acryl-datahub 1.0.0.1rc7__py3-none-any.whl → 1.0.0.2rc2__py3-none-any.whl

{acryl_datahub-1.0.0.1rc7.dist-info → acryl_datahub-1.0.0.2rc2.dist-info}/RECORD

@@ -1,7 +1,7 @@
-acryl_datahub-1.0.0.1rc7.dist-info/licenses/LICENSE,sha256=9xNHpsD0uYF5ONzXsKDCuHHB-xbiCrSbueWXqrTNsxk,11365
+acryl_datahub-1.0.0.2rc2.dist-info/licenses/LICENSE,sha256=9xNHpsD0uYF5ONzXsKDCuHHB-xbiCrSbueWXqrTNsxk,11365
 datahub/__init__.py,sha256=aq_i5lVREmoLfYIqcx_pEQicO855YlhD19tWc1eZZNI,59
 datahub/__main__.py,sha256=pegIvQ9hzK7IhqVeUi1MeADSZ2QlP-D3K0OQdEg55RU,106
-datahub/_version.py,sha256=AZj-rwp4edRcZvS9Mq4fxTeV64QHFW-6zysNAtjc2qg,323
+datahub/_version.py,sha256=lFv-ImaIXKL_EDY2GlHJHg9iVkj13C_xihZRNnxH3M8,323
 datahub/entrypoints.py,sha256=2TYgHhs3sCxJlojIHjqfxzt3_ImPwPzq4vBtsUuMqu4,8885
 datahub/errors.py,sha256=BzKdcmYseHOt36zfjJXc17WNutFhp9Y23cU_L6cIkxc,612
 datahub/py.typed,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
@@ -72,7 +72,7 @@ datahub/cli/env_utils.py,sha256=RQzjg4JE29hjPt4v7p-RuqoOr99w8E3DBHWiN2Sm7T4,252
 datahub/cli/exists_cli.py,sha256=IsuU86R-g7BJjAl1vULH6d-BWJHAKa4XHLZl5WxGUEM,1233
 datahub/cli/get_cli.py,sha256=VV80BCXfZ0-C8fr2k43SIuN9DB-fOYP9StWsTHnXwFw,2327
 datahub/cli/iceberg_cli.py,sha256=-XT3wpkr8b-HFMafYk7lSon3Lys6XjTQA8U1b698ByM,23003
-datahub/cli/ingest_cli.py,sha256=Xvdb0v-r_8rhlmZo161E043BMbLuyvNq1LLJWHdlXT0,20526
+datahub/cli/ingest_cli.py,sha256=Welutg0LOjuEiBOnQdNTnEZFDyKZiiWxqyQK-Go8dL8,20540
 datahub/cli/json_file.py,sha256=nWo-VVthaaW4Do1eUqgrzk0fShb29MjiKXvZVOTq76c,943
 datahub/cli/lite_cli.py,sha256=XKMejSuYUToKBvgN3YmmnxjRcaG5WPw23gJuQK8pgRc,13099
 datahub/cli/migrate.py,sha256=3orGfLNsdh1Q7gkPaCaf2bBWM5b3Ih4fGFw3poe0wiA,17937
@@ -122,7 +122,7 @@ datahub/emitter/generic_emitter.py,sha256=i37ZFm9VR_tmiZm9kIypEkQEB_cLKbzj_tJvVi
 datahub/emitter/kafka_emitter.py,sha256=Uix1W1WaXF8VqUTUfzdRZKca2XrR1w50Anx2LVkROlc,5822
 datahub/emitter/mce_builder.py,sha256=i-iLLdnuy7h1JrzwC2sCtQthbso-cNj1uijOQZKHbeA,16717
 datahub/emitter/mcp.py,sha256=hAAYziDdkwjazQU0DtWMbQWY8wS09ACrKJbqxoWXdgc,9637
-datahub/emitter/mcp_builder.py,sha256=JyAC8obvkf6ZpINJ8I2p-Ofr52-tuoQBDbxp-bhjyrM,11871
+datahub/emitter/mcp_builder.py,sha256=8IwJAlolQkPpMqQJPLtGrsUqAcuFNs98nrI5iYUxgaU,11920
 datahub/emitter/mcp_patch_builder.py,sha256=u7cpW6DkiN7KpLapmMaXgL_FneoN69boxiANbVgMdSI,4564
 datahub/emitter/request_helper.py,sha256=HpI9a9W0TzoVbrs584rF8P8w-IT_iKLmvYmO_6IHhXs,1008
 datahub/emitter/response_helper.py,sha256=h2hrZYiv4xfauD_lHPW_fN_AV8KhWNM4CVd-Lat2vT0,4608
@@ -171,7 +171,7 @@ datahub/ingestion/glossary/classifier.py,sha256=daLxnVv_JlfB_jBOxH5LrU_xQRndrsGo
 datahub/ingestion/glossary/classifier_registry.py,sha256=yFOYLQhDgCLqXYMG3L1BquXafeLcZDcmp8meyw6k9ts,307
 datahub/ingestion/glossary/datahub_classifier.py,sha256=O7wm6gQT1Jf2QSKdWjJQbS5oSzJwplXzfza26Gdq5Mg,7555
 datahub/ingestion/graph/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-datahub/ingestion/graph/client.py,sha256=Qtjf5YrQeQzcTb0qxr6-y4MSEKSJm8f0hO6BoeRA_yI,65916
+datahub/ingestion/graph/client.py,sha256=DUOy3fzwUMT0wJ2GAcLiYa5bzelwfZjBmzhbm95IIys,69918
 datahub/ingestion/graph/config.py,sha256=_oha8Je7P80ZmrkZUAaRHyYbdMmTkMI5JkYjEP2Ri1Q,751
 datahub/ingestion/graph/connections.py,sha256=9462L0ZWGKURyypAln25eMPhK3pcufBar9tNDoqspXs,741
