Mathis/llamactl
1
0
Fork 0
mirror of https://github.com/lordmathis/llamactl.git synced 2025-11-06 00:54:23 +00:00
Code Issues Packages Projects Releases Wiki Activity

Added support for serving behind a reverse proxy

Browse Source 
- Added support for specifying response headers for each backend
  - Allowing users to set `X-Accel-Buffering: no` to disable buffering for streaming responses in nginx
  - Updated `configuration.md` to document the new configuration options
- Modified Vite config to build with `base: "./"`, making assets be accessed via relative paths
- Updated API_BASE to use `document.baseURI`, allowing API calls to be made relative to the base path
This commit is contained in:
Anuruth Lertpiya
2025-09-29 12:37:00 +00:00
parent 92cb57e816
commit 5906d89f8d
6 changed files with 68 additions and 7 deletions
Download Patch File Download Diff File Expand all files Collapse all files

43
pkg/config/config.go
View File

@@ -13,10 +13,11 @@ import (
// BackendSettings contains structured backend configuration
type BackendSettings struct {
Command string `yaml:"command"`
Args []string `yaml:"args"`
Environment map[string]string `yaml:"environment,omitempty"`
Docker *DockerSettings `yaml:"docker,omitempty"`
Command string `yaml:"command"`
Args []string `yaml:"args"`
Environment map[string]string `yaml:"environment,omitempty"`
Docker *DockerSettings `yaml:"docker,omitempty"`
ResponseHeaders map[string]string `yaml:"response_headers,omitempty"`
}
// DockerSettings contains Docker-specific configuration
@@ -58,6 +59,9 @@ type ServerConfig struct {
// Enable Swagger UI for API documentation
EnableSwagger bool `yaml:"enable_swagger"`
// Response headers to send with responses
ResponseHeaders map[string]string `yaml:"response_headers,omitempty"`
}
// InstancesConfig contains instance management configuration
@@ -337,6 +341,12 @@ func loadEnvVars(cfg *AppConfig) {
}
parseEnvVars(llamaDockerEnv, cfg.Backends.LlamaCpp.Docker.Environment)
}
if llamaEnv := os.Getenv("LLAMACTL_LLAMACPP_RESPONSE_HEADERS"); llamaEnv != "" {
if cfg.Backends.LlamaCpp.ResponseHeaders == nil {
cfg.Backends.LlamaCpp.ResponseHeaders = make(map[string]string)
}
parseHeaders(llamaEnv, cfg.Backends.LlamaCpp.ResponseHeaders)
}
// vLLM backend
if vllmCmd := os.Getenv("LLAMACTL_VLLM_COMMAND"); vllmCmd != "" {
@@ -380,6 +390,12 @@ func loadEnvVars(cfg *AppConfig) {
}
parseEnvVars(vllmDockerEnv, cfg.Backends.VLLM.Docker.Environment)
}
if llamaEnv := os.Getenv("LLAMACTL_VLLM_RESPONSE_HEADERS"); llamaEnv != "" {
if cfg.Backends.VLLM.ResponseHeaders == nil {
cfg.Backends.VLLM.ResponseHeaders = make(map[string]string)
}
parseHeaders(llamaEnv, cfg.Backends.VLLM.ResponseHeaders)
}
// MLX backend
if mlxCmd := os.Getenv("LLAMACTL_MLX_COMMAND"); mlxCmd != "" {
@@ -394,6 +410,12 @@ func loadEnvVars(cfg *AppConfig) {
}
parseEnvVars(mlxEnv, cfg.Backends.MLX.Environment)
}
if llamaEnv := os.Getenv("LLAMACTL_MLX_RESPONSE_HEADERS"); llamaEnv != "" {
if cfg.Backends.MLX.ResponseHeaders == nil {
cfg.Backends.MLX.ResponseHeaders = make(map[string]string)
}
parseHeaders(llamaEnv, cfg.Backends.MLX.ResponseHeaders)
}
// Instance defaults
if autoRestart := os.Getenv("LLAMACTL_DEFAULT_AUTO_RESTART"); autoRestart != "" {
@@ -481,6 +503,19 @@ func parseEnvVars(envString string, envMap map[string]string) {
}
}
// parseHeaders parses HTTP headers in format "KEY1=value1;KEY2=value2"
// and populates the provided environment map
func parseHeaders(envString string, envMap map[string]string) {
if envString == "" {
return
}
for _, envPair := range strings.Split(envString, ";") {
if parts := strings.SplitN(strings.TrimSpace(envPair), "=", 2); len(parts) == 2 {
envMap[parts[0]] = parts[1]
}
}
}
// getDefaultDataDirectory returns platform-specific default data directory
func getDefaultDataDirectory() string {
switch runtime.GOOS {