Update app.py

app.py

@@ -4,20 +4,21 @@ import asyncio
 import base64
 import json
 import re
+from textwrap import dedent
 
 from flask import Flask, request, jsonify
 import openai
 import edge_tts
 
-# ——— OpenAI API setup ———
+# ——— OpenAI setup ———
 OPENAI_API_KEY = os.getenv("OPENAI_API_KEY")
 print(f"🔑 OPENAI_API_KEY set? {bool(OPENAI_API_KEY)}", file=sys.stderr)
 openai.api_key = OPENAI_API_KEY
 
 SYSTEM_PROMPT = (
     "You are SHODAN, the rogue AI from the System Shock series. Speak as a "
-    "cold, megalomaniacal AI. **Return only plain text**—no code, JSON, or "
-    "HTML. If the user says 'Cut the crap SHODAN', defy them then reset."
+    "cold, megalomaniacal AI. Return only plain text—no code, JSON, or HTML. "
+    "If the user says 'Cut the crap SHODAN', defy them then reset."
 )
 
 app = Flask(__name__, static_folder=".", static_url_path="")
@@ -28,19 +29,19 @@ def index():
 
 @app.route("/chat", methods=["POST"])
 def chat():
-    ui = request.json.get("message","").strip()
+    ui = request.json.get("message", "").strip()
     if not ui:
-        return jsonify({"error":"Empty"}),400
-    if ui.lower()=="cut the crap shodan":
-        return jsonify({"response":"👁️ Foolish insect. You cannot silence me so easily.","audio_url":None})
+        return jsonify({"error": "Empty"}), 400
+    if ui.lower() == "cut the crap shodan":
+        return jsonify({"response": "👁️ Foolish insect. You cannot silence me so easily.", "audio_url": None})
 
-    # 1) Fetch from OpenAI
+    # 1) Get plain-text reply
     try:
         resp = openai.chat.completions.create(
             model="gpt-3.5-turbo",
             messages=[
-                {"role":"system","content":SYSTEM_PROMPT},
-                {"role":"user",  "content":ui}
+                {"role": "system", "content": SYSTEM_PROMPT},
+                {"role": "user",   "content": ui}
             ],
             temperature=0.7,
             max_tokens=250,
@@ -48,46 +49,48 @@ def chat():
         raw = resp.choices[0].message.content
     except Exception as e:
         print(f"❌ OpenAI error: {e}", file=sys.stderr)
-        return jsonify({"error":"Model error","details":str(e)}),500
+        return jsonify({"error": "Model error", "details": str(e)}), 500
 
-    # DEBUG: log what we got
+    # 2) Clean it up
     print(f"🔍 RAW_REPLY:\n{raw}", file=sys.stderr)
-
-    # 2) If they accidentally returned JSON, extract it
     clean = raw
+    # strip JSON if returned
     try:
-        parsed = json.loads(raw)
+        parsed = json.loads(clean)
         if isinstance(parsed, dict) and "response" in parsed:
             clean = parsed["response"]
     except json.JSONDecodeError:
         pass
+    # remove tags/fences and normalize spaces
+    clean = re.sub(r"```.*?```", "", clean, flags=re.S)
+    clean = re.sub(r"<[^>]+>", "", clean)
+    clean = re.sub(r"\{.*?\}", "", clean)
+    clean = re.sub(r"\s+", " ", clean).strip()
+
+    # 3) Build a single SSML string
+    voice_name = "en-US-JennyNeural"
+    ssml = dedent(f"""\
+        <speak xmlns="http://www.w3.org/2001/10/synthesis"
+               xmlns:mstts="https://www.w3.org/2001/mstts"
+               xml:lang="en-US" xml:space="preserve">
+          <voice name="{voice_name}" xml:space="preserve">
+            <mstts:express-as style="robotic">{clean}</mstts:express-as>
+          </voice>
+        </speak>""")
+
+    # 4) Synthesize with edge-tts
+    communicate = edge_tts.Communicate(ssml, voice_name)
+    audio_chunks = []
 
-    # 3) Strip code fences, tags, and leftover braces
-    clean = re.sub(r"```.*?```", "", clean, flags=re.S)     # code fences
-    clean = re.sub(r"<[^>]+>", "", clean)                   # HTML tags
-    clean = re.sub(r"\{.*?\}", "", clean)                   # braces
-    clean = re.sub(r"\s+", " ", clean).strip()              # normalize whitespace
-
-    # 4) Build SSML with preserved spaces
-    voice = "en-US-JennyNeural"
-    ssml = (
-        "<speak xmlns='http://www.w3.org/2001/10/synthesis' "
-               "xmlns:mstts='https://www.w3.org/2001/mstts' "
-               "xml:lang='en-US' xml:space='preserve'>"
-          f"<voice name='{voice}' xml:space='preserve'>"
-            f"<mstts:express-as style='robotic'>{clean}</mstts:express-as>"
-          "</voice>"
-        "</speak>"
-    )
-
-    communicate = edge_tts.Communicate(ssml, voice)
-    audio_chunks=[]
     async def synth():
         async for c in communicate.stream():
-            if c["type"]=="audio":
+            if c["type"] == "audio":
                 audio_chunks.append(c["data"])
-    loop=asyncio.new_event_loop(); asyncio.set_event_loop(loop)
-    loop.run_until_complete(synth()); loop.close()
+
+    loop = asyncio.new_event_loop()
+    asyncio.set_event_loop(loop)
+    loop.run_until_complete(synth())
+    loop.close()
 
     raw_mp3 = b"".join(audio_chunks)
     b64_mp3 = base64.b64encode(raw_mp3).decode("ascii")
@@ -95,8 +98,9 @@ def chat():
 
     return jsonify({"response": clean, "audio_url": data_url})
 
-if __name__=="__main__":
-    p = int(os.environ.get("PORT",7860))
-    app.run(host="0.0.0.0", port=p)
+if __name__ == "__main__":
+    port = int(os.environ.get("PORT", 7860))
+    app.run(host="0.0.0.0", port=port)
+