OutputParser Improvement

[Amaaan09]

Checked

• I searched existing ideas and did not find a similar one
• I added a very descriptive title
• I've clearly described the feature request and motivation for it

Feature request

Improvement for OutputParser

Specifically this one

Motivation

When working with LLMs other than that of OPENAI, I felt that the output parser was not working adequately even with prompt engineering. So I figured out how to improve this for my use case and would like to share the result with the community.

Proposal (If applicable)

In the said file I would improve it in such a way that elif statement of the includes_answer variable which returns the AgentFinish object should come before the action_match if statement which would also remove the need of the FINAL_ANSWER_AND_PARSABLE_ACTION_ERROR_MESSAGE.

Also perhaps I would like to code it myself and contribute.