Fix #143: Add data creation CLI

[rchan26]

For #143:

• Create reginald_create_index CLI to create an index
• Create Dockerfile for reginald_create_index
• Split up data index creation from LlamaIndex to make a DataIndexCreator class
• Also add a setup_service_context helper function which gets called in both setting up LlamaIndex model and when creating the index on its own
• Some more logging with the parser
  • Created get_env_var helper function which just logs when it's trying to read a environment variable and says if it's able to find it. If not able to, default arguments are used
  • I made the default arguments in argparses lambda functions so that we only try to get environment variables if they are needed (a bit of a hack because otherwise, it'd log that it's trying to find an environment variable even if we pass in a value for the CLI argument)
  • get_args loops through the arguments and calls the arguments and calls any lambda functions to get the default arguments
• k, chunk_size, chunk_overlap_ratio, num_output are now able to be passed as CLI arguments in reginald_run, reginald_run_api_llm and reginald_create_index

[rchan26]

converting to draft as a few things to fix regarding the types of the default arguments. trying to fix this now

[rwood-97]

Looks good - I changed the instances of os.getenv to get_env_var which means it prints the "getting SLACK_APP_TOKEN " bit twice (since it does actually get it twice) but I think that is okay?