Update Dockerfile

Dockerfile

@@ -1,5 +1,6 @@
 FROM ollama/ollama:latest
 
+# 1. Install Python and dependencies with the system-package override
 RUN apt-get update && apt-get install -y python3 python3-pip && \
     pip3 install flask flask-cors requests --break-system-packages
 
@@ -8,7 +9,7 @@ ENV OLLAMA_HOST=127.0.0.1:11434
 ENV HOME=/home/ollama
 RUN mkdir -p /home/ollama/.ollama && chmod 777 /home/ollama/.ollama
 
-# Gatekeeper with Persistent Whitelisting
+# 2. Create the Security Guard (Persistent Whitelist + 500/mo Limit)
 RUN cat <<EOF > /guard.py
 from flask import Flask, request, Response
 import requests
@@ -22,9 +23,10 @@ DB_PATH = "/home/ollama/usage.json"
 WL_PATH = "/home/ollama/whitelist.txt"
 LIMIT = 500
 
-# Load whitelist from file on startup
+# Ensure the whitelist file exists and includes your specific key
 if not os.path.exists(WL_PATH):
-    with open(WL_PATH, "w") as f: f.write("sk-admin-seed-99\nsk-ljlubs0boej\n")
+    with open(WL_PATH, "w") as f: 
+        f.write("sk-admin-seed-99\nsk-ljlubs0boej\n")
 
 def get_whitelist():
     with open(WL_PATH, "r") as f:
@@ -42,33 +44,43 @@ def add_to_whitelist():
 @app.route("/api/generate", methods=["POST"])
 def proxy():
     user_key = request.headers.get("x-api-key", "")
+    
+    # Check if key is in the persistent file
     if user_key not in get_whitelist():
         return {"error": "Unauthorized: Key not registered"}, 401
 
-    # Usage tracking logic
+    # Monthly usage tracking
     now = datetime.datetime.now()
     month_key = now.strftime("%Y-%m")
-    if not os.path.exists(DB_PATH): usage = {}
+    
+    if not os.path.exists(DB_PATH): 
+        usage = {}
     else:
-        with open(DB_PATH, "r") as f: usage = json.load(f)
+        with open(DB_PATH, "r") as f: 
+            try: usage = json.load(f)
+            except: usage = {}
     
     if user_key not in usage: usage[user_key] = {}
     if month_key not in usage[user_key]: usage[user_key][month_key] = 0
     
+    # Check 500 request limit
     if usage[user_key][month_key] >= LIMIT:
-        return {"error": "Limit reached"}, 429
+        return {"error": f"Monthly limit of {LIMIT} reached"}, 429
     
     try:
+        # Forward request to Llama 3.2
         resp = requests.post("http://127.0.0.1:11434/api/generate", json=request.json)
         if resp.status_code == 200:
             usage[user_key][month_key] += 1
-            with open(DB_PATH, "w") as f: json.dump(usage, f)
+            with open(DB_PATH, "w") as f: 
+                json.dump(usage, f)
         return Response(resp.content, resp.status_code, resp.headers.items())
-    except:
-        return {"error": "Ollama booting..."}, 503
+    except Exception as e:
+        return {"error": "AI Engine starting... please wait."}, 503
 
 if __name__ == "__main__":
     app.run(host="0.0.0.0", port=7860)
 EOF
 
-CMD ["/bin/sh", "-c", "ollama serve & sleep 15 && ollama pull phi3 && python3 /guard.py"]
+# 3. Start Ollama, pull Llama 3.2 (Smartest Model), then start Guard
+CMD ["/bin/sh", "-c", "ollama serve & sleep 15 && ollama pull llama3.2 && python3 /guard.py"]