unstructured-ingest 0.3.0__py3-none-any.whl → 0.3.2__py3-none-any.whl

{unstructured_ingest-0.3.0.dist-info → unstructured_ingest-0.3.2.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: unstructured-ingest
-Version: 0.3.0
+Version: 0.3.2
 Summary: A library that prepares raw documents for downstream ML tasks.
 Home-page: https://github.com/Unstructured-IO/unstructured-ingest
 Author: Unstructured Technologies
@@ -22,37 +22,37 @@ Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence
 Requires-Python: >=3.9.0,<3.13
 Description-Content-Type: text/markdown
 License-File: LICENSE.md
-Requires-Dist: tqdm
-Requires-Dist: pandas
-Requires-Dist: dataclasses-json
-Requires-Dist: opentelemetry-sdk
 Requires-Dist: python-dateutil
 Requires-Dist: pydantic>=2.7
+Requires-Dist: opentelemetry-sdk
 Requires-Dist: click
+Requires-Dist: tqdm
+Requires-Dist: pandas
+Requires-Dist: dataclasses-json
 Provides-Extra: airtable
 Requires-Dist: pyairtable; extra == "airtable"
 Provides-Extra: astradb
 Requires-Dist: astrapy; extra == "astradb"
 Provides-Extra: azure
-Requires-Dist: adlfs; extra == "azure"
 Requires-Dist: fsspec; extra == "azure"
+Requires-Dist: adlfs; extra == "azure"
 Provides-Extra: azure-ai-search
 Requires-Dist: azure-search-documents; extra == "azure-ai-search"
 Provides-Extra: bedrock
 Requires-Dist: boto3; extra == "bedrock"
 Provides-Extra: biomed
-Requires-Dist: requests; extra == "biomed"
 Requires-Dist: bs4; extra == "biomed"
+Requires-Dist: requests; extra == "biomed"
 Provides-Extra: box
-Requires-Dist: fsspec; extra == "box"
 Requires-Dist: boxfs; extra == "box"
+Requires-Dist: fsspec; extra == "box"
 Provides-Extra: chroma
 Requires-Dist: chromadb; extra == "chroma"
 Provides-Extra: clarifai
 Requires-Dist: clarifai; extra == "clarifai"
 Provides-Extra: confluence
-Requires-Dist: requests; extra == "confluence"
 Requires-Dist: atlassian-python-api; extra == "confluence"
+Requires-Dist: requests; extra == "confluence"
 Provides-Extra: couchbase
 Requires-Dist: couchbase; extra == "couchbase"
 Provides-Extra: csv
@@ -69,8 +69,8 @@ Requires-Dist: unstructured[docx]; extra == "doc"
 Provides-Extra: docx
 Requires-Dist: unstructured[docx]; extra == "docx"
 Provides-Extra: dropbox
-Requires-Dist: fsspec; extra == "dropbox"
 Requires-Dist: dropboxdrivefs; extra == "dropbox"
+Requires-Dist: fsspec; extra == "dropbox"
 Provides-Extra: elasticsearch
 Requires-Dist: elasticsearch[async]; extra == "elasticsearch"
 Provides-Extra: embed-huggingface
@@ -78,8 +78,8 @@ Requires-Dist: sentence-transformers; extra == "embed-huggingface"
 Provides-Extra: embed-mixedbreadai
 Requires-Dist: mixedbread-ai; extra == "embed-mixedbreadai"
 Provides-Extra: embed-octoai
-Requires-Dist: openai; extra == "embed-octoai"
 Requires-Dist: tiktoken; extra == "embed-octoai"
+Requires-Dist: openai; extra == "embed-octoai"
 Provides-Extra: embed-vertexai
 Requires-Dist: vertexai; extra == "embed-vertexai"
 Provides-Extra: embed-voyageai
@@ -87,19 +87,19 @@ Requires-Dist: voyageai; extra == "embed-voyageai"
 Provides-Extra: epub
 Requires-Dist: unstructured[epub]; extra == "epub"
 Provides-Extra: gcs
-Requires-Dist: fsspec; extra == "gcs"
 Requires-Dist: bs4; extra == "gcs"
 Requires-Dist: gcsfs; extra == "gcs"
+Requires-Dist: fsspec; extra == "gcs"
 Provides-Extra: github
-Requires-Dist: pygithub>1.58.0; extra == "github"
 Requires-Dist: requests; extra == "github"
+Requires-Dist: pygithub>1.58.0; extra == "github"
 Provides-Extra: gitlab
 Requires-Dist: python-gitlab; extra == "gitlab"
 Provides-Extra: google-drive
 Requires-Dist: google-api-python-client; extra == "google-drive"
 Provides-Extra: hubspot
-Requires-Dist: hubspot-api-client; extra == "hubspot"
