1.0.5

.github/dependabot.yml

@@ -13,7 +13,3 @@ updates:
     directory: "/" # Location of package manifests
     schedule:
       interval: "weekly"
-  - package-ecosystem: "cargo" # See documentation for possible values
-    directory: "/" # Location of package manifests
-    schedule:
-      interval: "weekly"

.github/workflows/compose.yaml

@@ -0,0 +1,21 @@
+services:
+  azure-oai-proxy:
+    # env_file: .env
+    image: 'gyarbij/azure-oai-proxy:latest'
+    # container_name: azure-oai-proxy
+    # Alternatively, use GitHub Container Registry:
+    # image: 'ghcr.io/gyarbij/azure-oai-proxy:latest'
+    restart: always
+    environment:
+      - AZURE_OPENAI_ENDPOINT=https://your-endpoint.openai.azure.com/
+    # - AZURE_OPENAI_PROXY_ADDRESS=0.0.0.0:11437
+    # - AZURE_OPENAI_PROXY_MODE=azure
+    # - AZURE_OPENAI_APIVERSION=2024-06-01
+    # - AZURE_OPENAI_MODEL_MAPPER=gpt-3.5-turbo=gpt-35-turbo,gpt-4=gpt-4-turbo
+    # - AZURE_AI_STUDIO_DEPLOYMENTS=mistral-large-2407=Mistral-large2:swedencentral,llama-3.1-405B=Meta-Llama-3-1-405B-Instruct:northcentralus,llama-3.1-70B=Llama-31-70B:swedencentral
+    # - AZURE_OPENAI_KEY_MISTRAL-LARGE-2407=your-api-key-1
+    # - AZURE_OPENAI_KEY_LLAMA-3.1-8B=your-api-key-2
+    # - AZURE_OPENAI_KEY_LLAMA-3.1-70B=your-api-key-3
+    ports:
+      - '11437:11437'
+    # Uncomment the following line to use an .env file:

.gitignore

@@ -14,4 +14,32 @@
 # Dependency directories (remove the comment below to include it)
 # vendor/
 
-.idea
+.idea
+
+# General
+.DS_Store
+.AppleDouble
+.LSOverride
+
+# Icon must end with two \r
+Icon
+
+
+# Thumbnails
+._*
+
+# Files that might appear in the root of a volume
+.DocumentRevisions-V100
+.fseventsd
+.Spotlight-V100
+.TemporaryItems
+.Trashes
+.VolumeIcon.icns
+.com.apple.timemachine.donotpresent
+
+# Directories potentially created on remote AFP share
+.AppleDB
+.AppleDesktop
+Network Trash Folder
+Temporary Items
+.apdisk

Dockerfile

@@ -1,6 +1,7 @@
 FROM golang:1.22.5 AS builder
 WORKDIR /build
 COPY . .
+RUN go get github.com/joho/godotenv
 RUN CGO_ENABLED=0 GOOS=linux go build -a -installsuffix cgo -o azure-oai-proxy .
 
 FROM gcr.io/distroless/base-debian12

example.env

@@ -0,0 +1,7 @@
+AZURE_OPENAI_APIVERSION=2024-06-01
+AZURE_OPENAI_ENDPOINT=https://your-azure-openai-resource.openai.azure.com/
+AZURE_OPENAI_API_KEY=your-azure-openai-api-key
+AZURE_OPENAI_MODEL_MAPPER=gpt-3.5-turbo=gpt-35-turbo,gpt-4=gpt-4-0613
+AZURE_AI_STUDIO_DEPLOYMENTS=mistral-large=Mistral-large2:swedencentral,llama-3=Meta-Llama-31-405B-Instruct:northcentralus
+AZURE_OPENAI_PROXY_ADDRESS=0.0.0.0:11437
+AZURE_OPENAI_PROXY_MODE=azure