 datahub/ingestion/graph/entity_versioning.py,sha256=nrcNz0Qm6kpE6oTu_mrYUQDx14KPspBTc6R9SyFUY6c,6901
@@ -205,7 +205,7 @@ datahub/ingestion/source/ge_profiling_config.py,sha256=FlWfXoVoayabVXNMB9qETEU0G
 datahub/ingestion/source/glue_profiling_config.py,sha256=vpMJH4Lf_qgR32BZy58suabri1yV5geaAPjzg2eORDc,2559
 datahub/ingestion/source/ldap.py,sha256=CNr3foofIpoCXu_GGqfcajlQE2qkHr5isYwVcDutdkk,18695
 datahub/ingestion/source/metabase.py,sha256=j8DRV2GvisezidL1JZ5HJLF_hdFdtvaoyDoEdEyh0Ks,32603
-datahub/ingestion/source/mlflow.py,sha256=pWhBEIXxI19-yTFeZU72Pgc3uY9nK0-w2TEpG803HNI,32485
+datahub/ingestion/source/mlflow.py,sha256=6uN1fjyubs9rjAsdtkSRMKf7h3_89UvFfWDqkgdvPdY,32422
 datahub/ingestion/source/mode.py,sha256=20vWflnRIMWqK8q2Mt2PorMBLnzRAA4bMjcWEyqhTio,64506
 datahub/ingestion/source/mongodb.py,sha256=2C2Cxn8DXL53IbNiywIuKt8UT_EMcPg9f8su-OPSNGU,21237
 datahub/ingestion/source/nifi.py,sha256=w5TPnqPmpotvzSsJROi6nUiHWPUVC6u1g0CzXIE6FNs,56903
@@ -217,7 +217,7 @@ datahub/ingestion/source/redash.py,sha256=YxjSad-X_wPmxYH8dJmFz_VCFhiLTCTSlK99Wd
 datahub/ingestion/source/salesforce.py,sha256=CQtDFv1OsbC1vyzNbKOc6GxhFQ5GdYj45hgAF0-oIcw,40487
 datahub/ingestion/source/source_registry.py,sha256=a2mLjJPLkSI-gYCTb_7U7Jo4D8jGknNQ_yScPIihXFk,1208
 datahub/ingestion/source/sql_queries.py,sha256=Ip7UZub7fgMh7P5jL_zJPY7lSkc9GGTy8GJ8lqZrcsE,9502
-datahub/ingestion/source/superset.py,sha256=FRZ7cCURW6NHUOKaFicdAZq2caXektvO9rJE4tO9scU,40336
+datahub/ingestion/source/superset.py,sha256=bMfvm9HgUoS3T7BjHsDrrOodc8iBRrJRQYv2D66bABo,41194
 datahub/ingestion/source/abs/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 datahub/ingestion/source/abs/config.py,sha256=mBQe0JTaP-Rcv4HnMUUySoYbSr4r3jDEMioxaXHnxXU,6709
 datahub/ingestion/source/abs/datalake_profiler_config.py,sha256=Rkf64evufyVGPiE4VK8QAjzBiJFu85tOGMmJ0lJZ2Og,3600
@@ -328,12 +328,13 @@ datahub/ingestion/source/grafana/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRk
 datahub/ingestion/source/grafana/grafana_source.py,sha256=3pU3xodPgS5lmnjuQ_u7F0XPzD_Y8MnPlMxRJ86qz4g,4960
 datahub/ingestion/source/hex/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 datahub/ingestion/source/hex/api.py,sha256=JfFPD8O4z16fwZE_BdX5aCQztEq-tbzxJJ7aofH4DE4,12274
-datahub/ingestion/source/hex/constants.py,sha256=NuBjxgJpIt598Cyn_9IcZ158PqBdn5vNjw8T92sTQck,115
-datahub/ingestion/source/hex/hex.py,sha256=DPpsi5e-sdUgbS0Okyvx1mvc00Adu47zA65oFnRP74A,6510
-datahub/ingestion/source/hex/mapper.py,sha256=6dsGvvhPAOAbAG1ayxLwipgJGt1q7YanWYfMX3rZeiM,12603
-datahub/ingestion/source/hex/model.py,sha256=hmMfOLEGZcKjwy2DW29OPf_9_Q_TesgnUTCen2br_fA,1471
+datahub/ingestion/source/hex/constants.py,sha256=8hUTMWyG5keTNfXoLu_Dh413Hw_mGGJX1atiiDZyKtg,271
+datahub/ingestion/source/hex/hex.py,sha256=PIRl8fPkKtlHV7cqR4H8RKVYdTLgEFXHFzc3QAqJLhE,12733
+datahub/ingestion/source/hex/mapper.py,sha256=N3mTlEcrOmhv9ia1dnHGFgFJD2ddyTtU3H5IUbb-UxU,13344
+datahub/ingestion/source/hex/model.py,sha256=S9bUhfFcjzuio2dBS6HzSyRVPiSJvRvMQ0qyVrjV5-E,1766
+datahub/ingestion/source/hex/query_fetcher.py,sha256=5r065vL7XohcgZ_fj-1h6o8cxrPin37IeYsC99GU6LA,12287
 datahub/ingestion/source/iceberg/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-datahub/ingestion/source/iceberg/iceberg.py,sha256=i9o0ia2vQUGqoagN7GgsoaUlhjj9xKBNP-3ia2cMgHY,30762