 Requires-Dist: urllib3; extra == "hubspot"
+Requires-Dist: hubspot-api-client; extra == "hubspot"
 Provides-Extra: jira
 Requires-Dist: atlassian-python-api; extra == "jira"
 Provides-Extra: kafka
@@ -115,19 +115,19 @@ Requires-Dist: pymongo; extra == "mongodb"
 Provides-Extra: msg
 Requires-Dist: unstructured[msg]; extra == "msg"
 Provides-Extra: notion
-Requires-Dist: httpx; extra == "notion"
 Requires-Dist: backoff; extra == "notion"
 Requires-Dist: htmlBuilder; extra == "notion"
 Requires-Dist: notion-client; extra == "notion"
+Requires-Dist: httpx; extra == "notion"
 Provides-Extra: odt
 Requires-Dist: unstructured[odt]; extra == "odt"
 Provides-Extra: onedrive
+Requires-Dist: bs4; extra == "onedrive"
 Requires-Dist: msal; extra == "onedrive"
 Requires-Dist: Office365-REST-Python-Client; extra == "onedrive"
-Requires-Dist: bs4; extra == "onedrive"
 Provides-Extra: openai
-Requires-Dist: openai; extra == "openai"
 Requires-Dist: tiktoken; extra == "openai"
+Requires-Dist: openai; extra == "openai"
 Provides-Extra: opensearch
 Requires-Dist: opensearch-py; extra == "opensearch"
 Provides-Extra: org
@@ -156,8 +156,8 @@ Requires-Dist: unstructured[rst]; extra == "rst"
 Provides-Extra: rtf
 Requires-Dist: unstructured[rtf]; extra == "rtf"
 Provides-Extra: s3
-Requires-Dist: s3fs; extra == "s3"
 Requires-Dist: fsspec; extra == "s3"
+Requires-Dist: s3fs; extra == "s3"
 Provides-Extra: salesforce
 Requires-Dist: simple-salesforce; extra == "salesforce"
 Provides-Extra: sftp
@@ -171,8 +171,8 @@ Requires-Dist: singlestoredb; extra == "singlestore"
 Provides-Extra: slack
 Requires-Dist: slack-sdk[optional]; extra == "slack"
 Provides-Extra: snowflake
-Requires-Dist: snowflake-connector-python; extra == "snowflake"
 Requires-Dist: psycopg2-binary; extra == "snowflake"
+Requires-Dist: snowflake-connector-python; extra == "snowflake"
 Provides-Extra: togetherai
 Requires-Dist: together; extra == "togetherai"
 Provides-Extra: tsv

{unstructured_ingest-0.3.0.dist-info → unstructured_ingest-0.3.2.dist-info}/RECORD

@@ -5,27 +5,37 @@ test/integration/chunkers/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJW
 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=6dVNMBrL6WIO4KXA-0nf2tNrPYk_tsor8uomi6fbi3Q,727
-test/integration/connectors/test_astradb.py,sha256=Zy0xVOV60HOsvGKM4ankBE_X5ST37PBzR3iusk7DsEc,3492
-test/integration/connectors/test_azure_cog_search.py,sha256=dae4GifRiKue5YpsxworDiaMQoMsxcPDBithb6OFkx4,8876
+test/integration/connectors/test_astradb.py,sha256=QPFrODXmOHagpuKaiooxXb3OEW93w2g4fmq8BkaBCnY,5303
+test/integration/connectors/test_azure_ai_search.py,sha256=dae4GifRiKue5YpsxworDiaMQoMsxcPDBithb6OFkx4,8876
 test/integration/connectors/test_confluence.py,sha256=xcPmZ_vi_pkCt-tUPn10P49FH9i_9YUbrAPO6fYk5rU,3521
-test/integration/connectors/test_delta_table.py,sha256=4_KPyQJpd6DmyIjjtXWPMw6NNf7xULRkxmqfbvmZ80g,5018
-test/integration/connectors/test_kafka.py,sha256=3-OtZFZ93aCfmP0fUJzHJG7BBOfM5uCOtCLVHarsnMs,5869
+test/integration/connectors/test_delta_table.py,sha256=GSzWIkbEUzOrRPt2F1uO0dabcp7kTFDj75BhhI2y-WU,6856
+test/integration/connectors/test_kafka.py,sha256=j7jsNWZumNBv9v-5Bpx8geUUXpxxad5EuA4CMRsl4R8,7104
+test/integration/connectors/test_lancedb.py,sha256=O3YF6MVBkCsCgklXCJe8Kpy8aKGfafASVH4PspmpcYs,7628
+test/integration/connectors/test_milvus.py,sha256=CVmYw9iEeKT_0OtShxye2E6i1LbWzzDA8JtwJRkYQlA,4763
+test/integration/connectors/test_mongodb.py,sha256=YeS_DUnVYN02F76j87W8RhXGHnJMzQYb3n-L1-oWGXI,12254
 test/integration/connectors/test_onedrive.py,sha256=KIkBwKh1hnv203VCL2UABnDkS_bP4NxOFm1AL8EPGLA,3554
-test/integration/connectors/test_pinecone.py,sha256=809YADKRrdYnoXAd7HYaNCP3XJG7nb24NzOJkNu44nI,5535
