v1.0.7
Features
Multi Threaded Intent Detection
The DialogflowConversation
Class now supports Multi Threaded Intent Detection via Python threading!
This will enable hundreds or thousands of tests to be run in parallel, significantly expediting Intent Detection tasks for QA and Operation teams. See DialogflowConversation.run_intent_detection()
for more information.
Intents Language Code Support
All methods in the Intents
Class have been updated to include language_code
as an optional arg (where applicable) to allow for Intent CRUD operations across multi-lingual CX Agents.
Bug Fixes
- Fixed an issue in
StatsUtil.stats()
that caused theagent_id
to not be passed into the underlying functions appropriately - Fixed broken link in
examples/template.ipynb
Enhancements
- Additional error handling in
ChangeHistory.change_history_to_dataframe()
to account for the scenario where no Change History results are present in the provided CX Agent - Added
chunk_size
andrate_limit
args toDialogflowConversation.run_intent_detection()
to allow for more granular control over the amount of data sent to the Detect Intent API in parallel - Added Boolean flag for
checkpoints
inDialogflowConversation.reply()
to provide an option for users to obtain timestamp information for debugging purposes
Docs
- Updated docstrings for
DataframeFunctions.bulk_update_intent_from_dataframe()
which was missinglanguage_code
as an optional arg
Misc.
- Updated Docker image to be specific vs. latest due to issues in
3.10.x
- Updated Makefile