go.mod

@@ -4,6 +4,7 @@ go 1.22.4
 
 require (
 	github.com/gin-gonic/gin v1.10.0
+	github.com/joho/godotenv v1.5.1
 	github.com/tidwall/gjson v1.17.1
 )
 

go.sum

@@ -28,6 +28,8 @@ github.com/goccy/go-json v0.10.2/go.mod h1:6MelG93GURQebXPDq3khkgXZkazVtN9CRI+MG
 github.com/google/go-cmp v0.5.5 h1:Khx7svrCpmxxtHBq5j2mp/xVjsi8hQMfNLvJFAlrGgU=
 github.com/google/go-cmp v0.5.5/go.mod h1:v8dTdLbMG2kIc/vJvl+f65V22dbkXbowE6jgT/gNBxE=
 github.com/google/gofuzz v1.0.0/go.mod h1:dBl0BpW6vV/+mYPU4Po3pmUjxk6FQPldtuIdl/M65Eg=
+github.com/joho/godotenv v1.5.1 h1:7eLL/+HRGLY0ldzfGMeQkb7vMd0as4CfYvUVzLqw0N0=
+github.com/joho/godotenv v1.5.1/go.mod h1:f4LDr5Voq0i2e/R5DDNOoa2zzDfwtkZa6DnEwAbqwq4=
 github.com/json-iterator/go v1.1.12 h1:PV8peI4a0ysnczrg+LtxykD8LfKY9ML6u2jnxaEnrnM=
 github.com/json-iterator/go v1.1.12/go.mod h1:e30LSqwooZae/UwlEbR2852Gd8hjQvJoHmT4TnhNGBo=
 github.com/klauspost/cpuid/v2 v2.0.9/go.mod h1:FInQzS24/EEf25PyTYn52gqo7WaD8xa0213Md/qVLRg=

main.go

@@ -7,10 +7,12 @@ import (
 	"log"
 	"net/http"
 	"os"
+	"strings"
 
 	"github.com/gin-gonic/gin"
 	"github.com/gyarbij/azure-oai-proxy/pkg/azure"
 	"github.com/gyarbij/azure-oai-proxy/pkg/openai"
+	"github.com/joho/godotenv"
 )
 
 var (
@@ -49,6 +51,11 @@ type Deprecation struct {
 }
 
 func init() {
+	// Load .env file if it exists
+	if err := godotenv.Load(); err != nil {
+		log.Println("No .env file found")
+	}
+
 	gin.SetMode(gin.ReleaseMode)
 	if v := os.Getenv("AZURE_OPENAI_PROXY_ADDRESS"); v != "" {
 		Address = v
@@ -58,6 +65,16 @@ func init() {
 	}
 	log.Printf("loading azure openai proxy address: %s", Address)
 	log.Printf("loading azure openai proxy mode: %s", ProxyMode)
+
+	// Load Azure OpenAI Model Mapper
+	if v := os.Getenv("AZURE_OPENAI_MODEL_MAPPER"); v != "" {
+		for _, pair := range strings.Split(v, ",") {
+			info := strings.Split(pair, "=")
+			if len(info) == 2 {
+				azure.AzureOpenAIModelMapper[info[0]] = info[1]
+			}
+		}
+	}
 }
 
 func main() {
@@ -109,6 +126,22 @@ func handleGetModels(c *gin.Context) {
 		c.JSON(http.StatusInternalServerError, gin.H{"error": "failed to fetch deployed models"})
 		return
 	}
+
+	// Add serverless deployments to the models list
+	for deploymentName := range azure.ServerlessDeploymentInfo {
+		models = append(models, Model{
+			ID:     deploymentName,
+			Object: "model",
+			Capabilities: Capabilities{
+				Completion:     true,
+				ChatCompletion: true,
+				Inference:      true,
+			},
+			LifecycleStatus: "active",
+			Status:          "ready",
+		})
+	}
+
 	result := ModelList{
 		Object: "list",
 		Data:   models,