+datahub/ingestion/source/iceberg/iceberg.py,sha256=PhLLXWgBdfZ3hL7LgLvDr6aTK-QKmiZCFNz5jD-mxZM,30773
 datahub/ingestion/source/iceberg/iceberg_common.py,sha256=VGosqYPmn_j6GETSnDHZ8Ay1BVOedmx2x5LHxw16I3A,12278
 datahub/ingestion/source/iceberg/iceberg_profiler.py,sha256=9iwp2vpQTi4OMbIKoDZV5lAdvjMR0ls6Llpck9grJIE,9875
 datahub/ingestion/source/identity/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
@@ -378,7 +379,7 @@ datahub/ingestion/source/powerbi/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRk
 datahub/ingestion/source/powerbi/config.py,sha256=5rG62dspGF9jIo8l6HLpB6ECv5n-t1un2ZyGiisD784,24219
 datahub/ingestion/source/powerbi/dataplatform_instance_resolver.py,sha256=-njW1kJOy-LY5JFwJLhVQ0bMBj9NQz5TZhQqsSi_KsM,2285
 datahub/ingestion/source/powerbi/powerbi-lexical-grammar.rule,sha256=5df3qvalCS9hZ46DPXs6XDcw9-IofGf8Eol_rUC7LHI,20329
-datahub/ingestion/source/powerbi/powerbi.py,sha256=a5itVuGmg-0xAQK5a-cXB5UxpR3rLJx0o2x_lz-8ox8,55955
+datahub/ingestion/source/powerbi/powerbi.py,sha256=f0vwzVdKAU7Qp7gRHgTOPq45ThUmsbFXCwcIDaS1S34,56464
 datahub/ingestion/source/powerbi/m_query/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 datahub/ingestion/source/powerbi/m_query/data_classes.py,sha256=yDi0C13ko2dVxdLJBYvUuGbT4Q2hxQRse3sL7Ul1ZU0,2050
 datahub/ingestion/source/powerbi/m_query/native_sql_parser.py,sha256=zzKVDGeUM3Yv3-zNah4D6mSnr6jXsstNuLmzczcPQEE,3683
@@ -488,7 +489,7 @@ datahub/ingestion/source/sql/sql_utils.py,sha256=q-Bsk6WxlsRtrw9RXBxvqI3zuaMTC_F
 datahub/ingestion/source/sql/sqlalchemy_data_reader.py,sha256=FvHZ4JEK3aR2DYOBZiT_ZsAy12RjTu4t_KIR_92B11k,2644
 datahub/ingestion/source/sql/sqlalchemy_uri_mapper.py,sha256=KOpbmDIE2h1hyYEsbVHJi2B7FlsyUMTXZx4diyzltQg,1826
 datahub/ingestion/source/sql/teradata.py,sha256=9WdrxDy02lRJi9IZgsAATFsmxcQnIw5Gr6yCqHJQy5k,33507
-datahub/ingestion/source/sql/trino.py,sha256=8viVOu67mhDnsO3LuPSRi1WDR5MLdOXu7HOo1vtHVmo,17882
+datahub/ingestion/source/sql/trino.py,sha256=gSLDyETKavSVR8l9wdebrfoc41cqAWz6ApqIicW0BF8,17892
 datahub/ingestion/source/sql/two_tier_sql_source.py,sha256=YDrGBb5WKVls6qv17QU5foKrf71SydzEltc3WsVAhQc,5732
 datahub/ingestion/source/sql/vertica.py,sha256=_9OgSgIgqBml0av063rb8nACiT3SAmzpw0ouyF91wv8,33382
 datahub/ingestion/source/sql/mssql/__init__.py,sha256=1agpl8S_uDW40olkhCX_W19dbr5GO9qgjS3R7pLRZSk,87
@@ -504,7 +505,7 @@ datahub/ingestion/source/state/profiling_state.py,sha256=lsWu7oZhB9nSlqoklvjs-Lj
 datahub/ingestion/source/state/profiling_state_handler.py,sha256=jDMiIrAq8k4GrYoh9Ymh0ZAmzejYFk8E1W7-kuw6lXg,4295
 datahub/ingestion/source/state/redundant_run_skip_handler.py,sha256=h28twxcsMNvI74bUjAKleRYid8kfIyWS7Y11aBldDlY,9435
 datahub/ingestion/source/state/sql_common_state.py,sha256=OtJpJfMTBSgyR37dn3w-nnZwlc0nFNb2GoUzIWhnyAc,143
-datahub/ingestion/source/state/stale_entity_removal_handler.py,sha256=-KZjprFEO2tFtt2j236uRV1GVZEL5Q7Mt7TCZWfcxs8,14921
+datahub/ingestion/source/state/stale_entity_removal_handler.py,sha256=Lr2HYGx_b2FQ8A36s7s11tl-4-mGIM13bfy5JbQ3LtM,14890
 datahub/ingestion/source/state/stateful_ingestion_base.py,sha256=N0Qzp2t2qAf27WndhsvGbCYCd29dSrLY3TSfbO0hoKA,17369
 datahub/ingestion/source/state/usage_common_state.py,sha256=TJyb0CpwibsduJYI854EFdtrwWnz7JC-IkzKUXVGDx0,983
