Update app.py

app.py

@@ -20,8 +20,6 @@ Major updates:
 import sys
 import subprocess
 from typing import Dict, Optional, Tuple, List
-import time
-import random
 
 def install(packages: List[str]):
     for package in packages:
@@ -39,6 +37,7 @@ install([
 # -----------------------------------------------------------------------------
 # 2. Static imports
 # -----------------------------------------------------------------------------
+import random
 import requests
 import json
 import tempfile
@@ -269,7 +268,7 @@ def get_technique_based_on_level(level: str) -> str:
     return random.choice(techniques.get(level, ["with slurs"]))
 
 # -----------------------------------------------------------------------------
-# 9. LLM Query Function (with enhanced error handling)
+# 9. LLM Query Function (supports multiple models)
 # -----------------------------------------------------------------------------
 def query_llm(model_name: str, prompt: str, instrument: str, level: str, key: str,
               time_sig: str, measures: int) -> str:
@@ -303,90 +302,7 @@ def query_llm(model_name: str, prompt: str, instrument: str, level: str, key: st
             "Sum must be exactly as specified. ONLY output the JSON array. No prose."
         )
 
-    # Retry up to 3 times for rate limited models
-    max_retries = 3
-    retry_delay = 5  # seconds
-    
-    for attempt in range(max_retries):
-        try:
-            if model_name == "Mistral":
-                headers = {
-                    "Authorization": f"Bearer {MISTRAL_API_KEY}",
-                    "Content-Type": "application/json",
-                }
-                payload = {
-                    "model": "mistral-medium",
-                    "messages": [
-                        {"role": "system", "content": system_prompt},
-                        {"role": "user", "content": user_prompt},
-                    ],
-                    "temperature": 0.7 if level == "Advanced" else 0.5,
-                    "max_tokens": 1000,
-                    "top_p": 0.95,
-                    "frequency_penalty": 0.2,
-                    "presence_penalty": 0.2,
-                }
-                response = requests.post(MISTRAL_API_URL, headers=headers, json=payload)
-                response.raise_for_status()
-                content = response.json()["choices"][0]["message"]["content"]
-                return content.replace("```json","").replace("```","").strip()
-            
-            elif model_name in ["DeepSeek", "Claude", "Gemma", "Kimi", "Llama 3.1"]:
-                client = OpenAI(
-                    base_url="https://openrouter.ai/api/v1",
-                    api_key=OPENROUTER_API_KEYS[model_name],
-                )
-                
-                model_map = {
-                    "DeepSeek": "deepseek/deepseek-chat-v3-0324:free",
-                    "Claude": "anthropic/claude-3.5-sonnet:beta",
-                    "Gemma": "google/gemma-3n-e2b-it:free",
-                    "Kimi": "moonshotai/kimi-dev-72b:free",
-                    "Llama 3.1": "meta-llama/llama-3.1-405b-instruct:free"
-                }
-                
-                # Special handling for Gemma API structure
-                if model_name == "Gemma":
-                    messages = [
-                        {"role": "user", "content": user_prompt}
-                    ]
-                else:
-                    messages = [
-                        {"role": "system", "content": system_prompt},
-                        {"role": "user", "content": user_prompt},
-                    ]
-                
-                completion = client.chat.completions.create(
-                    extra_headers={
-                        "HTTP-Referer": "https://github.com/AdaptiveMusicExerciseGenerator",
-                        "X-Title": "Music Exercise Generator",
-                    },
-                    model=model_map[model_name],
-                    messages=messages,
-                    temperature=0.7 if level == "Advanced" else 0.5,
-                    max_tokens=1000,
-                    top_p=0.95,
-                    frequency_penalty=0.2,
-                    presence_penalty=0.2,
-                )
-                content = completion.choices[0].message.content
-                return content.replace("```json","").replace("```","").strip()
-            
-            else:
-                return get_fallback_exercise(instrument, level, key, time_sig, measures)
-                
-        except Exception as e:
-            print(f"Error querying {model_name} API (attempt {attempt+1}): {e}")
-            if "429" in str(e) or "Rate limit" in str(e):
-                print(f"Rate limited, retrying in {retry_delay} seconds...")
-                time.sleep(retry_delay)
-                retry_delay *= 2  # Exponential backoff
-            else:
-                break
-    
-    # Fallback to Mistral if other APIs fail
-    print(f"All attempts failed for {model_name}, using Mistral fallback")
-    try:
+    if model_name == "Mistral":
         headers = {
             "Authorization": f"Bearer {MISTRAL_API_KEY}",
             "Content-Type": "application/json",
@@ -403,12 +319,53 @@ def query_llm(model_name: str, prompt: str, instrument: str, level: str, key: st
             "frequency_penalty": 0.2,
             "presence_penalty": 0.2,
         }
-        response = requests.post(MISTRAL_API_URL, headers=headers, json=payload)
-        response.raise_for_status()
-        content = response.json()["choices"][0]["message"]["content"]
-        return content.replace("```json","").replace("```","").strip()
-    except Exception as e:
-        print(f"Error querying Mistral fallback: {e}")
+        try:
+            response = requests.post(MISTRAL_API_URL, headers=headers, json=payload)
+            response.raise_for_status()
+            content = response.json()["choices"][0]["message"]["content"]
+            return content.replace("```json","").replace("```","").strip()
+        except Exception as e:
+            print(f"Error querying Mistral API: {e}")
+            return get_fallback_exercise(instrument, level, key, time_sig, measures)
+    
+    elif model_name in ["DeepSeek", "Claude", "Gemma", "Kimi", "Llama 3.1"]:
+        try:
+            client = OpenAI(
+                base_url="https://openrouter.ai/api/v1",
+                api_key=OPENROUTER_API_KEYS[model_name],
+            )
+            
+            model_map = {
+                "DeepSeek": "deepseek/deepseek-chat-v3-0324:free",
+                "Claude": "anthropic/claude-3.5-sonnet:beta",
+                "Gemma": "google/gemma-3n-e2b-it:free",
+                "Kimi": "moonshotai/kimi-dev-72b:free",
+                "Llama 3.1": "meta-llama/llama-3.1-405b-instruct:free"
+            }
+            
+            completion = client.chat.completions.create(
+                extra_headers={
+                    "HTTP-Referer": "https://github.com/AdaptiveMusicExerciseGenerator",
+                    "X-Title": "Music Exercise Generator",
+                },
+                model=model_map[model_name],
+                messages=[
+                    {"role": "system", "content": system_prompt},
+                    {"role": "user", "content": user_prompt},
+                ],
+                temperature=0.7 if level == "Advanced" else 0.5,
+                max_tokens=1000,
+                top_p=0.95,
+                frequency_penalty=0.2,
+                presence_penalty=0.2,
+            )
+            content = completion.choices[0].message.content
+            return content.replace("```json","").replace("```","").strip()
+        except Exception as e:
+            print(f"Error querying {model_name} API: {e}")
+            return get_fallback_exercise(instrument, level, key, time_sig, measures)
+    
+    else:
         return get_fallback_exercise(instrument, level, key, time_sig, measures)
 
