Update app/services/reviewer_service.py

app/services/reviewer_service.py

@@ -1,3 +1,4 @@
+import time
 import json
 import logging
 import re
@@ -19,7 +20,6 @@ class AIReviewerService:
         self._init_db()
 
     def _init_db(self):
-        """Initializes SQLite cache table with proper timeout."""
         with self.db_lock:
             with sqlite3.connect(DB_PATH, timeout=30.0) as conn:
                 cursor = conn.cursor()
@@ -69,7 +69,6 @@ class AIReviewerService:
         uncached_files = []
         uncached_indices = []
 
-        # 1. Check SQLite cache
         for idx, f in enumerate(files):
             file_hash = self._generate_file_hash(f)
             cached_res = self._get_cached_result(file_hash)
@@ -83,14 +82,14 @@ class AIReviewerService:
             logger.info(f"⚡ Served all {len(files)} files instantly from SQLite cache.")
             return results
 
-        # 2. Parallel Sub-Batch Processing
         batches = [
             uncached_files[i : i + self.batch_size]
             for i in range(0, len(uncached_files), self.batch_size)
         ]
 
         new_results = []
-        with ThreadPoolExecutor(max_workers=5) as pool:
+        # Reduced max_workers to 2 to prevent hitting API Rate Limits
+        with ThreadPoolExecutor(max_workers=2) as pool:
             future_to_batch = {pool.submit(self._process_batch, batch): batch for batch in batches}
             for future in as_completed(future_to_batch):
                 try:
@@ -105,13 +104,12 @@ class AIReviewerService:
                             "vulnerabilities": [],
                             "improvement_suggestions": [],
                             "metrics": {"complexity": 0, "maintainability": 0.0},
-                            "error": str(e),
+                            "error": "RATE_LIMIT_EXCEEDED" if "429" in str(e) else str(e),
                         })
 
         new_results_map = {item.get("fileName"): item for item in new_results if isinstance(item, dict)}
         to_cache = []
 
-        # 3. Assemble final response and update cache
         for idx, f in zip(uncached_indices, uncached_files):
             res = new_results_map.get(
                 f.fileName,
@@ -131,6 +129,22 @@ class AIReviewerService:
         self._save_cached_results(to_cache)
         return results
 
+    def _call_llm_with_retry(self, prompt: str, max_tokens: int = 2048, retries: int = 3) -> str:
+        """Helper to retry LLM generation automatically when 429 rate limits occur."""
+        delay = 3.0
+        for attempt in range(retries):
+            try:
+                return llm_engine.generate(prompt, max_tokens=max_tokens)
+            except Exception as e:
+                err_str = str(e).upper()
+                if "429" in err_str or "RATE" in err_str or "LIMIT" in err_str:
+                    if attempt < retries - 1:
+                        logger.warning(f"⚠️ Rate limited (429). Retrying in {delay}s... (Attempt {attempt + 1}/{retries})")
+                        time.sleep(delay)
+                        delay *= 2  # Exponential backoff
+                        continue
+                raise e
+
     def _process_batch(self, batch: list) -> list:
         combined_code = ""
         file_names = []
@@ -156,7 +170,7 @@ Return ONLY valid JSON array:
 ]"""
         batch_results = []
         try:
-            response_text = llm_engine.generate(prompt, max_tokens=2048)
+            response_text = self._call_llm_with_retry(prompt, max_tokens=2048)
             parsed = self._parse_json(response_text)
             processed_map = {item.get("fileName"): item for item in parsed if isinstance(item, dict)}
 
@@ -182,7 +196,7 @@ Return ONLY valid JSON array:
                     "vulnerabilities": [],
                     "improvement_suggestions": [],
                     "metrics": {"complexity": 0, "maintainability": 0.0},
-                    "error": str(e),
+                    "error": "RATE_LIMIT_EXCEEDED" if "429" in str(e) else str(e),
                 })
         return batch_results
 
@@ -209,9 +223,7 @@ Return ONLY valid JSON array:
         if not text:
             return []
 
-        # Strip markdown code blocks safely without corrupting string content
         text = re.sub(r"```(?:json)?", "", text, flags=re.IGNORECASE).strip()
-
         start_idx = text.find('[')
         end_idx = text.rfind(']')