 datahub/ingestion/source/state/use_case_handler.py,sha256=3g8ddTvGXHe0dCiyTkyFeNmR8a3bhwywtIt8EpK5oQs,1271
@@ -535,7 +536,7 @@ datahub/ingestion/source/usage/clickhouse_usage.py,sha256=jJ-EUJdS7t4d9RVjLWQQ2e
 datahub/ingestion/source/usage/starburst_trino_usage.py,sha256=O3EDEZsXPNdsKGD-jStREA8e4-iTlnqd3ocqtAYFKNA,10544
 datahub/ingestion/source/usage/usage_common.py,sha256=uuCgIduhlRL2zIAN8rymZ5cZn1WF6akZ-ZbbaVYo9_w,9813
 datahub/ingestion/source/vertexai/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-datahub/ingestion/source/vertexai/vertexai.py,sha256=TXTa-Cm1C7xkbTNXNmKr_hi2FDJ9VG-ahrxk9yuxaTg,43635
+datahub/ingestion/source/vertexai/vertexai.py,sha256=ipq2Zb2lXTZkUg9r78kvuIIhX7mc-5hr-o83395IWpo,43589
 datahub/ingestion/source/vertexai/vertexai_config.py,sha256=uMnsv3b6TsPRH26u_JE_v1u0db7ANEAFlVxU5A6ELRM,989
 datahub/ingestion/source/vertexai/vertexai_result_type_utils.py,sha256=fE2l_xXvKONqb4jabl4LtKRBZDnP3koMLJV520wEAMg,2555
 datahub/ingestion/source_config/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
@@ -970,7 +971,7 @@ datahub/utilities/file_backed_collections.py,sha256=zW-Xy6zO1E6R-FRJKjlimMKr-emW
 datahub/utilities/global_warning_util.py,sha256=adrEl3WhetQ-bymrPINjd976ZFndhbvk3QosUYGsos8,261
 datahub/utilities/groupby.py,sha256=pe6rP4ZCttYB98yjbs0Aey8C32aLb7rq-NJ_BFky0H4,524
 datahub/utilities/hive_schema_to_avro.py,sha256=1MP0a6FFVEYxLg_4lKF7hPxbHJJy0uRQYkML5zRwV3Q,11622
-datahub/utilities/ingest_utils.py,sha256=znIuvFkCdOAOg1dkF-mJn03A2YYFPHlDPZsfCPxKkaQ,3117
+datahub/utilities/ingest_utils.py,sha256=OgETzX_9g6DcYlxGH0_xgAmAlWLSROr25ydDL-mBhKM,3137
 datahub/utilities/is_pytest.py,sha256=2m9T4S9IIKhI5RfTqrB2ZmumzHocdxBHpM1HroWj2XQ,138
 datahub/utilities/logging_manager.py,sha256=bc-x5VZGvFUHT0HD-TF3Uz_nzw3dpKdJSbz6kjpAqAQ,10073
 datahub/utilities/lossy_collections.py,sha256=5rdtfK2pjwvOrrzLf_KGFOMiVvLLmoXj5EVQXTFSR3E,5704
@@ -1043,8 +1044,8 @@ datahub_provider/operators/datahub_assertion_operator.py,sha256=uvTQ-jk2F0sbqqxp
 datahub_provider/operators/datahub_assertion_sensor.py,sha256=lCBj_3x1cf5GMNpHdfkpHuyHfVxsm6ff5x2Z5iizcAo,140
 datahub_provider/operators/datahub_operation_operator.py,sha256=aevDp2FzX7FxGlXrR0khoHNbxbhKR2qPEX5e8O2Jyzw,174
 datahub_provider/operators/datahub_operation_sensor.py,sha256=8fcdVBCEPgqy1etTXgLoiHoJrRt_nzFZQMdSzHqSG7M,168
-acryl_datahub-1.0.0.1rc7.dist-info/METADATA,sha256=QkeMAnAXXez9FFTnJVpxhOJMNDadF0gfVQ3uz3Fh4i8,176849
-acryl_datahub-1.0.0.1rc7.dist-info/WHEEL,sha256=CmyFI0kx5cdEMTLiONQRbGQwjIoR1aIYB7eCAQ4KPJ0,91
-acryl_datahub-1.0.0.1rc7.dist-info/entry_points.txt,sha256=o3mDeJXSKhsy7XLkuogihraiabBdLn9HaizYXPrxmk0,9710
-acryl_datahub-1.0.0.1rc7.dist-info/top_level.txt,sha256=iLjSrLK5ox1YVYcglRUkcvfZPvKlobBWx7CTUXx8_GI,25
-acryl_datahub-1.0.0.1rc7.dist-info/RECORD,,
+acryl_datahub-1.0.0.2rc2.dist-info/METADATA,sha256=VuKbVh0Lt8z7Jik8lZ39CF56PZHqn_oIwn2LBmYzrVc,176849
+acryl_datahub-1.0.0.2rc2.dist-info/WHEEL,sha256=CmyFI0kx5cdEMTLiONQRbGQwjIoR1aIYB7eCAQ4KPJ0,91
+acryl_datahub-1.0.0.2rc2.dist-info/entry_points.txt,sha256=o3mDeJXSKhsy7XLkuogihraiabBdLn9HaizYXPrxmk0,9710
+acryl_datahub-1.0.0.2rc2.dist-info/top_level.txt,sha256=iLjSrLK5ox1YVYcglRUkcvfZPvKlobBWx7CTUXx8_GI,25
+acryl_datahub-1.0.0.2rc2.dist-info/RECORD,,

