Refactor: Unified package structure, Development workflow enhancements, and fixes on issues.

[TheMightyRaider]

This PR is focussed on improving the ease of setting up the application for development along with fixes on some issues.

Single packaging

Currently, There are two packages that needs to be build for the functioning of the application,

• gh_processor
• gh_explainer

Testing out changes to the files inside gh_explainer and gh_processor folder becomes difficult because some of the scripts inside them depend on the build packages itself instead of the local function.

For example,
summarise.py uses a function called download_github_repo which is being used from the gh_processor package.

Let's say I want to update this function,

• Currently , I would have to change the function, which is declared in github_downloader.py.
• Create a new build of the gh_processor package.
• And now, Test the functionality.

This sounds cumbersome. Hence these are the following I have done in this PR,

• I have migrated the package imports to local imports, This would allow developers to just setup the environment once and work on the developments.
• Created a common .toml file for building a single package containing both gh_processor and gh_explainer.
• Created a .sh file for building the package. [TODO: The end goal should be to push it in the pip repository.]

Enhancements

• The repo which are inferenced are persistently stored locally in the application and this causes few issues. In this PR, I'm deleting the repos once the inference is done. This allows inferencing the same repo twice which wasn't happening earlier plus saving storage space.

Fixes

• Dynamically setting the max_length value based on the model chosen, which fixes this issue [ ValueError due to Input Length Exceeding max_length in Summarization #22 ].