+test/integration/connectors/test_pinecone.py,sha256=X10OWZ6IrO6YyhuR3ydMAZOQq3u2f5u_lCjKNYUUcnI,7558
 test/integration/connectors/test_qdrant.py,sha256=ASvO-BNyhv8m8or28KljrJy27Da0uaTNeoR5w_QsvFg,5121
 test/integration/connectors/test_s3.py,sha256=YHEYMqWTKTfR7wlL4VoxtgMs1YiYKyhLIBdG-anaQGo,6896
 test/integration/connectors/databricks_tests/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 test/integration/connectors/databricks_tests/test_volumes_native.py,sha256=k4lALbwNtlyuI3wd3OHoBULI21E3Ck2Fo8EJXaVfwgw,5812
+test/integration/connectors/elasticsearch/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+test/integration/connectors/elasticsearch/conftest.py,sha256=-i4_7MkIxSQENz7nuD2uHuhGU9mZ33vpeTPhHtRpQfs,989
+test/integration/connectors/elasticsearch/test_elasticsearch.py,sha256=nqdHwBpvgk_74orzDaQIKALK5cb0YloxSdt7QDJX0r0,11169
+test/integration/connectors/elasticsearch/test_opensearch.py,sha256=Rk4tQ_Qv5icycDWMUpnzTbg-QzwGyb6nKqB0gDef9D0,10555
 test/integration/connectors/sql/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-test/integration/connectors/sql/test_postgres.py,sha256=gDBuNyvWmpVPmDrSSYC99z3t17B_a196P1MwIAOp5Dk,6584
-test/integration/connectors/sql/test_singlestore.py,sha256=wGI3-lc6qh0qN4-WD9VtiXBB9MlekeqK402_9EXQyX0,5876
-test/integration/connectors/sql/test_snowflake.py,sha256=XXU2-2z_k8jHWP684v2IuaGOlV3cmPpg3RxkwMp08v8,6998
-test/integration/connectors/sql/test_sqlite.py,sha256=51QrFufAq-XxNjHAkmPWxdJUkGdIRRIGKeRT09A5pkA,5704
+test/integration/connectors/sql/test_postgres.py,sha256=lrymDI7bVX_4qij5gsUc_bTvHPeelu6hpJemQ6WWmlY,6783
+test/integration/connectors/sql/test_singlestore.py,sha256=iCp9q6tzhNIUCUubCPiRKj6VmJnwot4JGo9fkkTHg_U,5960
+test/integration/connectors/sql/test_snowflake.py,sha256=DqQIV9H5Uv7HaHtDyrAPdqefd316oVt5lKtdJ2Zdk6Q,7082
+test/integration/connectors/sql/test_sqlite.py,sha256=gSfp2hXAb5BGknzZXVa7K5bBwEb5Li4k5493mQCFjBQ,5719
 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.py,sha256=JxfX8u46YwpqUnVGd4syI0SrqGqvGQx9yBN0Xq-bIKE,2328
+test/integration/connectors/utils/docker.py,sha256=lnSjRgYoQa5c5nBdg2eLkB8KJVOjk4eyqq_C6PtTkME,4806
 test/integration/connectors/utils/docker_compose.py,sha256=GVTB6Cel05c0VQ2n4AwkQQx_cBfz13ZTs1HpbaYipNU,2223
-test/integration/connectors/utils/validation.py,sha256=5rQOBJyu1etvuwJmkH6xvKUPF08AKwJRxlN4L7-nw9w,13894
+test/integration/connectors/utils/validation.py,sha256=SwvPVuHjJxTo8xEUwnuL9FZNpu3sZZ8iouOz5xh_kB8,14272
+test/integration/connectors/weaviate/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+test/integration/connectors/weaviate/conftest.py,sha256=6Q6QdrLJmGHowRFSmoVSzup2EX6qASfS2Z5tqlpTm9M,387
+test/integration/connectors/weaviate/test_local.py,sha256=SK6iEwQUKiCd0X99BEk8GlQoLaCcJcFPt09NN526Ct0,4508
 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
@@ -71,7 +81,7 @@ test/unit/v2/partitioners/test_partitioner.py,sha256=iIYg7IpftV3LusoO4H8tr1IHY1U
 test/unit/v2/utils/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 test/unit/v2/utils/data_generator.py,sha256=UoYVNjG4S4wlaA9gceQ82HIpF9_6I1UTHD1_GrQBHp0,973
 unstructured_ingest/__init__.py,sha256=U4S_2y3zgLZVfMenHRaJFBW8yqh2mUBuI291LGQVOJ8,35
-unstructured_ingest/__version__.py,sha256=lWtlg90A2bUoi9oMXDJVdgZ8UO2vchSsWKV19YBO4f0,42
+unstructured_ingest/__version__.py,sha256=Js7MXQhyIj1akVjPNsLkmZxqoOHDGOr2opEPgFOSTZQ,42
 unstructured_ingest/error.py,sha256=qDncnJgbf5ils956RcO2CGlAKYDT5OaEM9Clv1JVTNc,1448
 unstructured_ingest/interfaces.py,sha256=OYVUP0bzBJpT-Lz92BDyz_hLBvyfxkuSwWHhUdnUayA,31493
 unstructured_ingest/logger.py,sha256=S5nSqGcABoQyeicgRnBQFjDScCaTvFVivOCvbo-laL0,4479
@@ -267,7 +277,7 @@ unstructured_ingest/pipeline/utils.py,sha256=RNx4bv2FhKOhaK_YTiRubta7n9wmJwqzznF