datahub/_version.py

@@ -1,6 +1,6 @@
 # Published at https://pypi.org/project/acryl-datahub/.
 __package_name__ = "acryl-datahub"
-__version__ = "1.0.0.1rc7"
+__version__ = "1.0.0.2rc2"
 
 
 def is_dev_mode() -> bool:

datahub/cli/ingest_cli.py

@@ -216,9 +216,9 @@ def run(
 @click.option(
     "--executor-id",
     type=str,
-    default="default",
     help="Executor id to route execution requests to. Do not use this unless you have configured a custom executor.",
     required=False,
+    default=None,
 )
 @click.option(
     "--cli-version",
@@ -239,7 +239,7 @@ def run(
     type=str,
     help="Timezone for the schedule in 'America/New_York' format. Uses UTC by default.",
     required=False,
-    default="UTC",
+    default=None,
 )
 @click.option(
     "--debug", type=bool, help="Should we debug.", required=False, default=False
@@ -255,10 +255,10 @@ def deploy(
     name: Optional[str],
     config: str,
     urn: Optional[str],
-    executor_id: str,
+    executor_id: Optional[str],
     cli_version: Optional[str],
     schedule: Optional[str],
-    time_zone: str,
+    time_zone: Optional[str],
     extra_pip: Optional[str],
     debug: bool = False,
 ) -> None:

datahub/emitter/mcp_builder.py

@@ -137,6 +137,10 @@ class ProjectIdKey(ContainerKey):
     project_id: str
 
 
+class ExperimentKey(ContainerKey):
+    id: str
+
+
 class MetastoreKey(ContainerKey):
     metastore: str
 

datahub/ingestion/graph/client.py

@@ -27,6 +27,7 @@ from pydantic import BaseModel
 from requests.models import HTTPError
 from typing_extensions import deprecated
 
+from datahub._codegen.aspect import _Aspect
 from datahub.cli import config_utils
 from datahub.configuration.common import ConfigModel, GraphError, OperationalError
 from datahub.emitter.aspect import TIMESERIES_ASPECT_MAP
@@ -1697,6 +1698,7 @@ class DataHubGraph(DatahubRestEmitter, EntityVersioningAPI):
 
         return res["runAssertionsForAsset"]
 
+    @deprecated("Use get_entities instead which returns typed aspects")
     def get_entities_v2(
         self,
         entity_name: str,
@@ -1736,6 +1738,108 @@ class DataHubGraph(DatahubRestEmitter, EntityVersioningAPI):
                     retval[entity_urn][aspect_key] = aspect_value
         return retval
 