 # -----------------------------------------------------------------------------
@@ -457,7 +414,7 @@ def generate_exercise(instrument: str, level: str, key: str, tempo: int, time_si
         return f"Error: {str(e)}", None, str(tempo), None, "0", time_signature, 0
 
 # -----------------------------------------------------------------------------
-# 12. AI chat assistant with enhanced error handling
+# 12. AI chat assistant
 # -----------------------------------------------------------------------------
 def handle_chat(message: str, history: List, instrument: str, level: str, ai_model: str):
     if not message.strip():
@@ -468,81 +425,54 @@ def handle_chat(message: str, history: List, instrument: str, level: str, ai_mod
         messages.append({"role": "assistant", "content": assistant_msg})
     messages.append({"role": "user", "content": message})
     
-    max_retries = 3
-    retry_delay = 3  # seconds
+    if ai_model == "Mistral":
+        headers = {"Authorization": f"Bearer {MISTRAL_API_KEY}", "Content-Type": "application/json"}
+        payload = {"model": "mistral-medium", "messages": messages, "temperature": 0.7, "max_tokens": 500}
+        try:
+            response = requests.post(MISTRAL_API_URL, headers=headers, json=payload)
+            response.raise_for_status()
+            content = response.json()["choices"][0]["message"]["content"]
+            history.append((message, content))
+            return "", history
+        except Exception as e:
+            history.append((message, f"Error: {str(e)}"))
+            return "", history
     
