- 2.3: unknown model errors (400 + RFC 9457 problem+json with valid_models) - 2.5: GET /v1/models endpoint (derived from config, not hardcoded) - 2.6: POST /v1/embeddings passthrough (body-based dispatch, no rewrite) - 2.7: POST /v1/rerank with path rewrite (/v1/rerank → /rerank) - wire proxy.Handler in main.go (was using dummy handler) - 140+ tests passing, race detector clean - all requests: client → nginx → gateway → upstreams - ready for config deployment to go live
157 lines
4.4 KiB
Go
157 lines
4.4 KiB
Go
// Package proxy provides request routing and forwarding.
|
|
package proxy
|
|
|
|
import (
|
|
"bytes"
|
|
"encoding/json"
|
|
"fmt"
|
|
"io"
|
|
"net/http"
|
|
"net/url"
|
|
|
|
"github.com/Riotpiaole/homelab-frontend/internal/config"
|
|
)
|
|
|
|
// RouteRequest determines which upstream should handle the request.
|
|
// For /v1/chat/completions, /v1/embeddings, and /v1/rerank, it uses body-based dispatch.
|
|
// For other routes, it looks up by path in the configured routes.
|
|
func (h *Handler) RouteRequest(r *http.Request) (*Route, error) {
|
|
// For /v1/chat/completions, /v1/embeddings, /v1/rerank use body-based dispatch
|
|
if r.Method == "POST" && (r.URL.Path == "/v1/chat/completions" || r.URL.Path == "/v1/embeddings" || r.URL.Path == "/v1/rerank") {
|
|
return h.routeByModel(r, r.URL.Path)
|
|
}
|
|
|
|
// For other paths, try to find a matching route by path
|
|
// Look for exact path match or path prefix match
|
|
for _, route := range h.routes {
|
|
// Check if route's pathRewrite matches request path
|
|
if route.Upstream.PathRewrite != "" && route.Upstream.PathRewrite == r.URL.Path {
|
|
return route, nil
|
|
}
|
|
}
|
|
|
|
// If no routes configured, return error
|
|
if len(h.routes) == 0 {
|
|
return nil, fmt.Errorf("no routes configured")
|
|
}
|
|
|
|
// Return the first configured route as fallback
|
|
for _, route := range h.routes {
|
|
return route, nil
|
|
}
|
|
|
|
return nil, fmt.Errorf("no route available")
|
|
}
|
|
|
|
// routeByModel reads the request body to find the "model" field and routes accordingly.
|
|
// The body is preserved for forwarding to the upstream.
|
|
// Returns a modelValidationError for client errors (invalid JSON, missing/unknown model).
|
|
// The path parameter indicates which endpoint is being called (/v1/chat/completions, /v1/embeddings, /v1/rerank)
|
|
func (h *Handler) routeByModel(r *http.Request, path string) (*Route, error) {
|
|
// If there's no body, we can't determine the model
|
|
if r.Body == nil {
|
|
return nil, &modelValidationError{
|
|
Kind: "missing_model",
|
|
Message: "request body required",
|
|
}
|
|
}
|
|
|
|
// Read the body to extract the model name
|
|
// We need to be careful to preserve the body for the upstream
|
|
bodyBytes, err := io.ReadAll(r.Body)
|
|
if err != nil {
|
|
return nil, &modelValidationError{
|
|
Kind: "invalid_request",
|
|
Message: fmt.Sprintf("failed to read request body: %v", err),
|
|
}
|
|
}
|
|
|
|
// Restore the body so it can be read again by the upstream
|
|
r.Body = io.NopCloser(bytes.NewReader(bodyBytes))
|
|
|
|
// Parse the JSON to find the model field
|
|
var payload map[string]interface{}
|
|
if err := json.Unmarshal(bodyBytes, &payload); err != nil {
|
|
return nil, &modelValidationError{
|
|
Kind: "invalid_json",
|
|
Message: "request body is not valid JSON",
|
|
}
|
|
}
|
|
|
|
// Extract the model name
|
|
modelVal, hasModel := payload["model"]
|
|
if !hasModel {
|
|
return nil, &modelValidationError{
|
|
Kind: "missing_model",
|
|
Message: "'model' field is missing",
|
|
}
|
|
}
|
|
|
|
// Handle null model
|
|
if modelVal == nil {
|
|
return nil, &modelValidationError{
|
|
Kind: "null_model",
|
|
Message: "'model' field is null",
|
|
}
|
|
}
|
|
|
|
// Extract as string
|
|
modelName, ok := modelVal.(string)
|
|
if !ok {
|
|
return nil, &modelValidationError{
|
|
Kind: "missing_model",
|
|
Message: "'model' field must be a string",
|
|
}
|
|
}
|
|
|
|
// Handle empty string
|
|
if modelName == "" {
|
|
return nil, &modelValidationError{
|
|
Kind: "empty_model",
|
|
Message: "'model' field cannot be empty",
|
|
}
|
|
}
|
|
|
|
// Look up the model in the registry
|
|
modelUpstream := h.config.LookupModel(modelName)
|
|
if modelUpstream == nil {
|
|
return nil, &modelValidationError{
|
|
Kind: "unknown_model",
|
|
Message: fmt.Sprintf("unknown model: %q", modelName),
|
|
Model: modelName,
|
|
}
|
|
}
|
|
|
|
// Determine the upstream path based on the request path
|
|
upstreamPath := path
|
|
if path == "/v1/rerank" {
|
|
// Rerank endpoint uses /rerank path on upstream
|
|
upstreamPath = "/rerank"
|
|
}
|
|
|
|
// Create a route for this model with appropriate timeouts
|
|
// These are sensible defaults for LLM models
|
|
upstreamCfg := &config.Upstream{
|
|
Address: modelUpstream.Address,
|
|
PathRewrite: upstreamPath,
|
|
ConnectTimeout: h.defaultConnectTimeout,
|
|
ReadTimeout: h.defaultReadTimeout,
|
|
WriteTimeout: h.defaultWriteTimeout,
|
|
MaxBodySize: h.defaultMaxBodySize,
|
|
AuthRequired: false,
|
|
}
|
|
|
|
targetURL, _ := url.Parse("http://" + modelUpstream.Address)
|
|
|
|
route := &Route{
|
|
Name: "v1-chat-" + modelName,
|
|
Upstream: upstreamCfg,
|
|
Transport: h.getOrCreateTransport(modelUpstream.Address, upstreamCfg),
|
|
Director: func(req *http.Request) {
|
|
directorFunc(req, targetURL, upstreamCfg)
|
|
},
|
|
}
|
|
|
|
return route, nil
|
|
}
|