzhSV9ZByrrp302g,5476
+unstructured_ingest/v2/processes/chunker.py,sha256=1bfJ2qgl6qu2HvClzHbC7-q5QtUp7mrlNxZxnPGYTm0,5479
 unstructured_ingest/v2/processes/connector_registry.py,sha256=vkEe6jpgdYtZCxMj59s5atWGgmPuxAEXRUoTt-MJ7wc,2198
-unstructured_ingest/v2/processes/embedder.py,sha256=nFYiOmIJwWLodBt_cC-E5h7zmYB9t3hLu2BWtBStm3g,5977
+unstructured_ingest/v2/processes/embedder.py,sha256=PQn0IO8xbGRQHpcT2VVl-J8gTJ5HGGEP9gdEAwMVK3U,6498
 unstructured_ingest/v2/processes/filter.py,sha256=kjUmMw2SDq2bme0JCAOxs6cJriIG6Ty09KOznS-xz08,2145
-unstructured_ingest/v2/processes/partitioner.py,sha256=bpqmZDsKKi6qtxNWdIWBfQmr1ccQUhU0axecpGAUf_4,7739
+unstructured_ingest/v2/processes/partitioner.py,sha256=2Lhztd730soVC2TOqrn_ba7CGZna8AHHpqJY2ZUYVxE,7776
 unstructured_ingest/v2/processes/uncompress.py,sha256=Z_XfsITGdyaRwhtNUc7bMj5Y2jLuBge8KoK4nxhqKag,2425
-unstructured_ingest/v2/processes/connectors/__init__.py,sha256=XZWdbUKXioO4vfCYjgNNV4ZDNPQ_VrAUcHMjHGIys3E,5334
+unstructured_ingest/v2/processes/connectors/__init__.py,sha256=glyowqb93_NNreQXoRLbF0PvzMc6Ptv0ARfl3xfSH4E,4967
 unstructured_ingest/v2/processes/connectors/airtable.py,sha256=Yi7PEv_FejZ9_y3BPY3gu5YGVfeLh-9YX-qLyQHjJsY,8921
 unstructured_ingest/v2/processes/connectors/astradb.py,sha256=ZctZRfXcOAMBGPkKgHvhTmV_-2F0YN5vqwfY9UCHIlU,5791
 unstructured_ingest/v2/processes/connectors/azure_cognitive_search.py,sha256=S55v7TXu30rEdgythMBB_2VcuomyMPmcPtLYykbhw_E,8466
@@ -317,10 +362,10 @@ unstructured_ingest/v2/processes/connectors/couchbase.py,sha256=SONLywyEfoAlLc-H
 unstructured_ingest/v2/processes/connectors/databricks_volumes.py,sha256=BQHHpCDwE51inD3pZF4tL4zLr7lv6iBcwnA1NazrHqY,9423
 unstructured_ingest/v2/processes/connectors/elasticsearch.py,sha256=ojxMUHkLa6ZG50aTGn2YWhDHZ1n38uFRn5p8_ghAIvM,16762
 unstructured_ingest/v2/processes/connectors/google_drive.py,sha256=7xOQthcqBd9auJxB0nxZlhh1vdjXpMX_CtQZa6YfZz0,13088
-unstructured_ingest/v2/processes/connectors/kdbai.py,sha256=D71gt8fsPOXi2-Rir8mATw6dRM3BdzYGnn62qG1iaBw,5586
+unstructured_ingest/v2/processes/connectors/kdbai.py,sha256=8bGHbZctJ_Tl1AUSMnI7CCZ7CgEtTRVcRuvlB1HPlqQ,5907
 unstructured_ingest/v2/processes/connectors/local.py,sha256=a3stgnIkhBbXPIQD0O-RaRM-Eb-szHj9Yy4Fz881-9c,6723
 unstructured_ingest/v2/processes/connectors/milvus.py,sha256=ZUlyAQyTt0U1JoapFYHQW3IIaGYY50b3URDSLEAFjtk,7687
-unstructured_ingest/v2/processes/connectors/mongodb.py,sha256=2_R_hrEAaTU4vJTCK9oKblWTgv6BKjyUhFtC7uq3q2w,4859
+unstructured_ingest/v2/processes/connectors/mongodb.py,sha256=A0pt6JcNTD5bEu79jZ8KhnHcBQ2VUJ2AjtQAtdFr_Lo,13175
 unstructured_ingest/v2/processes/connectors/onedrive.py,sha256=ZiUo-dFo1LMOvFwphSLRZiR1PcrN8GWLTHhsh4TU6n0,9207
 unstructured_ingest/v2/processes/connectors/opensearch.py,sha256=dfDSNrWIEk19wuHdlMJpp_SLMOteNPlkDBPlAwu1LVY,6767
 unstructured_ingest/v2/processes/connectors/outlook.py,sha256=NK67Pd8Nk5oUIXTK-sK18K7rZ_Cl0UuCbeF2ExBEZho,9294
@@ -328,21 +373,30 @@ unstructured_ingest/v2/processes/connectors/pinecone.py,sha256=k_GH55S_OQ6-wCLC6
 unstructured_ingest/v2/processes/connectors/salesforce.py,sha256=2CiO2ZZiZ1Y1-nB7wcDlDVcpW2B7ut9wCj66rkkqho0,11616
 unstructured_ingest/v2/processes/connectors/sharepoint.py,sha256=hOaV5gBcHFc6N5Rbu3MgM-5Aol1ht-QkNIN4PqjvfxE,19665
 unstructured_ingest/v2/processes/connectors/singlestore.py,sha256=4rVvWKK2iQr03Ff6cB5zjfE1MpN0JyIGpCxxFCDI6hc,5563
