mirror of
https://gitcode.com/JianFeeeee/ModelRouter.git
synced 2026-09-24 19:08:04 +00:00
feat: ModelRouter — unified OpenAI-compatible multi-source LLM gateway
- Lua adapters per upstream (transform_request/response/stream_chunk, build_headers signing hooks) - AUTO priority routing with per-model kind (chat/image), explicit source/model routing - Per-source concurrency caps with queueing, exponential backoff, AUTO failover - OpenAI-compatible API: chat completions, SSE streaming, image generations, models - Gateway key auth, web UI for adapter/source management, runtime persistence - e2e test running the real binary against mocked upstreams
This commit is contained in:
124
internal/config/config.go
Normal file
124
internal/config/config.go
Normal file
@ -0,0 +1,124 @@
|
||||
// Package config loads the gateway YAML configuration plus a runtime overlay
|
||||
// (web UI edits) and resolves them into sources with per-model priority.
|
||||
package config
|
||||
|
||||
import (
|
||||
"fmt"
|
||||
"os"
|
||||
"time"
|
||||
|
||||
"gopkg.in/yaml.v3"
|
||||
)
|
||||
|
||||
// Config is the top-level gateway configuration.
|
||||
type Config struct {
|
||||
Listen string `yaml:"listen"`
|
||||
GatewayKeys []string `yaml:"gateway_keys"`
|
||||
DefaultModel string `yaml:"default_model"` // e.g. "AUTO" or a model id
|
||||
AdapterDir string `yaml:"adapter_dir"`
|
||||
RuntimeFile string `yaml:"runtime_file"`
|
||||
MaxConcurrent int `yaml:"max_concurrent"` // global inflight cap, 0 = unlimited
|
||||
Sources []Source `yaml:"sources"`
|
||||
}
|
||||
|
||||
// Model is a single exposed model id bound to a source, with priority used by
|
||||
// AUTO model selection (higher number = preferred).
|
||||
type Model struct {
|
||||
ID string `yaml:"id"`
|
||||
Priority int `yaml:"priority"`
|
||||
Kind string `yaml:"kind"` // "chat" (default) | "image"
|
||||
Meta map[string]interface{} `yaml:"meta"`
|
||||
}
|
||||
|
||||
// Source describes a single upstream LLM provider.
|
||||
type Source struct {
|
||||
Name string `yaml:"name"`
|
||||
BaseURL string `yaml:"base_url"`
|
||||
APIKey string `yaml:"api_key"`
|
||||
Adapter string `yaml:"adapter"`
|
||||
Endpoint string `yaml:"endpoint"` // chat endpoint override
|
||||
ImageEndpoint string `yaml:"image_endpoint"` // image endpoint override
|
||||
Models []Model `yaml:"models"`
|
||||
Headers map[string]string `yaml:"headers"`
|
||||
Meta map[string]interface{} `yaml:"meta"`
|
||||
Temperature float64 `yaml:"temperature"`
|
||||
MaxTokens int `yaml:"max_tokens"`
|
||||
Timeout time.Duration `yaml:"timeout"`
|
||||
MaxConcurrent int `yaml:"max_concurrent"` // per-source inflight cap
|
||||
QueueTimeout time.Duration `yaml:"queue_timeout"` // wait for slot before failing
|
||||
}
|
||||
|
||||
// Load reads and validates a config file.
|
||||
func Load(path string) (*Config, error) {
|
||||
data, err := os.ReadFile(path)
|
||||
if err != nil {
|
||||
return nil, err
|
||||
}
|
||||
var cfg Config
|
||||
if err := yaml.Unmarshal(data, &cfg); err != nil {
|
||||
return nil, fmt.Errorf("parse config: %w", err)
|
||||
}
|
||||
if err := cfg.ApplyDefaults(); err != nil {
|
||||
return nil, err
|
||||
}
|
||||
return &cfg, nil
|
||||
}
|
||||
|
||||
// ApplyDefaults sets missing values and validates the config.
|
||||
func (c *Config) ApplyDefaults() error {
|
||||
if c.Listen == "" {
|
||||
c.Listen = ":8080"
|
||||
}
|
||||
if c.AdapterDir == "" {
|
||||
c.AdapterDir = "adapters"
|
||||
}
|
||||
if c.RuntimeFile == "" {
|
||||
c.RuntimeFile = "runtime.json"
|
||||
}
|
||||
if c.DefaultModel == "" {
|
||||
c.DefaultModel = "AUTO"
|
||||
}
|
||||
seen := map[string]bool{}
|
||||
modelOwners := map[string]string{}
|
||||
for i := range c.Sources {
|
||||
s := &c.Sources[i]
|
||||
if s.Name == "" {
|
||||
return fmt.Errorf("config: sources[%d] missing name", i)
|
||||
}
|
||||
if s.BaseURL == "" {
|
||||
return fmt.Errorf("config: source %s missing base_url", s.Name)
|
||||
}
|
||||
if s.Adapter == "" {
|
||||
s.Adapter = "openai"
|
||||
}
|
||||
if s.Timeout == 0 {
|
||||
s.Timeout = 120 * time.Second
|
||||
}
|
||||
if s.QueueTimeout == 0 {
|
||||
s.QueueTimeout = 60 * time.Second
|
||||
}
|
||||
if s.MaxConcurrent == 0 {
|
||||
s.MaxConcurrent = 8
|
||||
}
|
||||
if seen[s.Name] {
|
||||
return fmt.Errorf("config: duplicate source name %q", s.Name)
|
||||
}
|
||||
seen[s.Name] = true
|
||||
for j := range s.Models {
|
||||
m := &s.Models[j]
|
||||
if m.ID == "" {
|
||||
return fmt.Errorf("config: source %s has a model without id", s.Name)
|
||||
}
|
||||
if owner, ok := modelOwners[m.ID]; ok {
|
||||
return fmt.Errorf("config: model %q defined by both %s and %s", m.ID, owner, s.Name)
|
||||
}
|
||||
modelOwners[m.ID] = s.Name
|
||||
}
|
||||
}
|
||||
return nil
|
||||
}
|
||||
|
||||
// RuntimeConfig is the persisted web-UI editable slice (sources added/edited).
|
||||
type RuntimeConfig struct {
|
||||
Sources []Source `json:"sources"`
|
||||
}
|
||||
Reference in New Issue
Block a user