 unstructured_ingest/pipeline/write.py,sha256=xmDjmbieGRrcI342he7PkgxWaMoSJ5nWPmP5AM2xloU,669
 unstructured_ingest/pipeline/reformat/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 unstructured_ingest/pipeline/reformat/chunking.py,sha256=vbJgi2Yl9Rd9yZxIf64Nxj6cjUJnJWRpDCagswQmrLw,6040
-unstructured_ingest/pipeline/reformat/embedding.py,sha256=ohNvW9MhVGKVCx8ZlnLlLgkFQ_6UYLA7yUwT7Bzj9I8,2522
+unstructured_ingest/pipeline/reformat/embedding.py,sha256=vyRgrNvz50eMOCO00YdV9ODK0LRIB3_NF6t1mWD01uc,2525
 unstructured_ingest/runner/__init__.py,sha256=FO0X_jBIMilXdyjBajyFmzHoC3eVypNMGlhdOW4mcCM,2859
 unstructured_ingest/runner/airtable.py,sha256=1ndJ6PKT63E0gZN3KYFBj4Yo94zQYsIvSjC6ro2nIPE,1115
 unstructured_ingest/runner/astradb.py,sha256=FSBtQrsdC9E3eHUcAuQ0apcCnWolz-9tkvy-Uf7QeKg,1102
@@ -329,7 +339,7 @@ unstructured_ingest/runner/writers/fsspec/s3.py,sha256=kHJq2O3864QBd_tL2SKb0mdyw
 unstructured_ingest/utils/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 unstructured_ingest/utils/chunking.py,sha256=efWEfMcCukG5zASZrXhkNgAX8AzHa6t3rClMzm2TwFE,1521
 unstructured_ingest/utils/compression.py,sha256=NNiY-2S2Gf3at7zC1PYxMijaEza9vVSzRn5mdFf6mHo,4434
-unstructured_ingest/utils/data_prep.py,sha256=9UKewDHB8-cMlQ8POvokhjVsy-ksiSqAAW2ibqPYAfk,4400
+unstructured_ingest/utils/data_prep.py,sha256=IDAedOSBdgZpD9IY4tLJT-rmKGV7GHtU6KRj6VM-_tE,4666
 unstructured_ingest/utils/dep_check.py,sha256=SXXcUna2H0RtxA6j1S2NGkvQa9JP2DujWhmyBa7776Y,2400
 unstructured_ingest/utils/google_filetype.py,sha256=YVspEkiiBrRUSGVeVbsavvLvTmizdy2e6TsjigXTSRU,468
 unstructured_ingest/utils/string_and_date_utils.py,sha256=LwcbLmWpwt1zEabLlyUd5kIf9oOWcZxsRzxDglLCMeU,1375
@@ -382,36 +392,36 @@ unstructured_ingest/v2/processes/embedder.py,sha256=PQn0IO8xbGRQHpcT2VVl-J8gTJ5H
 unstructured_ingest/v2/processes/filter.py,sha256=kjUmMw2SDq2bme0JCAOxs6cJriIG6Ty09KOznS-xz08,2145
 unstructured_ingest/v2/processes/partitioner.py,sha256=agpHwB9FR8OZVQqE7zFEb0IcDPCOPA_BZjLzLF71nOY,8194
 unstructured_ingest/v2/processes/uncompress.py,sha256=Z_XfsITGdyaRwhtNUc7bMj5Y2jLuBge8KoK4nxhqKag,2425
-unstructured_ingest/v2/processes/connectors/__init__.py,sha256=ORSxrryPZErHAZTC3sp3UhWCh3G1B2SzTIM4H4OdVCc,5862
+unstructured_ingest/v2/processes/connectors/__init__.py,sha256=8M3aYYNbOkS2SYG2B_HLHMgX4V69-Oz1VqpQcRQMiVg,5167
 unstructured_ingest/v2/processes/connectors/airtable.py,sha256=eeZJe-bBNxt5Sa-XEFCdcGeJCguJU5WN2Mv9kLp5dVQ,8917
-unstructured_ingest/v2/processes/connectors/astradb.py,sha256=n5RT1l8pHbZG7m-CLKhWGCuWgfpeuIzvOZv7UAmTE6c,14683
-unstructured_ingest/v2/processes/connectors/azure_ai_search.py,sha256=5EXu565yDxjg7Iz0PO2mljwPnZVGYuWomNsbnMUOW_I,11813
+unstructured_ingest/v2/processes/connectors/astradb.py,sha256=zsIElFNNqVCXcLqBw6C8bRoyPQDrGNPkTWeA0FYYO94,14703
+unstructured_ingest/v2/processes/connectors/azure_ai_search.py,sha256=-6IijSWGqj-85vD0c4l5wdMHp-LF371jO8j53PPRB4I,12002
 unstructured_ingest/v2/processes/connectors/chroma.py,sha256=skrxRPHZ8y3JxNa0dt5SVitHiDQ5WVxLvY_kh2-QUrQ,8029
 unstructured_ingest/v2/processes/connectors/confluence.py,sha256=qQApDcmPBGg4tHXwSOj4JPkAbrO9GQ4NRlaETjhp25U,7003
 unstructured_ingest/v2/processes/connectors/couchbase.py,sha256=yhMDbpkZXs-Kis7tFlgjvNemU-MdWMdpCZDrpZNFaU4,12180
-unstructured_ingest/v2/processes/connectors/delta_table.py,sha256=ckdM6Z_hcltbtHdgkPi7_wntUvZSumAt7eQCxbmM4rQ,7480
-unstructured_ingest/v2/processes/connectors/elasticsearch.py,sha256=lNvUbbTMv2ZKxRN6cesfD2AeQc1kQG9AKqY9RHBfVXs,16796
+unstructured_ingest/v2/processes/connectors/delta_table.py,sha256=1yS7ivEyiucwd_kv6LL5HQdGabT43yeG6XCdwiz89hc,8019
 unstructured_ingest/v2/processes/connectors/gitlab.py,sha256=yBgCeLy9iCVI8bBDcHHuHB0H3BO05e9E1OccbHwvKAo,9724