+    def get_entities(
+        self,
+        entity_name: str,
+        urns: List[str],
+        aspects: Optional[List[str]] = None,
+        with_system_metadata: bool = False,
+    ) -> Dict[str, Dict[str, Tuple[_Aspect, Optional[SystemMetadataClass]]]]:
+        """
+        Get entities using the OpenAPI v3 endpoint, deserializing aspects into typed objects.
+
+        Args:
+            entity_name: The entity type name
+            urns: List of entity URNs to fetch
+            aspects: Optional list of aspect names to fetch. If None, all aspects will be fetched.
+            with_system_metadata: If True, return system metadata along with each aspect.
+
+        Returns:
+            A dictionary mapping URNs to a dictionary of aspect name to tuples of
+            (typed aspect object, system metadata). If with_system_metadata is False,
+            the system metadata in the tuple will be None.
+        """
+        aspects = aspects or []
+
+        request_payload = []
+        for urn in urns:
+            entity_request: Dict[str, Any] = {"urn": urn}
+            for aspect_name in aspects:
+                entity_request[aspect_name] = {}
+            request_payload.append(entity_request)
+
+        headers: Dict[str, Any] = {
+            "Accept": "application/json",
+            "Content-Type": "application/json",
+        }
+
+        url = f"{self.config.server}/openapi/v3/entity/{entity_name}/batchGet"
+        if with_system_metadata:
+            url += "?systemMetadata=true"
+
+        response = self._session.post(
+            url, data=json.dumps(request_payload), headers=headers
+        )
+        response.raise_for_status()
+        entities = response.json()
+
+        result: Dict[str, Dict[str, Tuple[_Aspect, Optional[SystemMetadataClass]]]] = {}
+
+        for entity in entities:
+            entity_urn = entity.get("urn")
+            if entity_urn is None:
+                logger.warning(
+                    f"Missing URN in entity response: {entity}, skipping deserialization"
+                )
+                continue
+
+            entity_aspects: Dict[
+                str, Tuple[_Aspect, Optional[SystemMetadataClass]]
+            ] = {}
+
+            for aspect_name, aspect_obj in entity.items():
+                if aspect_name == "urn":
+                    continue
+
+                aspect_class = ASPECT_NAME_MAP.get(aspect_name)
+                if aspect_class is None:
+                    logger.warning(
+                        f"Unknown aspect type {aspect_name}, skipping deserialization"
+                    )
+                    continue
+
+                aspect_value = aspect_obj.get("value")
+                if aspect_value is None:
+                    logger.warning(
+                        f"Unknown aspect value for aspect {aspect_name}, skipping deserialization"
+                    )
+                    continue
+
+                try:
+                    post_json_obj = post_json_transform(aspect_value)
+                    typed_aspect = aspect_class.from_obj(post_json_obj)
+                    assert isinstance(typed_aspect, aspect_class) and isinstance(
+                        typed_aspect, _Aspect
+                    )
+
+                    system_metadata = None
+                    if with_system_metadata:
+                        system_metadata_obj = aspect_obj.get("systemMetadata")
+                        if system_metadata_obj:
+                            system_metadata = SystemMetadataClass.from_obj(
+                                system_metadata_obj
+                            )
+
+                    entity_aspects[aspect_name] = (typed_aspect, system_metadata)
+                except Exception as e:
+                    logger.error(f"Error deserializing aspect {aspect_name}: {e}")
+                    raise
+
+            if entity_aspects:
+                result[entity_urn] = entity_aspects
+
+        return result
+
     def upsert_custom_assertion(
         self,
         urn: Optional[str],

datahub/ingestion/source/hex/constants.py

@@ -1,3 +1,8 @@
+from datahub.metadata.urns import DataPlatformUrn
+
 HEX_PLATFORM_NAME = "hex"
+HEX_PLATFORM_URN = DataPlatformUrn(platform_name=HEX_PLATFORM_NAME)
 HEX_API_BASE_URL_DEFAULT = "https://app.hex.tech/api/v1"
 HEX_API_PAGE_SIZE_DEFAULT = 100
+
+DATAHUB_API_PAGE_SIZE_DEFAULT = 100

datahub/ingestion/source/hex/hex.py

@@ -1,9 +1,12 @@
+from dataclasses import dataclass
+from datetime import datetime, timedelta, timezone
 from typing import Any, Dict, Iterable, List, Optional
 
-from pydantic import Field, SecretStr
+from pydantic import Field, SecretStr, root_validator
 from typing_extensions import assert_never
 
 from datahub.configuration.common import AllowDenyPattern
+from datahub.configuration.datetimes import parse_user_datetime
 from datahub.configuration.source_common import (
     EnvConfigMixin,
     PlatformInstanceConfigMixin,
@@ -21,12 +24,17 @@ from datahub.ingestion.api.source import MetadataWorkUnitProcessor
 from datahub.ingestion.api.workunit import MetadataWorkUnit
 from datahub.ingestion.source.hex.api import HexApi, HexApiReport
 from datahub.ingestion.source.hex.constants import (
+    DATAHUB_API_PAGE_SIZE_DEFAULT,
     HEX_API_BASE_URL_DEFAULT,
     HEX_API_PAGE_SIZE_DEFAULT,
     HEX_PLATFORM_NAME,
 )
 from datahub.ingestion.source.hex.mapper import Mapper
 from datahub.ingestion.source.hex.model import Component, Project
+from datahub.ingestion.source.hex.query_fetcher import (
+    HexQueryFetcher,
+    HexQueryFetcherReport,
+)
 from datahub.ingestion.source.state.stale_entity_removal_handler import (
     StaleEntityRemovalHandler,
     StaleEntityRemovalSourceReport,
@@ -34,9 +42,10 @@ from datahub.ingestion.source.state.stale_entity_removal_handler import (
 )
 from datahub.ingestion.source.state.stateful_ingestion_base import (
     StatefulIngestionConfigBase,
-    StatefulIngestionReport,
     StatefulIngestionSourceBase,
 )
+from datahub.ingestion.source_report.ingestion_stage import IngestionStageReport
+from datahub.sdk.main_client import DataHubClient
 
