HITL - Session Data Improvements

[0mdc]

Motivation and Context

This changeset changes the data format of the HITL output to improve its functionality.
In this context, a "session" is a sequence of N episodes to be completed.

Instead of recording a single file for a session, the system now records a single file per episode, and a session metadata file. The following content is uploaded to S3:

output_{port}/
    {episode_id}.json.gz
    {episode_id}.json.gz
    session.json.gz

Changes:

• Episodes are now recorded in individual files.
  • If the session fails (e.g. a user disconnects after completing some episodes), the completed episodes are uploaded.
• A lightweight session.json.jz file is created for each session, even if there is no episode.
  • The session contains vital data such as errors and total time spent.
• Data collection output is now structured.
• Empty frames are now recorded.
• Miscellaneous small improvements.

How Has This Been Tested

Tested on single-user and multi-user sessions.

Types of changes

• [Development]

Checklist

• My code follows the code style of this project.
• I have updated the documentation if required.
• I have read the CONTRIBUTING document.
• I have completed my CLA (see CONTRIBUTING)
• I have added tests to cover my changes if required.

[zephirefaith]

Makes sense. Thank you for improving the logic here.
Nit about same naming of local and class variable to avoid confusion.