Version 0.9.8 of the relationalai Python package is now available.
This release includes improved error messages for configuration issues and enhanced accuracy in data stream status reporting in the CLI.
To upgrade, activate your virtual environment and run the following command:
Improved error messages for missing or invalid configuration files to make setup and troubleshooting easier:
If the raiconfig.toml file is missing, you’ll now see a clear message prompting you to run rai init to create one.
If the active profile specified in your config is missing or invalid, the error message guides you to check your profiles and use rai profile:switch or rai init as needed.
When required configuration values are missing, the error message now includes the profile name and config file path to help you quickly locate and fix the issue.
Fixed incorrect status reporting in the rai imports:list command.
In some cases, streams were shown as “completed” while they were still processing or quarantined.
The RelationalAI Python client now automatically retries requests that fail due to ConnectionResetError (such as “connection reset by peer”) when fetching query results.
This enhancement makes query result retrieval more resilient to transient network or server issues, reducing the likelihood of failures and improving overall reliability.
Fixed an issue where suspended engines incorrectly showed a timestamp in the SUSPENDS_AT field in the api.engines view and api.get_engine() procedure.
Now, for engines that are already suspended, the SUSPENDS_AT field will be NULL.