--set-<value> should have a --value counter part for session based options

[nopeless]

--set-<value> writes to file. I wouldn't want that.

[kardolus]

@nopeless flags would be one way of doing that, however you can already use the environment variables. Every configuration option has an associated environment variable, eg:

OPENAI_MODEL=gpt-3.5-turbo chatgpt what is the capital of Belgium?

or

OPENAI_ROLE='you extract wisdom' cat README.md | chatgpt summarize

[nopeless]

@kardolus environment variables are hard to keep in closures especially in powershell they don't have something like env to keep them temporarily :/

[kardolus]

@nopeless ahh gotcha. I didn't know that about powershell. I can add the --value flags. May take me a week or two.

[nopeless]

btw by --value i was thinking about --max-tokens for example. i hope it isn't that confusing

e.g.

chatgpt-cli --role "You are developer" -i

[kardolus]

I am finally ready to pick this one up. I am reinventing a lot of the Viper functionality in my CLI. I want to get rid of my own logic as much as I can and fall back to Viper. Will work on this issue in parallel to this one: #68

I think both can be solved within the same refactor.

[kardolus]

FYI, started working on it here: https://github.com/kardolus/chatgpt-cli/tree/viper-refactor

It's a massive refactor. It's not done yet (I want setters for all values). Here's a preview though:

 2024-10-06 09:35:12 ⌚ (☸️ |kind-apache:N/A) Guillermos-MacBook-Pro in ~/workspace/chatgpt-cli
± |viper-refactor {3} ✓| → ./bin/chatgpt --help
A powerful ChatGPT client that enables seamless interactions with the GPT model. Provides multiple modes and context management features, including the ability to pipe custom context into the conversation.

Usage:
  chatgpt [flags]

Flags:
      --api-key string             Set the API key for authentication
      --auth-header string         Set the authorization header
      --auth-token-prefix string   Set the authorization token prefix
      --auto-create-new-thread     Automatically create a new thread for each session
      --clear-history              Clear all prior conversation context for the current thread
      --command-prompt string      Set the command prompt format
      --completions-path string    Set the completions API endpoint
  -c, --config                     Display the configuration
      --debug                      Enable debug mode
      --delete-thread string       Delete the specified thread
      --frequency-penalty float    Set the frequency penalty
  -h, --help                       help for chatgpt
  -i, --interactive                Use interactive mode
  -l, --list-models                List available models
      --list-threads               List available threads
      --models-path string         Set the models API endpoint
      --omit-history               Omit history in the conversation
      --output-prompt string       Set the output prompt format
      --presence-penalty float     Set the presence penalty
  -q, --query                      Use query mode instead of stream mode
      --set-completions string     Generate autocompletion script for your current shell
      --set-context-window int     Set a new default context window size
      --set-max-tokens int         Set a new default max token size by specifying the max tokens
      --set-model string           Set a new default GPT model by specifying the model name
      --set-thread string          Set a new active thread by specifying the thread name
      --skip-tls-verify            Skip TLS certificate verification
      --temperature float          Set the sampling temperature (default 1)
      --top-p float                Set the top-p value for nucleus sampling (default 1)
      --track-token-usage          Track token usage
      --url string                 Set the API base URL
  -v, --version                    Display the version information

[kardolus]

Yikes, it's pretty cluttered haha. Getting there though.

 2024-10-06 21:24:20 ⌚ (☸️ |kind-cloudera:N/A) Guillermos-MacBook-Pro in ~/workspace/chatgpt-cli
± |viper-refactor {3} U:1 ✗| → ./bin/chatgpt --help
ChatGPT CLI - A powerful client for interacting with GPT models.

Usage:
  chatgpt [flags]

General Flags:
  -q, --query                    Use query mode instead of stream mode
  -i, --interactive              Use interactive mode
  -c, --config                   Display the configuration
  -v, --version                  Display the version information
  --list-models                  List available models
  --list-threads                 List available threads
  --clear-history                Clear the history of the current thread
  --delete-thread                Delete the specified thread
  --set-completions              Generate autocompletion script for your current shell

Configuration Setters:
  --set-api-key                  Set the API key for authentication
  --set-auth-header              Set the authorization header
  --set-auth-token-prefix        Set the authorization token prefix
  --set-auto-create-new-thread   Automatically create a new thread for each session
  --set-command-prompt           Set the command prompt format
  --set-completions              Generate autocompletion script for your current shell
  --set-completions-path         Set the completions API endpoint
  --set-context-window           Set a new default context window size
  --set-debug                    Enable debug mode
  --set-frequency-penalty        Set the frequency penalty
  --set-max-tokens               Set a new default max token size
  --set-model                    Set a new default GPT model by specifying the model name
  --set-models-path              Set the models API endpoint
  --set-omit-history             Omit history in the conversation
  --set-output-prompt            Set the output prompt format
  --set-presence-penalty         Set the presence penalty
  --set-skip-tls-verify          Skip TLS certificate verification
  --set-temperature              Set the sampling temperature
  --set-thread                   Set a new active thread by specifying the thread name
  --set-top-p                    Set the top-p value for nucleus sampling
  --set-track-token-usage        Track token usage
  --set-url                      Set the API base URL