 
 class HexSourceConfig(
@@ -93,9 +102,73 @@ class HexSourceConfig(
         default=True,
         description="Set ownership identity from owner/creator email",
     )
+    include_lineage: bool = Field(
+        default=True,
+        description='Include Hex lineage, being fetched from DataHub. See "Limitations" section in the docs for more details about the limitations of this feature.',
+    )
+    lineage_start_time: Optional[datetime] = Field(
+        default=None,
+        description="Earliest date of lineage to consider. Default: 1 day before lineage end time. You can specify absolute time like '2023-01-01' or relative time like '-7 days' or '-7d'.",
+    )
+    lineage_end_time: Optional[datetime] = Field(
+        default=None,
+        description="Latest date of lineage to consider. Default: Current time in UTC. You can specify absolute time like '2023-01-01' or relative time like '-1 day' or '-1d'.",
+    )
+    datahub_page_size: int = Field(
+        default=DATAHUB_API_PAGE_SIZE_DEFAULT,
+        description="Number of items to fetch per DataHub API call.",
+    )
+
+    @root_validator(pre=True)
+    def validate_lineage_times(cls, data: Dict[str, Any]) -> Dict[str, Any]:
+        # lineage_end_time default = now
+        if "lineage_end_time" not in data or data["lineage_end_time"] is None:
+            data["lineage_end_time"] = datetime.now(tz=timezone.utc)
+        # if string is given, parse it
+        if isinstance(data["lineage_end_time"], str):
+            data["lineage_end_time"] = parse_user_datetime(data["lineage_end_time"])
+        # if no timezone is given, assume UTC
+        if data["lineage_end_time"].tzinfo is None:
+            data["lineage_end_time"] = data["lineage_end_time"].replace(
+                tzinfo=timezone.utc
+            )
+        # at this point, we ensure there is a non null datetime with UTC timezone for lineage_end_time
+        assert (
+            data["lineage_end_time"]
+            and isinstance(data["lineage_end_time"], datetime)
+            and data["lineage_end_time"].tzinfo is not None
+            and data["lineage_end_time"].tzinfo == timezone.utc
+        )
+
+        # lineage_start_time default = lineage_end_time - 1 day
+        if "lineage_start_time" not in data or data["lineage_start_time"] is None:
+            data["lineage_start_time"] = data["lineage_end_time"] - timedelta(days=1)
+        # if string is given, parse it
+        if isinstance(data["lineage_start_time"], str):
+            data["lineage_start_time"] = parse_user_datetime(data["lineage_start_time"])
+        # if no timezone is given, assume UTC
+        if data["lineage_start_time"].tzinfo is None:
+            data["lineage_start_time"] = data["lineage_start_time"].replace(
+                tzinfo=timezone.utc
+            )
+        # at this point, we ensure there is a non null datetime with UTC timezone for lineage_start_time
+        assert (
+            data["lineage_start_time"]
+            and isinstance(data["lineage_start_time"], datetime)
+            and data["lineage_start_time"].tzinfo is not None
+            and data["lineage_start_time"].tzinfo == timezone.utc
+        )
+
+        return data
 
 
-class HexReport(StaleEntityRemovalSourceReport, HexApiReport):
+@dataclass
+class HexReport(
+    StaleEntityRemovalSourceReport,
+    HexApiReport,
+    IngestionStageReport,
+    HexQueryFetcherReport,
+):
     pass
 
 
@@ -110,7 +183,7 @@ class HexSource(StatefulIngestionSourceBase):
     def __init__(self, config: HexSourceConfig, ctx: PipelineContext):
         super().__init__(config, ctx)
         self.source_config = config
-        self.report = HexReport()
+        self.report: HexReport = HexReport()
         self.platform = HEX_PLATFORM_NAME
         self.hex_api = HexApi(
             report=self.report,
@@ -129,6 +202,28 @@ class HexSource(StatefulIngestionSourceBase):
             categories_as_tags=self.source_config.categories_as_tags,
             set_ownership_from_email=self.source_config.set_ownership_from_email,
         )
+        self.project_registry: Dict[str, Project] = {}
+        self.component_registry: Dict[str, Component] = {}
+
+        self.datahub_client: Optional[DataHubClient] = None
+        self.query_fetcher: Optional[HexQueryFetcher] = None
+        if self.source_config.include_lineage:
+            graph = ctx.require_graph("Lineage")
+            assert self.source_config.lineage_start_time and isinstance(
+                self.source_config.lineage_start_time, datetime
+            )
+            assert self.source_config.lineage_end_time and isinstance(
+                self.source_config.lineage_end_time, datetime
+            )
+            self.datahub_client = DataHubClient(graph=graph)
+            self.query_fetcher = HexQueryFetcher(
+                datahub_client=self.datahub_client,
+                workspace_name=self.source_config.workspace_name,
+                start_datetime=self.source_config.lineage_start_time,
+                end_datetime=self.source_config.lineage_end_time,
+                report=self.report,
+                page_size=self.source_config.datahub_page_size,
+            )
 
     @classmethod
     def create(cls, config_dict: Dict[str, Any], ctx: PipelineContext) -> "HexSource":
@@ -143,25 +238,58 @@ class HexSource(StatefulIngestionSourceBase):
             ).workunit_processor,
         ]
 