-unstructured_ingest/v2/processes/connectors/sql.py,sha256=srj2ECKnkGR_iEFBdpa8sxw3ACCvJ5L0uoKCuHxKUe4,9204
 unstructured_ingest/v2/processes/connectors/utils.py,sha256=8kd0g7lo9NqnpaIkjeO-Ut6erhwUNH_gS9koevpe3WE,878
 unstructured_ingest/v2/processes/connectors/weaviate.py,sha256=Ss0YyD5T6k-00eJ6dr5lSo2H0LcOjVTMmozehyTvnAo,8866
+unstructured_ingest/v2/processes/connectors/databricks/__init__.py,sha256=jO71UTC7bLA_N12CrLWJzh_yZML5gfT7VohxzCpUGWg,1848
+unstructured_ingest/v2/processes/connectors/databricks/volumes.py,sha256=db4PxE1LiKWUq0b9THABFRChArAfHps89pZBglqEg3c,6521
+unstructured_ingest/v2/processes/connectors/databricks/volumes_aws.py,sha256=I1MJwe5LOxoPLjwo00H0XbXO6u_SJHWYgsj4s6ePoyI,2754
+unstructured_ingest/v2/processes/connectors/databricks/volumes_azure.py,sha256=P4rfcE3td7WyuuguRgUnGQytCMDpfeYrrpshBZuVynY,3539
+unstructured_ingest/v2/processes/connectors/databricks/volumes_gcp.py,sha256=UUotY_-HpgSEJkvdQfZTlbxY7CRLZ4ctL8TlryeFvxk,2790
+unstructured_ingest/v2/processes/connectors/databricks/volumes_native.py,sha256=Wk7s2_u5G0BOV5slvGc8IlUf7ivznY9PrgPqe6nlJKM,2897
 unstructured_ingest/v2/processes/connectors/fsspec/__init__.py,sha256=TtdeImM7Ypl_n6sl7I1JqX6bGSG0t_FqvCqE3Cy24og,1846
 unstructured_ingest/v2/processes/connectors/fsspec/azure.py,sha256=Y01BuVRql0Kvzc_cdaZE9dDGYjJzrwJu-etfUrEGcUU,7061
 unstructured_ingest/v2/processes/connectors/fsspec/box.py,sha256=Cjk0LUxqOCDbme0GmnD_5_b1hfStjI23cKw6BquKNrg,5488
 unstructured_ingest/v2/processes/connectors/fsspec/dropbox.py,sha256=NNAxIRdOQxUncfwhu7J7SnQRM6BSStNOyQZi-4E51iY,5816
-unstructured_ingest/v2/processes/connectors/fsspec/fsspec.py,sha256=usLzU2NA5D_a1juhja4jyJP_CzW4h-5rZ22bWVwvZGQ,10853
+unstructured_ingest/v2/processes/connectors/fsspec/fsspec.py,sha256=eFcrpSAB8wbLHuCiDb-2QpEUtgEEUA_iSqcT81H2-3Q,11472
 unstructured_ingest/v2/processes/connectors/fsspec/gcs.py,sha256=-_pYHbsBG9FyRyNIaf_xyFbPiiR7pnWEEg_8mp0rIZ8,7053
 unstructured_ingest/v2/processes/connectors/fsspec/s3.py,sha256=je1BDqFWlyMfPa4oAMMNFQLLQtCY9quuqx3xjTwF8OQ,6251
 unstructured_ingest/v2/processes/connectors/fsspec/sftp.py,sha256=dwpyqDq0qceCBWX3zM1hiUlgXB4hzX6ObOr-sh-5CJs,6926
 unstructured_ingest/v2/processes/connectors/fsspec/utils.py,sha256=jec_Qfe2hbfahBuY-u8FnvHuv933AI5HwPFjOL3kEEY,456
-unstructured_ingest-0.0.25.dist-info/LICENSE.md,sha256=SxkKP_62uIAKb9mb1eH7FH4Kn2aYT09fgjKpJt5PyTk,11360
-unstructured_ingest-0.0.25.dist-info/METADATA,sha256=NdNIJw4d0nu0NKP_FD5c8RZ2Tt3hWMMm0pJNdKGZdQU,7108
-unstructured_ingest-0.0.25.dist-info/WHEEL,sha256=GV9aMThwP_4oNCtvEC2ec3qUYutgWeAzklro_0m4WJQ,91
-unstructured_ingest-0.0.25.dist-info/entry_points.txt,sha256=gUAAFnjFPnBgThJSEbw0N5ZjxtaKlT1s9e05_arQrNw,70
-unstructured_ingest-0.0.25.dist-info/top_level.txt,sha256=QaTxTcjfM5Hr9sZJ6weOJvSe5ESQc0F8AWkhHInTCf8,20
-unstructured_ingest-0.0.25.dist-info/RECORD,,
+unstructured_ingest/v2/processes/connectors/sql/__init__.py,sha256=tr3SZH0tz04XSxqGRkUu__tL_0zn0bSms2jILE-3Rug,543
+unstructured_ingest/v2/processes/connectors/sql/postgres.py,sha256=oMwfYCycX-jTSKW-c6o6K09aU74Wn1B_G3Ib20oYi1A,6050
+unstructured_ingest/v2/processes/connectors/sql/sql.py,sha256=MbSvYSjhgGj8HHI7P-gH5bQ0Lqxtf8BEFsKNmCUfzug,9807
+unstructured_ingest/v2/processes/connectors/sql/sqlite.py,sha256=LxC2Q_rPHytbTDflmWzj4H5Jx-41phKnfp6FCpDe-UY,5701
+unstructured_ingest-0.1.1.dist-info/LICENSE.md,sha256=SxkKP_62uIAKb9mb1eH7FH4Kn2aYT09fgjKpJt5PyTk,11360