-unstructured_ingest/v2/processes/connectors/google_drive.py,sha256=a1BAvhX3nsgghjuR5CJ1lOwMtJ5ZJwimg6VtDYvluxA,13104
+unstructured_ingest/v2/processes/connectors/google_drive.py,sha256=EEwXK1Anlu-eXl5qxmdDIqPYW7eMSez6WGlTPG2vSn8,13121
 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=A0pt6JcNTD5bEu79jZ8KhnHcBQ2VUJ2AjtQAtdFr_Lo,13175
+unstructured_ingest/v2/processes/connectors/milvus.py,sha256=Bzv2fa852BcM4_Pr-I_DPvLmjPoXv0Z7BeEA8qSKCDc,9725
+unstructured_ingest/v2/processes/connectors/mongodb.py,sha256=XLuprTCY0D9tAh_qn81MjJrDN9YaNqMlKe7BJl3eTZc,14998
 unstructured_ingest/v2/processes/connectors/onedrive.py,sha256=heZMtOIrCySi552ldIk8iH0pSRXZ0W2LeD-CcNOwCFQ,15979
-unstructured_ingest/v2/processes/connectors/opensearch.py,sha256=dfDSNrWIEk19wuHdlMJpp_SLMOteNPlkDBPlAwu1LVY,6767
 unstructured_ingest/v2/processes/connectors/outlook.py,sha256=KgNGM8hImRhy6_SpswRP2VwRD4VOrqqJoySgxf2oduI,9290
-unstructured_ingest/v2/processes/connectors/pinecone.py,sha256=Fh7G0gam49HSxn6SoWIIgqYTBKkY34u6LzjZmJB7fMI,10762
+unstructured_ingest/v2/processes/connectors/pinecone.py,sha256=hWkXgVDAzCtrBxf7A4HoexBACGAfVf_Qvn9YHbeiBSY,11505
 unstructured_ingest/v2/processes/connectors/salesforce.py,sha256=2CiO2ZZiZ1Y1-nB7wcDlDVcpW2B7ut9wCj66rkkqho0,11616
 unstructured_ingest/v2/processes/connectors/sharepoint.py,sha256=Ndn2Wm7RupfjAtlLxxQwJueeE0V8aGMbNVPuFq9nqdQ,19730
 unstructured_ingest/v2/processes/connectors/slack.py,sha256=Z73VmQ3oUY09KoLEi5OBdQeDt4ONEY_02SglWQc6HXE,9252
 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=p7sjCYZb7JmY3v3Xy1gm-q0O7oamLTsSFf2EWXYfXYQ,6447
 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/elasticsearch/__init__.py,sha256=Zzc0JNPP-eFqpwWw1Gp-XC8H-s__IgkYKzoagECycZY,829
+unstructured_ingest/v2/processes/connectors/elasticsearch/elasticsearch.py,sha256=lzbrQ66zz3Dh_G29XFkyzQ84St8H_xfQVsYV4mTf32c,19141
+unstructured_ingest/v2/processes/connectors/elasticsearch/opensearch.py,sha256=qRz8Fyr2RSZIPZGkhPeme6AZxM0aX-c_xOa1ZtSr2Kg,6781
 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
@@ -423,8 +433,14 @@ unstructured_ingest/v2/processes/connectors/fsspec/sftp.py,sha256=dwpyqDq0qceCBW
 unstructured_ingest/v2/processes/connectors/fsspec/utils.py,sha256=jec_Qfe2hbfahBuY-u8FnvHuv933AI5HwPFjOL3kEEY,456
 unstructured_ingest/v2/processes/connectors/kafka/__init__.py,sha256=mQJ9Ex-QCfhz-BB5YWTfbPf7xGLd1i7FpjRr0ukbhNw,754
 unstructured_ingest/v2/processes/connectors/kafka/cloud.py,sha256=qprsfI8VH0mVTa1MOCpa2D4coyopinQ5ag2KXcAecXE,3296
-unstructured_ingest/v2/processes/connectors/kafka/kafka.py,sha256=BkbozVTrDBingDuH8gTRiF5rceHoM1D3eibhl1pKgZQ,9092
+unstructured_ingest/v2/processes/connectors/kafka/kafka.py,sha256=qEv_yaG94KekFtfS06KgpTTbqeJkje0hn5uOjsMMngw,9414
 unstructured_ingest/v2/processes/connectors/kafka/local.py,sha256=vwLZjvc_C17zOqcrzic0aIoPwS98sqYiwiMknw2IcK4,2586
+unstructured_ingest/v2/processes/connectors/lancedb/__init__.py,sha256=lHUPCOiyOGu1IME1QiyFBZaB8z8e3bP8Y8TkqKs32Qk,906
+unstructured_ingest/v2/processes/connectors/lancedb/aws.py,sha256=yR8V4O-oI_nUKJtHTLxhteEJpPDPn-_d2IkkXvgThJ0,1406
+unstructured_ingest/v2/processes/connectors/lancedb/azure.py,sha256=Ms5vQVRIpTF1Q2qBl_bET9wbgaf4diPaH-iR8kJlr4E,1461
+unstructured_ingest/v2/processes/connectors/lancedb/gcp.py,sha256=p5BPaFtS3y3Yh8PIr3tUqsAXrUYu4QYYAWQNh5W2ucE,1361
+unstructured_ingest/v2/processes/connectors/lancedb/lancedb.py,sha256=7WIShs2V3dpN6wUhDTt1j2rvdiPp6yopbh7XYkb9T3s,5129
+unstructured_ingest/v2/processes/connectors/lancedb/local.py,sha256=_7-6iO6B60gAWwJUUrmlsRzYMFIBeZgu_QT3mhw5L0I,1272
 unstructured_ingest/v2/processes/connectors/qdrant/__init__.py,sha256=xM19uYzAuGizVoZIM_hnVZ5AcBN69aOBGpqZcpWPtuE,760
 unstructured_ingest/v2/processes/connectors/qdrant/cloud.py,sha256=accJ4sNWBVWV-KiVBDBDBYYx5A9CUoikP5NCErRmfik,1624
