Documentation
¶
Overview ¶
Package cmds provides CLI commands for llmcli.
Index ¶
Constants ¶
This section is empty.
Variables ¶
This section is empty.
Functions ¶
func NewInferCmd ¶
NewInferCmd returns the infer command.
func NewModelsCmd ¶
NewModelsCmd returns the models command.
Types ¶
type RootFlags ¶ added in v0.23.0
type RootFlags struct {
// LogHTTP enables HTTP request/response logging at Debug level.
LogHTTP bool
// LogHTTPDebug additionally logs request/response headers and bodies.
// Has no effect unless LogHTTP is also true.
LogHTTPDebug bool
// LogHTTPAllHeaders disables the response header allowlist and prints all headers.
// Has no effect unless LogHTTPDebug is also true.
LogHTTPAllHeaders bool
// LogEvents prints each StreamEvent as JSON to stderr as it is consumed
// from the channel, before any other handling. Useful for debugging the
// provider event stream at the application layer.
LogEvents bool
}
RootFlags holds flags defined on the root command that are shared across all subcommands.
func (*RootFlags) BuildHTTPClient ¶ added in v0.23.0
BuildHTTPClient constructs an *http.Client from the root flags. Returns nil when no HTTP logging is requested, causing providers to fall back to llm.DefaultHttpClient(). The returned handler may be non-nil even when the client is nil — callers should check the client first.
func (*RootFlags) BuildLLMOptions ¶ added in v0.23.0
BuildLLMOptions returns llm.Option values for the root flags to be passed to providers that don't use the HTTP transport for logging (e.g. Bedrock).
Click to show internal directories.
Click to hide internal directories.