+unstructured_ingest-0.1.1.dist-info/METADATA,sha256=LQ_M1kX7q7rGBvslwml9KbrJGJHAaA_SLWM64BBaZrg,7188
+unstructured_ingest-0.1.1.dist-info/WHEEL,sha256=GV9aMThwP_4oNCtvEC2ec3qUYutgWeAzklro_0m4WJQ,91
+unstructured_ingest-0.1.1.dist-info/entry_points.txt,sha256=gUAAFnjFPnBgThJSEbw0N5ZjxtaKlT1s9e05_arQrNw,70
+unstructured_ingest-0.1.1.dist-info/top_level.txt,sha256=DMuDMHZRMdeay8v8Kdi855muIv92F0OkutvBCaBEW6M,25
+unstructured_ingest-0.1.1.dist-info/RECORD,,

{unstructured_ingest-0.0.25.dist-info → unstructured_ingest-0.1.1.dist-info}/top_level.txt

@@ -1 +1,2 @@
+test
 unstructured_ingest

unstructured_ingest/v2/processes/connectors/sql.py

@@ -1,275 +0,0 @@
-import json
-import uuid
-from dataclasses import dataclass, field
-from datetime import date, datetime
-from pathlib import Path
-from typing import TYPE_CHECKING, Any, Callable, Literal, Optional, Union
-
-import numpy as np
-import pandas as pd
-from dateutil import parser
-from pydantic import Field, Secret
-
-from unstructured_ingest.error import DestinationConnectionError
-from unstructured_ingest.utils.dep_check import requires_dependencies
-from unstructured_ingest.v2.interfaces import (
-    AccessConfig,
-    ConnectionConfig,
-    FileData,
-    Uploader,
-    UploaderConfig,
-    UploadStager,
-    UploadStagerConfig,
-)
-from unstructured_ingest.v2.logger import logger
-from unstructured_ingest.v2.processes.connector_registry import DestinationRegistryEntry
-
-if TYPE_CHECKING:
-    from sqlite3 import Connection as SqliteConnection
-
-    from psycopg2.extensions import connection as PostgresConnection
-
-CONNECTOR_TYPE = "sql"
-ELEMENTS_TABLE_NAME = "elements"
-SQLITE_DB = "sqlite"
-POSTGRESQL_DB = "postgresql"
-
-
-class SQLAccessConfig(AccessConfig):
-    username: Optional[str] = Field(default=None, description="DB username")
-    password: Optional[str] = Field(default=None, description="DB password")
-
-
-class SQLConnectionConfig(ConnectionConfig):
-    db_type: Literal["sqlite", "postgresql"] = Field(
-        default=SQLITE_DB, description="Type of the database backend"
-    )
-    database: Optional[str] = Field(
-        default=None,
-        description="Database name. For sqlite databases, this is the path to the .db file.",
-    )
-    host: Optional[str] = Field(default=None, description="DB host")
-    port: Optional[int] = Field(default=5432, description="DB host connection port")
-    access_config: Secret[SQLAccessConfig] = Field(default=SQLAccessConfig(), validate_default=True)
-    connector_type: str = Field(default=CONNECTOR_TYPE, init=False)
-
-    def __post_init__(self):
-        if (self.db_type == SQLITE_DB) and (self.database is None):
-            raise ValueError(
-                "A sqlite connection requires a path to a *.db file "
-                "through the `database` argument"
-            )
-
-
-class SQLUploadStagerConfig(UploadStagerConfig):
-    pass
-
-
-_COLUMNS = (
-    "id",
-    "element_id",
-    "text",
-    "embeddings",
-    "type",
-    "system",
-    "layout_width",
-    "layout_height",
-    "points",
-    "url",
-    "version",
-    "date_created",
-    "date_modified",
-    "date_processed",
-    "permissions_data",
-    "record_locator",
-    "category_depth",
-    "parent_id",
-    "attached_filename",
-    "filetype",
-    "last_modified",
-    "file_directory",
-    "filename",
-    "languages",
