Update api.py

api.py

@@ -1,28 +1,28 @@
 from flask import Flask, request, jsonify
-from app import generate_text
+from app import generate_chat_completion
+import time
 
 app = Flask(__name__)
 
-@app.route('/generate', methods=['POST'])
-def generate():
+@app.route('/v1/chat/completions', methods=['POST'])
+def chat_completions():
     data = request.json
-    prompt = data.get('prompt', '')
+    messages = data.get('messages', [])
+    max_tokens = data.get('max_tokens', 560)
+    temperature = data.get('temperature', 0.7)
     
-    if not prompt:
-        return jsonify({"error": "Prompt is required"}), 400
+    if not messages:
+        return jsonify({"error": "messages array is required"}), 400
     
     try:
-        result = generate_text(
-            prompt,
-            max_new_tokens=data.get('max_new_tokens', 560),
-            max_context=data.get('max_context', 1080)
+        result = generate_chat_completion(
+            messages=messages,
+            max_tokens=max_tokens,
+            temperature=temperature
         )
-        return jsonify({
-            "generated_text": result,
-            "tokens_used": len(prompt.split()) + len(result.split())
-        })
+        return jsonify(result)
     except Exception as e:
         return jsonify({"error": str(e)}), 500
 
 if __name__ == '__main__':
-    app.run(host='0.0.0.0', port=8080)
+    app.run(host='0.0.0.0', port=8081)  # Note port 8081