 unstructured_ingest/v2/processes/connectors/qdrant/local.py,sha256=cGEyv3Oy6y4BQ4DU8yhJWMpL82QYwBVdPTxxNuV127U,1588
@@ -434,11 +450,16 @@ unstructured_ingest/v2/processes/connectors/sql/__init__.py,sha256=D43wrV2ADvQsT
 unstructured_ingest/v2/processes/connectors/sql/postgres.py,sha256=rHaSb1MtdWMY6eQL2i2cWSL4w0VApFTChzmWtyfvFTI,5140
 unstructured_ingest/v2/processes/connectors/sql/singlestore.py,sha256=YrmhAL1RQ1c5-2fnR3UAyj_4KfvjYTQ2cWzpvsdJOnU,5535
 unstructured_ingest/v2/processes/connectors/sql/snowflake.py,sha256=jl524VudwmFK63emCT7DmZan_EWJAMiGir5_zoO9FuY,5697
-unstructured_ingest/v2/processes/connectors/sql/sql.py,sha256=Jwu3ZC4PGEw9la72cOwC3tclYAoBXFQTII9Mhh8ziP4,11571
+unstructured_ingest/v2/processes/connectors/sql/sql.py,sha256=LFzGeAUagLknK07DsXg2oSG7ZAgR6VqT9wfI_tYlHUg,14782
 unstructured_ingest/v2/processes/connectors/sql/sqlite.py,sha256=9605K36nQ5-gBxzt1daYKYotON1SE85RETusqCJrbdk,5230
-unstructured_ingest-0.3.0.dist-info/LICENSE.md,sha256=SxkKP_62uIAKb9mb1eH7FH4Kn2aYT09fgjKpJt5PyTk,11360
-unstructured_ingest-0.3.0.dist-info/METADATA,sha256=nn2t6UfzgYb6sr02uA_ixY-OQmcMwokknQ07Q9Kzdq0,7326
-unstructured_ingest-0.3.0.dist-info/WHEEL,sha256=GV9aMThwP_4oNCtvEC2ec3qUYutgWeAzklro_0m4WJQ,91
-unstructured_ingest-0.3.0.dist-info/entry_points.txt,sha256=gUAAFnjFPnBgThJSEbw0N5ZjxtaKlT1s9e05_arQrNw,70
-unstructured_ingest-0.3.0.dist-info/top_level.txt,sha256=DMuDMHZRMdeay8v8Kdi855muIv92F0OkutvBCaBEW6M,25
-unstructured_ingest-0.3.0.dist-info/RECORD,,
+unstructured_ingest/v2/processes/connectors/weaviate/__init__.py,sha256=eXamSnQdzzMvt62z80B8nmlkwDKO-Pogln_K_zLz53A,1067
+unstructured_ingest/v2/processes/connectors/weaviate/cloud.py,sha256=2g1Fm2J0ppfy2jCw4b5YtrsWrSD3VcrAaqiE7FlpIAg,6236
+unstructured_ingest/v2/processes/connectors/weaviate/embedded.py,sha256=S8Zg8StuZT-k7tCg1D5YShO1-vJYYk9-M1bE1fIqx64,3014
+unstructured_ingest/v2/processes/connectors/weaviate/local.py,sha256=LuTBKPseVewsz8VqxRPRLfGEm3BeI9nBZxpy7ZU5tOA,2201
+unstructured_ingest/v2/processes/connectors/weaviate/weaviate.py,sha256=ln1p9ahFTaT-qsL7p4bgw_IqnU60As_l6vVAqUWyQVE,11655
+unstructured_ingest-0.3.2.dist-info/LICENSE.md,sha256=SxkKP_62uIAKb9mb1eH7FH4Kn2aYT09fgjKpJt5PyTk,11360
+unstructured_ingest-0.3.2.dist-info/METADATA,sha256=rqTWqewB8eIrgrHJ-8AsNtehy35eSHKseCsveXTwN3Y,7326
+unstructured_ingest-0.3.2.dist-info/WHEEL,sha256=GV9aMThwP_4oNCtvEC2ec3qUYutgWeAzklro_0m4WJQ,91
+unstructured_ingest-0.3.2.dist-info/entry_points.txt,sha256=gUAAFnjFPnBgThJSEbw0N5ZjxtaKlT1s9e05_arQrNw,70
+unstructured_ingest-0.3.2.dist-info/top_level.txt,sha256=DMuDMHZRMdeay8v8Kdi855muIv92F0OkutvBCaBEW6M,25
+unstructured_ingest-0.3.2.dist-info/RECORD,,

unstructured_ingest/v2/processes/connectors/weaviate.py

@@ -1,242 +0,0 @@
-import json
-from dataclasses import dataclass, field
-from datetime import date, datetime
-from pathlib import Path
-from typing import TYPE_CHECKING, Any, Optional
-
-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 weaviate import Client
-
-CONNECTOR_TYPE = "weaviate"
-
-
-class WeaviateAccessConfig(AccessConfig):
-    access_token: Optional[str] = Field(
-        default=None, description="Used to create the bearer token."
-    )
-    api_key: Optional[str] = None
-    client_secret: Optional[str] = None
-    password: Optional[str] = None
-
-
-class WeaviateConnectionConfig(ConnectionConfig):
-    host_url: str = Field(description="Weaviate instance url")
-    class_name: str = Field(
-        description="Name of the class to push the records into, e.g: Pdf-elements"
-    )
-    access_config: Secret[WeaviateAccessConfig] = Field(
-        default=WeaviateAccessConfig(), validate_default=True
-    )
-    username: Optional[str] = None
