feat(dedup): implement two-stage duplicate detection and content rewriter

core/dedup.py

@@ -0,0 +1,35 @@
+import hashlib
+import re
+from typing import Optional
+
+def normalize_text(text: Optional[str]) -> str:
+    """Normalize text by removing URLs, telegram handles, hashtags, excessive punctuation/whitespace, and lowercasing."""
+    if not text:
+        return ""
+    # Remove URLs
+    text = re.sub(r'https?://\S+|www\.\S+', '', text)
+    # Remove Telegram @mentions / hashtags
+    text = re.sub(r'[@#]\w+', '', text)
+    # Normalize punctuation and whitespace
+    text = re.sub(r'[^\w\s]', ' ', text)
+    text = re.sub(r'\s+', ' ', text)
+    return text.strip().lower()
+
+def compute_content_hash(text: Optional[str], media_hash: Optional[str] = None) -> Optional[str]:
+    """Generate SHA256 hash from normalized text and/or media hash."""
+    norm_text = normalize_text(text)
+    if not norm_text and not media_hash:
+        return None
+    raw_key = f"{norm_text}|{media_hash or ''}"
+    return hashlib.sha256(raw_key.encode('utf-8')).hexdigest()
+
+def compute_file_hash(file_path: str) -> Optional[str]:
+    """Generate SHA256 hash of a media file."""
+    try:
+        hasher = hashlib.sha256()
+        with open(file_path, 'rb') as f:
+            while chunk := f.read(65536):
+                hasher.update(chunk)
+        return hasher.hexdigest()
+    except Exception:
+        return None

core/llm.py

@@ -0,0 +1,91 @@
+import os
+import json
+import time
+import httpx
+import logging
+from typing import Dict, Any, Optional
+from core.metrics import AI_REQUESTS_TOTAL, AI_LATENCY_SECONDS
+
+logger = logging.getLogger(__name__)
+
+class LLMClient:
+    def __init__(
+        self,
+        provider: Optional[str] = None,
+        api_key: Optional[str] = None,
+        model: Optional[str] = None,
+        base_url: Optional[str] = None,
+    ):
+        self.provider = provider or os.getenv("AI_PROVIDER", "gemini").lower()
+        self.api_key = api_key or os.getenv("AI_API_KEY", "")
+        self.model = model or os.getenv("AI_MODEL", "gemini-1.5-flash" if self.provider == "gemini" else "gpt-4o-mini")
+        self.base_url = base_url or os.getenv("AI_BASE_URL")
+
+    async def generate_json(self, prompt: str, system_prompt: Optional[str] = None, action_name: str = "general") -> Dict[str, Any]:
+        """Send prompt to LLM and parse JSON response."""
+        start_time = time.time()
+        status = "error"
+        try:
+            if self.provider == "gemini":
+                result = await self._call_gemini(prompt, system_prompt)
+            else:
+                result = await self._call_openai(prompt, system_prompt)
+            status = "success"
+            return result
+        except Exception as e:
+            logger.error(f"LLM generation failed ({self.provider}/{self.model}): {e}")
+            raise
+        finally:
+            duration = time.time() - start_time
+            AI_LATENCY_SECONDS.labels(action=action_name).observe(duration)
+            AI_REQUESTS_TOTAL.labels(action=action_name, status=status).inc()
+
+    async def _call_gemini(self, prompt: str, system_prompt: Optional[str] = None) -> Dict[str, Any]:
+        url = f"https://generativelanguage.googleapis.com/v1beta/models/{self.model}:generateContent?key={self.api_key}"
+        payload: Dict[str, Any] = {
+            "contents": [
+                {
+                    "parts": [{"text": prompt}]
+                }
+            ],
+            "generationConfig": {
+                "responseMimeType": "application/json",
+                "temperature": 0.2,
+            }
+        }
+        if system_prompt:
+            payload["systemInstruction"] = {
+                "parts": [{"text": system_prompt}]
+            }
+
+        async with httpx.AsyncClient(timeout=60.0) as client:
+            resp = await client.post(url, json=payload)
+            resp.raise_for_status()
+            data = resp.json()
+            raw_text = data["candidates"][0]["content"]["parts"][0]["text"]
+            return json.loads(raw_text)
+
+    async def _call_openai(self, prompt: str, system_prompt: Optional[str] = None) -> Dict[str, Any]:
+        url = self.base_url or "https://api.openai.com/v1/chat/completions"
+        headers = {
+            "Authorization": f"Bearer {self.api_key}",
+            "Content-Type": "application/json"
+        }
+        messages = []
+        if system_prompt:
+            messages.append({"role": "system", "content": system_prompt})
+        messages.append({"role": "user", "content": prompt})
+
+        payload = {
+            "model": self.model,
+            "messages": messages,
+            "response_format": {"type": "json_object"},
+            "temperature": 0.2,
+        }
+
+        async with httpx.AsyncClient(timeout=60.0) as client:
+            resp = await client.post(url, headers=headers, json=payload)
+            resp.raise_for_status()
+            data = resp.json()
+            content = data["choices"][0]["message"]["content"]
+            return json.loads(content)