Merge pull request #13 from mostlygeek/set-content-length

Dechunk HTTP requests by default (#11)

config.example.yaml

@@ -7,7 +7,7 @@ models:
     cmd: >
       models/llama-server-osx
       --port 8999
-      -m models/Llama-3.2-1B-Instruct-Q4_K_M.gguf
+      -m models/Llama-3.2-1B-Instruct-Q4_0.gguf
     proxy: http://127.0.0.1:8999
 
     # list of model name aliases this llama.cpp instance can serve
@@ -18,7 +18,7 @@ models:
     checkEndpoint: /health
 
   "qwen":
-    cmd: models/llama-server-osx --port 8999 -m models/Qwen2.5-1.5B-Instruct-Q4_K_M.gguf
+    cmd: models/llama-server-osx --port 8999 -m models/qwen2.5-0.5b-instruct-q8_0.gguf
     proxy: http://127.0.0.1:8999
     aliases:
     - gpt-3.5-turbo
@@ -40,5 +40,5 @@ models:
     cmd: models/llama-server-osx --port 8999 -m models/doesnotexist.gguf
     proxy: http://127.0.0.1:8999
   "broken_timeout":
-    cmd: models/llama-server-osx --port 8999 -m models/Qwen2.5-1.5B-Instruct-Q4_K_M.gguf
+    cmd: models/llama-server-osx --port 8999 -m models/qwen2.5-0.5b-instruct-q8_0.gguf
     proxy: http://127.0.0.1:9000

proxy/process.go

@@ -178,7 +178,7 @@ func (p *Process) ProxyRequest(w http.ResponseWriter, r *http.Request) {
 		http.Error(w, err.Error(), http.StatusInternalServerError)
 		return
 	}
-	req.Header = r.Header
+	req.Header = r.Header.Clone()
 	resp, err := client.Do(req)
 	if err != nil {
 		http.Error(w, err.Error(), http.StatusBadGateway)

proxy/proxymanager.go

@@ -6,6 +6,7 @@ import (
 	"fmt"
 	"io"
 	"net/http"
+	"strconv"
 	"sync"
 	"time"
 
@@ -121,6 +122,11 @@ func (pm *ProxyManager) proxyChatRequestHandler(c *gin.Context) {
 	}
 
 	c.Request.Body = io.NopCloser(bytes.NewBuffer(bodyBytes))
+
+	// dechunk it as we already have all the body bytes see issue #11
+	c.Request.Header.Del("transfer-encoding")
+	c.Request.Header.Add("content-length", strconv.Itoa(len(bodyBytes)))
+
 	pm.currentProcess.ProxyRequest(c.Writer, c.Request)
 }