-    for attempt in range(max_retries):
+    elif ai_model in ["DeepSeek", "Claude", "Gemma", "Kimi", "Llama 3.1"]:
         try:
-            if ai_model == "Mistral":
-                headers = {"Authorization": f"Bearer {MISTRAL_API_KEY}", "Content-Type": "application/json"}
-                payload = {"model": "mistral-medium", "messages": messages, "temperature": 0.7, "max_tokens": 500}
-                response = requests.post(MISTRAL_API_URL, headers=headers, json=payload)
-                response.raise_for_status()
-                content = response.json()["choices"][0]["message"]["content"]
-                history.append((message, content))
-                return "", history
+            client = OpenAI(
+                base_url="https://openrouter.ai/api/v1",
+                api_key=OPENROUTER_API_KEYS[ai_model],
+            )
             
-            elif ai_model in ["DeepSeek", "Claude", "Gemma", "Kimi", "Llama 3.1"]:
-                client = OpenAI(
-                    base_url="https://openrouter.ai/api/v1",
-                    api_key=OPENROUTER_API_KEYS[ai_model],
-                )
-                
-                model_map = {
-                    "DeepSeek": "deepseek/deepseek-chat-v3-0324:free",
-                    "Claude": "anthropic/claude-3.5-sonnet:beta",
-                    "Gemma": "google/gemma-3n-e2b-it:free",
-                    "Kimi": "moonshotai/kimi-dev-72b:free",
-                    "Llama 3.1": "meta-llama/llama-3.1-405b-instruct:free"
-                }
-                
-                # Special handling for Gemma API structure
-                if ai_model == "Gemma":
-                    adjusted_messages = [{"role": "user", "content": msg["content"]} for msg in messages]
-                else:
-                    adjusted_messages = messages
-                
-                completion = client.chat.completions.create(
-                    extra_headers={
-                        "HTTP-Referer": "https://github.com/AdaptiveMusicExerciseGenerator",
-                        "X-Title": "Music Exercise Generator",
-                    },
-                    model=model_map[ai_model],
-                    messages=adjusted_messages,
-                    temperature=0.7,
-                    max_tokens=500,
-                )
-                content = completion.choices[0].message.content
-                history.append((message, content))
-                return "", history
+            model_map = {
+                "DeepSeek": "deepseek/deepseek-chat-v3-0324:free",
+                "Claude": "anthropic/claude-3.5-sonnet:beta",
+                "Gemma": "google/gemma-3n-e2b-it:free",
+                "Kimi": "moonshotai/kimi-dev-72b:free",
+                "Llama 3.1": "meta-llama/llama-3.1-405b-instruct:free"
+            }
             
-            else:
-                history.append((message, "Error: Invalid AI model selected"))
-                return "", history
-                
+            completion = client.chat.completions.create(
+                extra_headers={
+                    "HTTP-Referer": "https://github.com/AdaptiveMusicExerciseGenerator",
+                    "X-Title": "Music Exercise Generator",
+                },
+                model=model_map[ai_model],
+                messages=messages,
+                temperature=0.7,
+                max_tokens=500,
+            )
+            content = completion.choices[0].message.content
+            history.append((message, content))
+            return "", history
         except Exception as e:
-            print(f"Chat error with {ai_model} (attempt {attempt+1}): {e}")
-            if "429" in str(e) or "Rate limit" in str(e):
-                print(f"Rate limited, retrying in {retry_delay} seconds...")
-                time.sleep(retry_delay)
-                retry_delay *= 2  # Exponential backoff
-            else:
-                # Fallback to Mistral
-                print(f"Using Mistral fallback for chat")
-                try:
-                    headers = {"Authorization": f"Bearer {MISTRAL_API_KEY}", "Content-Type": "application/json"}
-                    payload = {"model": "mistral-medium", "messages": messages, "temperature": 0.7, "max_tokens": 500}
-                    response = requests.post(MISTRAL_API_URL, headers=headers, json=payload)
-                    response.raise_for_status()
-                    content = response.json()["choices"][0]["message"]["content"]
-                    history.append((message, content))
-                    return "", history
-                except Exception as e:
-                    history.append((message, f"Error: {str(e)}"))
-                    return "", history
-                    
-    history.append((message, "Error: All API attempts failed"))
-    return "", history
+            history.append((message, f"Error: {str(e)}"))
+            return "", history
+    
+    else:
+        history.append((message, "Error: Invalid AI model selected"))
+        return "", history
 
 # -----------------------------------------------------------------------------
 # 13. Gradio user interface definition