v0.6.2

• additional debug output and unit tests

v0.6.1

• added assistant_message function

v0.6.0

Migrate to new openai API; see openai/openai-python#742.
As noted in the README this is for backward compatibility only. Consider rewriting your code to use the new openai API directly rather than relying on openai-pygenerator.

v0.5.7

• Removed unused dependency: returns

v0.5.6

• Decreased default timeout to 20 seconds (from 60 seconds)
• Increased default max tokens to 500 (from 100 tokens)

Note that both of the above are configurable, either through environment variables or programmatically - see the README.md.

v0.5.5

This release will also retry when an APIConnectionError occurs.

v0.5.4

Added new environment variable GPT_REQUEST_TIMEOUT_SECONDS which overrides the request_timeout parameter when calling the OpenAI API.

For more details see https://community.openai.com/t/configuring-timeout-for-chatcompletion-python/107226/24?page=2.

v0.5.2

Added new environment variable GPT_REQUEST_TIMEOUT_SECONDS which overrides the request_timeout parameter when calling the OpenAI API.

For more details see https://community.openai.com/t/configuring-timeout-for-chatcompletion-python/107226/24?page=2.

v0.5.1

• updated readme with explanations and links to functional-programming concepts

v0.5.0

• retry on 500 status