Changelog¶

in progress¶

Connectors: Added Azure Blob Storage (az://) and Azure Data Lake Storage Gen2 (adls://, abfss://) source and destination adapters, backed by adlfs. Supports account-key, SAS-token, and service-principal auth. Thanks, @hampsterx.
Filesystem destinations: Fixed blob-storage destinations (S3, GCS, Azure) aborting at the end of a load because the required post_load hook was missing from the shared base class. Thanks, @hampsterx.
Filesystem: Added #key=value reader hint channel to source URIs

Connectors: Added file:// source and destination adapters that read and write from/to local CSV / JSONL / Parquet files. Thanks, @hampsterx.
Connectors: Refactored filesystem internals
Connectors: Migrated filesystem CSV reader from pandas to Polars <<<<<<< HEAD
Filesystem readers: Added support for BSON. Thanks, @hampsterx. =======
Connectors: Added a file:// destination that writes local CSV / JSONL / Parquet files, the write-side twin of the file:// source. Thanks, @hampsterx.
Connectors: Added filesystem reader for the XLSX file format

f38a61cf (connector(xlsx): Add format reader for Excel XLSX files)

Connectors: Updated to Asana client v5
Core: Started using standard Python logger, removed --quiet option
API: Made option --source-table optional to prepare for streaming sources
Connectors: Added an mq-bridge source for streaming brokers (Kafka/NATS/AMQP/MQTT/ZeroMQ/IBM MQ/AWS SQS), via the <transport>+mqb:// URI scheme, with dotted query keys (e.g. tls.required=true) for nested config such as TLS. Thanks, @marcomq.
Maintenance: Started using orjson across the board
Connectors: Centralize file-format to reader mapping across all sources. Thanks, @hampsterx.

Maintenance: Refactored module namespace. If you are using omniload as a library, this introduces many breaking changes. However, the new layout is much more ergonomic.
Core: Added lazy-loading adapter module registry, to speed up startup times.