feat(llm): add 90s cooldown + auto-retry on OpenRouter rate limit (429)

app.py

@@ -39,6 +39,27 @@ from dotenv import load_dotenv
 
 load_dotenv()
 
+# ──────────────────────────────────────────────
+# OpenRouter Retry Helper
+# ──────────────────────────────────────────────
+
+def _openrouter_with_retry(api_call_fn, max_retries=1):
+    """
+    Retry an OpenRouter API call with 90-second cooldown on rate limit (HTTP 429).
+    Falls through to caller's exception handling after max retries exhausted.
+    """
+    import time
+    for attempt in range(max_retries + 1):
+        try:
+            return api_call_fn()
+        except Exception as e:
+            error_str = str(e)
+            if "429" in error_str and attempt < max_retries:
+                print(f"   [RATE LIMIT] OpenRouter rate limited (429). Cooling down for 90 seconds... (retry {attempt + 1}/{max_retries})")
+                time.sleep(90)
+            else:
+                raise
+
 # ──────────────────────────────────────────────
 # App Setup
 # ──────────────────────────────────────────────
@@ -328,7 +349,11 @@ def screen_article(title, summary, criteria):
         if not key: return True
         prompt = f"Decide if this article matches criteria: {criteria}\nTitle: {title}\nSummary: {summary}\nRespond ONLY with JSON: {{\"publish\": true/false, \"reason\": \"...\"}}"
         client = OpenAI(base_url="https://openrouter.ai/api/v1", api_key=key)
-        resp = client.chat.completions.create(model="qwen/qwen3.6-plus:free", messages=[{"role":"user","content":prompt}], response_format={"type":"json_object"})
+
+        def _do_screen():
+            return client.chat.completions.create(model="qwen/qwen3.6-plus:free", messages=[{"role":"user","content":prompt}], response_format={"type":"json_object"})
+
+        resp = _openrouter_with_retry(_do_screen)
         result = json.loads(resp.choices[0].message.content)
         print(f"[AGENT] Screening: {result.get('publish')} - {result.get('reason')}")
         return result.get("publish", True)

article_rewriter.py

@@ -51,6 +51,32 @@ load_dotenv()
 
 OPENROUTER_API_KEY = os.getenv("OPENROUTER_API_KEY", "")
 GROQ_API_KEY = os.getenv("GROQ_API_KEY", "")
+
+
+# ──────────────────────────────────────────────
+# 0b. OpenRouter Retry Helper
+# ──────────────────────────────────────────────
+
+def _openrouter_with_retry(api_call_fn, max_retries=1):
+    """
+    Retry an OpenRouter API call with 90-second cooldown on rate limit (HTTP 429).
+    Falls through to caller's exception handling after max retries exhausted.
+    """
+    for attempt in range(max_retries + 1):
+        try:
+            return api_call_fn()
+        except Exception as e:
+            error_str = str(e)
+            if "429" in error_str and attempt < max_retries:
+                print(f"   [RATE LIMIT] OpenRouter rate limited (429). Cooling down for 90 seconds... (retry {attempt + 1}/{max_retries})")
+                time.sleep(90)
+            else:
+                raise
+
+# ──────────────────────────────────────────────
+# Remaining Configuration
+# ──────────────────────────────────────────────
+
 WP_URL = os.getenv("WP_URL", "").rstrip("/")
 WP_USERNAME = os.getenv("WP_USERNAME", "")
 WP_APP_PASSWORD = os.getenv("WP_APP_PASSWORD", "")
@@ -590,21 +616,24 @@ def rewrite_article(original_title: str, original_body: str, source_url: str) ->
         try:
             print(f"✍️  Rewriting with OpenRouter ({LLM_MODEL})...")
             client = OpenAI(base_url="https://openrouter.ai/api/v1", api_key=OPENROUTER_API_KEY)
-            
-            completion = client.chat.completions.create(
-                model=LLM_MODEL,
-                messages=[
-                    {"role": "system", "content": "You are an expert content writer and SEO specialist."},
-                    {"role": "user", "content": prompt}
-                ],
-                temperature=0.8,
-                response_format={"type": "json_object"},
-                extra_body={
-                    "reasoning": {
-                        "max_tokens": 2048
+
+            def _do_rewrite():
+                return client.chat.completions.create(
+                    model=LLM_MODEL,
+                    messages=[
+                        {"role": "system", "content": "You are an expert content writer and SEO specialist."},
+                        {"role": "user", "content": prompt}
+                    ],
+                    temperature=0.8,
+                    response_format={"type": "json_object"},
+                    extra_body={
+                        "reasoning": {
+                            "max_tokens": 2048
+                        }
                     }
-                }
-            )
+                )
+
+            completion = _openrouter_with_retry(_do_rewrite)
             raw_text = completion.choices[0].message.content.strip()
             
