1.2.0

Feature:

• Remove dependency on AIRFLOW__CORE__ENABLE_XCOM_PICKLING. Users can set new environment variables, namely AIRFLOW__ASTRO_SDK__XCOM_STORAGE_CONN_ID and AIRFLOW__ASTRO_SDK__XCOM_STORAGE_URL and use a custom XCOM backend namely, AstroCustomXcomBackend which enables the XCOM data to be saved to an S3 or GCS location. #795, #997
• Added OpenLineage support for LoadFileOperator , AppendOperator , TransformOperator and MergeOperator #898, #899, #902, #901 and #900
• Add TransformFileOperator that
  • parses a SQL file with templating
  • applies all needed parameters
  • runs the SQL to return a table object to keep the aql.transform_file function, the function can return TransformFileOperator().output in a similar fashion to the merge operator. #892
• Add the implementation for row count for BaseTable. #1073

Improvement:

• Improved handling of snowflake identifiers for smooth experience with dataframe and run_raw_sql and load_file operators. #917, #1098
• Fix transform_file to not depend on transform decorator #1004
• Set the CI to run and publish benchmark reports once a week #443
• Fix cyclic dependency and improve import time. Reduces the import time for astro/databases/__init__.py from 23.254 seconds to 0.062 seconds #1013

Docs

• Create GETTING_STARTED.md #1036
• Document the Open Lineage facets published by Astro Python SDK. #1086
• Documentation changes to specify permissions needed for running BigQuery jobs. #896
• Document the details on custom XCOM. #1100
• Document the benchmarking process. #1017
• Include a detailed description on the default Dataset concept in Astro Python SDK. #1092

Misc

• NFS volume mount in Kubernetes to test benchmarking from local to databases. #883