-    def get_report(self) -> StatefulIngestionReport:
+    def get_report(self) -> HexReport:
         return self.report
 
     def get_workunits_internal(self) -> Iterable[MetadataWorkUnit]:
-        yield from self.mapper.map_workspace()
-
-        for project_or_component in self.hex_api.fetch_projects():
-            if isinstance(project_or_component, Project):
-                if self.source_config.project_title_pattern.allowed(
-                    project_or_component.title
-                ):
-                    yield from self.mapper.map_project(project=project_or_component)
-            elif isinstance(project_or_component, Component):
-                if (
-                    self.source_config.include_components
-                    and self.source_config.component_title_pattern.allowed(
+        with self.report.new_stage("Fetch Hex assets from Hex API"):
+            for project_or_component in self.hex_api.fetch_projects():
+                if isinstance(project_or_component, Project):
+                    if self.source_config.project_title_pattern.allowed(
                         project_or_component.title
-                    )
-                ):
-                    yield from self.mapper.map_component(component=project_or_component)
-            else:
-                assert_never(project_or_component)
+                    ):
+                        self.project_registry[project_or_component.id] = (
+                            project_or_component
+                        )
+                elif isinstance(project_or_component, Component):
+                    if (
+                        self.source_config.include_components
+                        and self.source_config.component_title_pattern.allowed(
+                            project_or_component.title
+                        )
+                    ):
+                        self.component_registry[project_or_component.id] = (
+                            project_or_component
+                        )
+                else:
+                    assert_never(project_or_component)
+
+        if self.source_config.include_lineage:
+            assert self.datahub_client and self.query_fetcher
+
+            with self.report.new_stage(
+                "Fetch Hex lineage from existing Queries in DataHub"
+            ):
+                for query_metadata in self.query_fetcher.fetch():
+                    project = self.project_registry.get(query_metadata.hex_project_id)
+                    if project:
+                        project.upstream_datasets.extend(
+                            query_metadata.dataset_subjects
+                        )
+                        project.upstream_schema_fields.extend(
+                            query_metadata.schema_field_subjects
+                        )
+                    else:
+                        self.report.report_warning(
+                            title="Missing project for lineage",
+                            message="Lineage missed because missed project, likely due to filter patterns or deleted project.",
+                            context=str(query_metadata),
+                        )
+
+        with self.report.new_stage("Emit"):
+            yield from self.mapper.map_workspace()
+
+            for project in self.project_registry.values():
+                yield from self.mapper.map_project(project=project)
+            for component in self.component_registry.values():
+                yield from self.mapper.map_component(component=component)

datahub/ingestion/source/hex/mapper.py

@@ -1,6 +1,6 @@
 import logging
 from datetime import datetime
-from typing import Iterable, List, Optional, Tuple
+from typing import Iterable, List, Optional, Tuple, Union
 
 from datahub._codegen.aspect import (
     _Aspect,  # TODO: is there a better import than this one?
@@ -46,6 +46,7 @@ from datahub.metadata.schema_classes import (
     DashboardInfoClass,
     DashboardUsageStatisticsClass,
     DataPlatformInstanceClass,
+    EdgeClass,
     GlobalTagsClass,
     OwnerClass,
     OwnershipClass,
@@ -53,7 +54,14 @@ from datahub.metadata.schema_classes import (
     TagAssociationClass,
     TimeWindowSizeClass,
 )
-from datahub.metadata.urns import ContainerUrn, CorpUserUrn, DashboardUrn, Urn
+from datahub.metadata.urns import (
+    ContainerUrn,
+    CorpUserUrn,
+    DashboardUrn,
+    DatasetUrn,
+    SchemaFieldUrn,
+    Urn,
+)
 
 logger = logging.getLogger(__name__)
 
@@ -116,6 +124,8 @@ class Mapper:
             ),
             externalUrl=f"{self._base_url}/{self._workspace_name}/hex/{project.id}",
             customProperties=dict(id=project.id),
+            datasetEdges=self._dataset_edges(project.upstream_datasets),
+            # TODO: support schema field upstream, maybe InputFields?
         )
 
         subtypes = SubTypesClass(
@@ -343,6 +353,22 @@ class Mapper:
             else None,
         )
 
+    def _dataset_edges(
+        self, upstream: List[Union[DatasetUrn, SchemaFieldUrn]]
+    ) -> Optional[List[EdgeClass]]:
+        # TBC: is there support for CLL in Dashboards? for the moment, skip SchemaFieldUrns
+        return (
+            [
+                EdgeClass(
+                    destinationUrn=upstream_urn.urn(),
+                )
+                for upstream_urn in upstream
+                if isinstance(upstream_urn, DatasetUrn)
+            ]
+            if upstream
+            else None
+        )
+
     def _yield_mcps(
         self, entity_urn: Urn, aspects: List[Optional[_Aspect]]
     ) -> Iterable[MetadataWorkUnit]:

datahub/ingestion/source/hex/model.py

@@ -1,6 +1,8 @@
-from dataclasses import dataclass
+from dataclasses import dataclass, field
 from datetime import datetime
-from typing import List, Optional
+from typing import List, Optional, Union
+
+from datahub.metadata.urns import DatasetUrn, SchemaFieldUrn
 
 
 @dataclass
@@ -51,6 +53,12 @@ class Project:
     creator: Optional[Owner] = None
     owner: Optional[Owner] = None
     analytics: Optional[Analytics] = None
+    upstream_datasets: List[Union[DatasetUrn, SchemaFieldUrn]] = field(
+        default_factory=list
+    )
+    upstream_schema_fields: List[Union[DatasetUrn, SchemaFieldUrn]] = field(
+        default_factory=list
+    )
 
 
 @dataclass