-    anonymous: bool = Field(default=False, description="if set, all auth values will be ignored")
-    scope: Optional[list[str]] = None
-    refresh_token: Optional[str] = Field(
-        default=None,
-        description="Will tie this value to the bearer token. If not provided, "
-        "the authentication will expire once the lifetime of the access token is up.",
-    )
-    connector_type: str = Field(default=CONNECTOR_TYPE, init=False)
-
-
-class WeaviateUploadStagerConfig(UploadStagerConfig):
-    pass
-
-
-@dataclass
-class WeaviateUploadStager(UploadStager):
-    upload_stager_config: WeaviateUploadStagerConfig = field(
-        default_factory=lambda: WeaviateUploadStagerConfig()
-    )
-
-    @staticmethod
-    def parse_date_string(date_string: str) -> date:
-        try:
-            timestamp = float(date_string)
-            return datetime.fromtimestamp(timestamp)
-        except Exception as e:
-            logger.debug(f"date {date_string} string not a timestamp: {e}")
-        return parser.parse(date_string)
-
-    @classmethod
-    def conform_dict(cls, data: dict) -> None:
-        """
-        Updates the element dictionary to conform to the Weaviate schema
-        """
-
-        # Dict as string formatting
-        if record_locator := data.get("metadata", {}).get("data_source", {}).get("record_locator"):
-            # Explicit casting otherwise fails schema type checking
-            data["metadata"]["data_source"]["record_locator"] = str(json.dumps(record_locator))
-
-        # Array of items as string formatting
-        if points := data.get("metadata", {}).get("coordinates", {}).get("points"):
-            data["metadata"]["coordinates"]["points"] = str(json.dumps(points))
-
-        if links := data.get("metadata", {}).get("links", {}):
-            data["metadata"]["links"] = str(json.dumps(links))
-
-        if permissions_data := (
-            data.get("metadata", {}).get("data_source", {}).get("permissions_data")
-        ):
-            data["metadata"]["data_source"]["permissions_data"] = json.dumps(permissions_data)
-
-        # Datetime formatting
-        if date_created := data.get("metadata", {}).get("data_source", {}).get("date_created"):
-            data["metadata"]["data_source"]["date_created"] = cls.parse_date_string(
-                date_created
-            ).strftime(
-                "%Y-%m-%dT%H:%M:%S.%fZ",
-            )
-
-        if date_modified := data.get("metadata", {}).get("data_source", {}).get("date_modified"):
-            data["metadata"]["data_source"]["date_modified"] = cls.parse_date_string(
-                date_modified
-            ).strftime(
-                "%Y-%m-%dT%H:%M:%S.%fZ",
-            )
-
-        if date_processed := data.get("metadata", {}).get("data_source", {}).get("date_processed"):
-            data["metadata"]["data_source"]["date_processed"] = cls.parse_date_string(
-                date_processed
-            ).strftime(
-                "%Y-%m-%dT%H:%M:%S.%fZ",
-            )
-
-        if last_modified := data.get("metadata", {}).get("last_modified"):
-            data["metadata"]["last_modified"] = cls.parse_date_string(last_modified).strftime(
-                "%Y-%m-%dT%H:%M:%S.%fZ",
-            )
-
-        # String casting
-        if version := data.get("metadata", {}).get("data_source", {}).get("version"):
-            data["metadata"]["data_source"]["version"] = str(version)
-
-        if page_number := data.get("metadata", {}).get("page_number"):
-            data["metadata"]["page_number"] = str(page_number)
-
-        if regex_metadata := data.get("metadata", {}).get("regex_metadata"):