             # Log reasoning tokens (OpenRouter specific)
@@ -853,13 +882,17 @@ def filter_irrelevant_images(image_urls: list[str], article_title: str, article_
             try:
                 print(f"   [AI-FILTER] Trying OpenRouter (qwen/qwen3.6-plus:free)...")
                 client = OpenAI(base_url="https://openrouter.ai/api/v1", api_key=openrouter_key)
-                completion = client.chat.completions.create(
-                    model="qwen/qwen3.6-plus:free",
-                    messages=[{"role": "user", "content": prompt}],
-                    temperature=0.1,
-                    response_format={"type": "json_object"},
-                    extra_body={"reasoning": {"max_tokens": 512}}
-                )
+
+                def _do_filter():
+                    return client.chat.completions.create(
+                        model="qwen/qwen3.6-plus:free",
+                        messages=[{"role": "user", "content": prompt}],
+                        temperature=0.1,
+                        response_format={"type": "json_object"},
+                        extra_body={"reasoning": {"max_tokens": 512}}
+                    )
+
+                completion = _openrouter_with_retry(_do_filter)
                 raw_text = completion.choices[0].message.content.strip()
             except Exception as e:
                 print(f"   [WARN] OpenRouter image filter failed: {e}")

main.py

@@ -707,6 +707,27 @@ def extract_frames(video_path: Path, count: int = 3) -> list[Image.Image]:
     return frames
 
 
+# ──────────────────────────────────────────────
+# 2b. OpenRouter Retry Helper
+# ──────────────────────────────────────────────
+
+def _openrouter_with_retry(api_call_fn, max_retries=1):
+    """
+    Retry an OpenRouter API call with 90-second cooldown on rate limit (HTTP 429).
+    Falls through to caller's exception handling after max retries exhausted.
+    """
+    for attempt in range(max_retries + 1):
+        try:
+            return api_call_fn()
+        except Exception as e:
+            error_str = str(e)
+            if "429" in error_str and attempt < max_retries:
+                print(f"   [RATE LIMIT] OpenRouter rate limited (429). Cooling down for 90 seconds... (retry {attempt + 1}/{max_retries})")
+                time.sleep(90)
+            else:
+                raise
+
+
 # ──────────────────────────────────────────────
 # 3. Video Analysis with Gemini
 # ──────────────────────────────────────────────
@@ -765,7 +786,7 @@ def _analyze_frames(frames: list[Image.Image]) -> dict:
         try:
             print(f"   [ENGINE] Trying OpenRouter ({LLM_MODEL})...")
             client = OpenAI(base_url="https://openrouter.ai/api/v1", api_key=openrouter_key)
-            
+
             messages_content = [{"type": "text", "text": prompt}]
             for frame in frames[:60]:
                 b64_image = _image_to_base64(frame)
@@ -774,17 +795,20 @@ def _analyze_frames(frames: list[Image.Image]) -> dict:
                     "image_url": {"url": f"data:image/jpeg;base64,{b64_image}"}
                 })
 
-            completion = client.chat.completions.create(
-                model=LLM_MODEL,
-                messages=[{"role": "user", "content": messages_content}],
-                temperature=0.2,
-                response_format={"type": "json_object"},
-                extra_body={
-                    "reasoning": {
-                        "max_tokens": 2048
+            def _do_analyze():
+                return client.chat.completions.create(
+                    model=LLM_MODEL,
+                    messages=[{"role": "user", "content": messages_content}],
+                    temperature=0.2,
+                    response_format={"type": "json_object"},
+                    extra_body={
+                        "reasoning": {
+                            "max_tokens": 2048
+                        }
                     }
-                }
-            )
+                )
+
+            completion = _openrouter_with_retry(_do_analyze)
             raw_text = completion.choices[0].message.content.strip()
             
             # Log reasoning tokens if available (OpenRouter specific)
@@ -897,18 +921,21 @@ def generate_content(analysis: any, source_url: str) -> dict:
         try:
             print(f"[CONTENT] Generating with OpenRouter ({LLM_MODEL})...")
             client = OpenAI(base_url="https://openrouter.ai/api/v1", api_key=openrouter_key)
-            
-            completion = client.chat.completions.create(
-                model=LLM_MODEL,
-                messages=[{"role": "user", "content": prompt}],
-                temperature=0.7,
-                response_format={"type": "json_object"},
-                extra_body={
-                    "reasoning": {
-                        "max_tokens": 2048
+
+            def _do_generate():
+                return client.chat.completions.create(
+                    model=LLM_MODEL,
+                    messages=[{"role": "user", "content": prompt}],
+                    temperature=0.7,
+                    response_format={"type": "json_object"},
+                    extra_body={
+                        "reasoning": {
+                            "max_tokens": 2048
+                        }
                     }
-                }
-            )
+                )
+
+            completion = _openrouter_with_retry(_do_generate)
             raw_text = completion.choices[0].message.content.strip()
             
             # Log reasoning tokens (OpenRouter specific)