-    "page_number",
-    "links",
-    "page_name",
-    "link_urls",
-    "link_texts",
-    "sent_from",
-    "sent_to",
-    "subject",
-    "section",
-    "header_footer_type",
-    "emphasized_text_contents",
-    "emphasized_text_tags",
-    "text_as_html",
-    "regex_metadata",
-    "detection_class_prob",
-)
-
-_DATE_COLUMNS = ("date_created", "date_modified", "date_processed", "last_modified")
-
-
-def parse_date_string(date_value: Union[str, int]) -> date:
-    try:
-        timestamp = float(date_value) / 1000 if isinstance(date_value, int) else float(date_value)
-        return datetime.fromtimestamp(timestamp)
-    except Exception as e:
-        logger.debug(f"date {date_value} string not a timestamp: {e}")
-    return parser.parse(date_value)
-
-
-@dataclass
-class SQLUploadStager(UploadStager):
-    upload_stager_config: SQLUploadStagerConfig = field(
-        default_factory=lambda: SQLUploadStagerConfig()
-    )
-
-    def run(
-        self,
-        elements_filepath: Path,
-        file_data: FileData,
-        output_dir: Path,
-        output_filename: str,
-        **kwargs: Any,
-    ) -> Path:
-        with open(elements_filepath) as elements_file:
-            elements_contents: list[dict] = json.load(elements_file)
-        output_path = Path(output_dir) / Path(f"{output_filename}.json")
-        output_path.parent.mkdir(parents=True, exist_ok=True)
-
-        output = []
-        for data in elements_contents:
-            metadata: dict[str, Any] = data.pop("metadata", {})
-            data_source = metadata.pop("data_source", {})
-            coordinates = metadata.pop("coordinates", {})
-
-            data.update(metadata)
-            data.update(data_source)
-            data.update(coordinates)
-
-            data["id"] = str(uuid.uuid4())
-
-            # remove extraneous, not supported columns
-            data = {k: v for k, v in data.items() if k in _COLUMNS}
-
-            output.append(data)
-
-        df = pd.DataFrame.from_dict(output)
-        for column in filter(lambda x: x in df.columns, _DATE_COLUMNS):
-            df[column] = df[column].apply(parse_date_string)
-        for column in filter(
-            lambda x: x in df.columns,
-            ("permissions_data", "record_locator", "points", "links"),
-        ):
-            df[column] = df[column].apply(
-                lambda x: json.dumps(x) if isinstance(x, (list, dict)) else None
-            )
-        for column in filter(
-            lambda x: x in df.columns,
-            ("version", "page_number", "regex_metadata"),
-        ):
-            df[column] = df[column].apply(str)
-
-        with output_path.open("w") as output_file:
-            df.to_json(output_file, orient="records", lines=True)
-        return output_path
-
-
-class SQLUploaderConfig(UploaderConfig):
-    batch_size: int = Field(default=50, description="Number of records per batch")
-
-
-@dataclass
-class SQLUploader(Uploader):
-    connector_type: str = CONNECTOR_TYPE
-    upload_config: SQLUploaderConfig
-    connection_config: SQLConnectionConfig
-
-    def precheck(self) -> None:
-        try:
-            cursor = self.connection().cursor()
-            cursor.execute("SELECT 1;")
-            cursor.close()
-        except Exception as e:
-            logger.error(f"failed to validate connection: {e}", exc_info=True)