-            data["metadata"]["regex_metadata"] = str(json.dumps(regex_metadata))
-
-    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 = json.load(elements_file)
-        for element in elements_contents:
-            self.conform_dict(data=element)
-        output_path = Path(output_dir) / Path(f"{output_filename}.json")
-        with open(output_path, "w") as output_file:
-            json.dump(elements_contents, output_file)
-        return output_path
-
-
-class WeaviateUploaderConfig(UploaderConfig):
-    batch_size: int = Field(default=100, description="Number of records per batch")
-
-
-@dataclass
-class WeaviateUploader(Uploader):
-    upload_config: WeaviateUploaderConfig
-    connection_config: WeaviateConnectionConfig
-    connector_type: str = CONNECTOR_TYPE
-
-    @requires_dependencies(["weaviate"], extras="weaviate")
-    def get_client(self) -> "Client":
-        from weaviate import Client
-
-        auth = self._resolve_auth_method()
-        return Client(url=self.connection_config.host_url, auth_client_secret=auth)
-
-    def precheck(self) -> None:
-        try:
-            self.get_client()
-        except Exception as e:
-            logger.error(f"Failed to validate connection {e}", exc_info=True)
-            raise DestinationConnectionError(f"failed to validate connection: {e}")
-
-    @requires_dependencies(["weaviate"], extras="weaviate")
-    def _resolve_auth_method(self):
-        access_configs = self.connection_config.access_config.get_secret_value()
-        connection_config = self.connection_config
-        if connection_config.anonymous:
-            return None
-
-        if access_configs.access_token:
-            from weaviate.auth import AuthBearerToken
-
-            return AuthBearerToken(
-                access_token=access_configs.access_token,
-                refresh_token=connection_config.refresh_token,
-            )
-        elif access_configs.api_key:
-            from weaviate.auth import AuthApiKey
-
-            return AuthApiKey(api_key=access_configs.api_key)
-        elif access_configs.client_secret:
-            from weaviate.auth import AuthClientCredentials
-
-            return AuthClientCredentials(
-                client_secret=access_configs.client_secret, scope=connection_config.scope
-            )
-        elif connection_config.username and access_configs.password:
-            from weaviate.auth import AuthClientPassword
-
-            return AuthClientPassword(
-                username=connection_config.username,
-                password=access_configs.password,
-                scope=connection_config.scope,
-            )
-        return None
-
-    def run(self, path: Path, file_data: FileData, **kwargs: Any) -> None:
-        with path.open("r") as file:
-            elements_dict = json.load(file)
-        logger.info(
-            f"writing {len(elements_dict)} objects to destination "
-            f"class {self.connection_config.class_name} "
-            f"at {self.connection_config.host_url}",
-        )
-
-        client = self.get_client()
-        client.batch.configure(batch_size=self.upload_config.batch_size)
-        with client.batch as b:
-            for e in elements_dict:
-                vector = e.pop("embeddings", None)
-                b.add_data_object(
-                    e,
-                    self.connection_config.class_name,
-                    vector=vector,
-                )
-
-
-weaviate_destination_entry = DestinationRegistryEntry(
-    connection_config=WeaviateConnectionConfig,
-    uploader=WeaviateUploader,
-    uploader_config=WeaviateUploaderConfig,
-    upload_stager=WeaviateUploadStager,
-    upload_stager_config=WeaviateUploadStagerConfig,
-)