Project Roadmap

[signebedi]

0.1.0 - add basic completion wrapper

• [logging] write tagged output to file #5
• [config] add support for a gptty.conf file #3
• [runtime] add context-sharing across chats with same tag #7

0.2.0 - add runtime commands

• [runtime] Add a CLI runtime with Click #2
• Install with pypi #11
• [runtime] add support for question chains / queues #17

0.2.1 - add log command

• [runtime] add a log subcommand #21
• [docs] add some guidelines for bash scripting #23

0.2.2 - add keyword tokenization

• [context] allow keyword-only context using RAKE to minimize API usage fees #25
• [context] weigh recent text more heavily in keyword tokenization #26

0.2.3 - model validation and chat completion

• [bug] openai.error.InvalidRequestError: This is a chat model and not supported in the v1/completions endpoint. Did you mean to use v1/chat/completions? #30
• [model] add support for chatcompletion models #31
• [model] add model validation #32
• [runtime] add response formatting options #24
• [model] allow forgetting older conversations within same context when approaching token limit #36

0.2.4 - debug and verbose modes

• [tests] add debug / verbose mode #28
• [tests] add tiktoken to show token count in debug / verbose mode #29
• [tests] write initial unit tests #40
• [tests] integrate with testing CI/CD software #41

0.2.5 - improved chatroom experience and query options

• [context] add support for --additional_context option for query subcommand #37
• [tests] test additional_context passed to get_context #54
• [tests] Add context.py tests when model_type == 'v1/chat/completions' #43
• [runtime] add a --json flag to return query data as a JSON object #46
• [runtime] add a --quiet tag for query subcommand #47
• [runtime] add a check for internet connectivity and return error if no connection #48
• Using arrow keys to edit question prints escape characters #50
• [runtime] gracefully handle keyboard interrupts in chat  #51
• [context] logic sometimes fails to constrain queries to max_context_length #53

0.2.6 - organization ID fix

• [bug] cannot connect to any model without org_id specified #63

0.2.7 - UniversalCompletion abstraction

• [scripting] expose universalCompletion class in __init__.py #39
• [docs] generate documentation #66

0.2.8 - daemons and add'l context tools

• [context] allow positional context passing #16
• [daemon] write a listener daemon to monitor a question queue #18

Backlog

• [runtime] add bool config log_reponses #35
• [tests] consider whether and how to test __main__.py and gptty.py #44
• [scripting] extend the UniversalCompletion class with context generation #68