-            raise DestinationConnectionError(f"failed to validate connection: {e}")
-
-    @property
-    def connection(self) -> Callable[[], Union["SqliteConnection", "PostgresConnection"]]:
-        if self.connection_config.db_type == POSTGRESQL_DB:
-            return self._make_psycopg_connection
-        elif self.connection_config.db_type == SQLITE_DB:
-            return self._make_sqlite_connection
-        raise ValueError(f"Unsupported database {self.connection_config.db_type} connection.")
-
-    def _make_sqlite_connection(self) -> "SqliteConnection":
-        from sqlite3 import connect
-
-        return connect(database=self.connection_config.database)
-
-    @requires_dependencies(["psycopg2"], extras="postgres")
-    def _make_psycopg_connection(self) -> "PostgresConnection":
-        from psycopg2 import connect
-
-        access_config = self.connection_config.access_config.get_secret_value()
-        return connect(
-            user=access_config.username,
-            password=access_config.password,
-            dbname=self.connection_config.database,
-            host=self.connection_config.host,
-            port=self.connection_config.port,
-        )
-
-    def prepare_data(
-        self, columns: list[str], data: tuple[tuple[Any, ...], ...]
-    ) -> list[tuple[Any, ...]]:
-        output = []
-        for row in data:
-            parsed = []
-            for column_name, value in zip(columns, row):
-                if self.connection_config.db_type == SQLITE_DB and isinstance(value, (list, dict)):
-                    value = json.dumps(value)
-                if column_name in _DATE_COLUMNS:
-                    if value is None:
-                        parsed.append(None)
-                    else:
-                        parsed.append(parse_date_string(value))
-                else:
-                    parsed.append(value)
-            output.append(tuple(parsed))
-        return output
-
-    def upload_contents(self, path: Path) -> None:
-        df = pd.read_json(path, orient="records", lines=True)
-        logger.debug(f"uploading {len(df)} entries to {self.connection_config.database} ")
-        df.replace({np.nan: None}, inplace=True)
-
-        columns = tuple(df.columns)
-        stmt = f"INSERT INTO {ELEMENTS_TABLE_NAME} ({','.join(columns)}) \
-                VALUES({','.join(['?' if self.connection_config.db_type==SQLITE_DB else '%s' for x in columns])})"  # noqa E501
-
-        for rows in pd.read_json(
-            path, orient="records", lines=True, chunksize=self.upload_config.batch_size
-        ):
-            with self.connection() as conn:
-                values = self.prepare_data(columns, tuple(rows.itertuples(index=False, name=None)))
-                if self.connection_config.db_type == SQLITE_DB:
-                    conn.executemany(stmt, values)
-                else:
-                    with conn.cursor() as cur:
-                        cur.executemany(stmt, values)
-
-                conn.commit()
-
-    def run(self, path: Path, file_data: FileData, **kwargs: Any) -> None:
-        self.upload_contents(path=path)
-
-
-sql_destination_entry = DestinationRegistryEntry(
-    connection_config=SQLConnectionConfig,
-    uploader=SQLUploader,
-    uploader_config=SQLUploaderConfig,
-    upload_stager=SQLUploadStager,
-    upload_stager_config=SQLUploadStagerConfig,
-)