Configuration Aliases:
  --api_key                      Alias for setting api key
  --auth_header                  Alias for setting auth header
  --auth_token_prefix            Alias for setting auth token prefix
  --auto_create_new_thread       Alias for setting auto create new thread
  --command_prompt               Alias for setting command prompt
  --completions_path             Alias for setting completions path
  --context_window               Alias for setting context window
  --debug                        Alias for setting debug
  --frequency_penalty            Alias for setting frequency penalty
  --help                         Alias for setting help
  --max_tokens                   Alias for setting max tokens
  --model                        Alias for setting model
  --models_path                  Alias for setting models path
  --omit_history                 Alias for setting omit history
  --output_prompt                Alias for setting output prompt
  --presence_penalty             Alias for setting presence penalty
  --skip_tls_verify              Alias for setting skip tls verify
  --temperature                  Alias for setting temperature
  --thread                       Alias for setting thread
  --top_p                        Alias for setting top p
  --track_token_usage            Alias for setting track token usage
  --url                          Alias for setting url

Environment Variables:
  You can also use environment variables to set config values. For example:
  OPENAI_API_KEY=your_api_key chatgpt --query 'Hello'

[kardolus]

Fixed in this commit: 1ad9eeb

Release: https://github.com/kardolus/chatgpt-cli/releases/tag/v1.7.0

[nopeless]

@kardolus I'm reading your second most recent comment, and it appears that the casing for the commands seem inconsistent? I don't think that's a good idea. Id like to hear your reasoning

[kardolus]

Your right! That’s fixed :)

[kardolus]

Here's the new output:

 2024-10-08 18:38:03 ⌚ (☸️ |kind-apache:N/A) Guillermos-MacBook-Pro in ~/workspace/chatgpt-cli
± |main {6} ✓| → chatgpt --help
ChatGPT CLI - A powerful client for interacting with GPT models.

Usage:
  chatgpt [flags]

General Flags:
  -q, --query                    Use query mode instead of stream mode
  -i, --interactive              Use interactive mode
  -c, --config                   Display the configuration
  -v, --version                  Display the version information
  -l, --list-models              List available models
  --list-threads                 List available threads
  --clear-history                Clear the history of the current thread
  --delete-thread                Delete the specified thread
  --set-completions              Generate autocompletion script for your current shell

Persistent Configuration Setters:
  --set-api-key                  Set the API key for authentication
  --set-auth-header              Set the authorization header
  --set-auth-token-prefix        Set the authorization token prefix
  --set-auto-create-new-thread   Automatically create a new thread for each session
  --set-command-prompt           Set the command prompt format
  --set-completions-path         Set the completions API endpoint
  --set-context-window           Set a new default context window size
  --set-debug                    Enable debug mode
  --set-frequency-penalty        Set the frequency penalty
  --set-max-tokens               Set a new default max token size
  --set-model                    Set a new default GPT model by specifying the model name
  --set-models-path              Set the models API endpoint
  --set-name                     The prefix for environment variable overrides
  --set-omit-history             Omit history in the conversation
  --set-output-prompt            Set the output prompt format
  --set-presence-penalty         Set the presence penalty
  --set-role                     Set the role of the AI assistant
  --set-skip-tls-verify          Skip TLS certificate verification
  --set-temperature              Set the sampling temperature
  --set-thread                   Set a new active thread by specifying the thread name
  --set-top-p                    Set the top-p value for nucleus sampling
  --set-track-token-usage        Track token usage
  --set-url                      Set the API base URL

Runtime Value Overrides:
  --api-key                      Override value for api-key
  --auth-header                  Override value for auth-header
  --auth-token-prefix            Override value for auth-token-prefix
  --auto-create-new-thread       Override value for auto-create-new-thread
  --command-prompt               Override value for command-prompt
  --completions-path             Override value for completions-path
  --context-window               Override value for context-window
  --debug                        Override value for debug
  --frequency-penalty            Override value for frequency-penalty
  --max-tokens                   Override value for max-tokens
  --model                        Override value for model
  --models-path                  Override value for models-path
  --name                         Override value for name
  --omit-history                 Override value for omit-history
  --output-prompt                Override value for output-prompt
  --presence-penalty             Override value for presence-penalty
  --role                         Override value for role
  --skip-tls-verify              Override value for skip-tls-verify
  --temperature                  Override value for temperature
  --thread                       Override value for thread
  --top-p                        Override value for top-p
  --track-token-usage            Override value for track-token-usage
  --url                          Override value for url

Environment Variables:
  You can also use environment variables to set config values. For example:
  OPENAI_API_KEY=your_api_key chatgpt --query 'Hello'

Configuration File:
  All configuration changes made with the setters will be saved in the config.yaml file.
  The config.yaml file is located in the following path: /Users/guillermo/.chatgpt-cli/config.yaml
  You can edit this file manually to change configuration settings as well.

[kardolus]

@nopeless unless you mean something else. I figured the snake vs kebab case was what you meant. Am I overlooking upper/lower casing?

[nopeless]

@kardolus I was referring to the mixed kebab and snake case. Thanks for fixing it 👍