superbrain-server 1.0.2-beta.0

package/payload/analyzers/text_analyzer.py

@@ -0,0 +1,117 @@
+#!/usr/bin/env python3
+"""
+Text Content Analyzer
+Analyzes text files (like info.txt from Instagram posts) using the ModelRouter.
+Routes to the best available free model: Groq → Gemini → OpenRouter → Local Ollama.
+"""
+
+import sys
+from pathlib import Path
+
+# Ensure backend root is in sys.path (needed when run as a subprocess)
+sys.path.insert(0, str(Path(__file__).resolve().parent.parent))
+
+from core.model_router import get_router
+
+
+def analyze_text(file_path: str) -> dict:
+    """
+    Analyze a text file using the best available AI model.
+
+    Returns:
+        dict with keys: success (bool), file (str), analysis (str), content (str)
+                     or error (str) on failure.
+    """
+    path = Path(file_path)
+
+    if not path.exists():
+        return {"success": False, "error": f"File not found: {file_path}"}
+    if not path.is_file():
+        return {"success": False, "error": f"Not a file: {file_path}"}
+
+    try:
+        with open(path, "r", encoding="utf-8") as f:
+            content = f.read().strip()
+    except Exception as e:
+        return {"success": False, "error": f"Could not read file: {e}"}
+
+    if not content:
+        return {"success": False, "error": "File is empty"}
+
+    prompt = f"""You are analyzing an Instagram post's metadata. Read the following information and provide insights.
+
+POST INFORMATION:
+{content}
+
+TASK: Analyze this post and provide:
+1. Main topic / theme
+2. Content type (travel guide, vlog, educational, product review, etc.)
+3. Key highlights and takeaways
+4. Notable details (locations, products, people, links, tips mentioned)
+5. Target audience
+
+Be specific and concise. Max 300 words."""
+
+    print("🔄 Generating text analysis...")
+
+    router = get_router()
+    try:
+        analysis = router.generate_text(prompt)
+    except RuntimeError as e:
+        return {"success": False, "error": f"All models failed: {e}"}
+
+    if not analysis:
+        return {"success": False, "error": "Model returned empty response"}
+
+    return {
+        "success": True,
+        "file": path.name,
+        "analysis": analysis,
+        "content": content,
+    }
+
+
+def main():
+    """Main entry point."""
+
+    if len(sys.argv) > 1:
+        file_path = sys.argv[1]
+    else:
+        print("=" * 70)
+        print("📄 TEXT ANALYZER")
+        print("=" * 70)
+        print()
+        file_path = input("Enter text file path: ").strip()
+
+    file_path = file_path.strip('"').strip("'").strip()
+
+    if not file_path:
+        print("❌ No path provided!")
+        return
+
+    print()
+    result = analyze_text(file_path)
+
+    if result["success"]:
+        print("=" * 70)
+        print("📊 TEXT ANALYSIS RESULTS")
+        print("=" * 70)
+        print()
+        print(f"📄 File: {result['file']}")
+        print()
+        print("📝 ORIGINAL CONTENT:")
+        print("-" * 70)
+        print(result["content"])
+        print("-" * 70)
+        print()
+        print("🔍 ANALYSIS:")
+        print("-" * 70)
+        print(result["analysis"])
+        print("-" * 70)
+        print()
+    else:
+        print(f"❌ Error: {result['error']}")
+
+
+if __name__ == "__main__":
+    main()

package/payload/analyzers/visual_analyze.py

@@ -0,0 +1,218 @@
+#!/usr/bin/env python3
+"""
+Video & Image Visual Analyzer
+Uses the ModelRouter to pick the best available free vision model
+(Gemini → Groq Vision → OpenRouter Vision → Local Ollama as last resort)
+
+For videos: extracts key frames using OpenCV (or ffmpeg fallback),
+then sends them to the vision model.
+"""
+
+import sys
+import base64
+from pathlib import Path
+from typing import List
+
+# Ensure backend root is in sys.path (needed when run as a subprocess)
+sys.path.insert(0, str(Path(__file__).resolve().parent.parent))
+
+from core.model_router import get_router
+
+
+# ─────────────────────────────────────────────────────────────────────────────
+#  Frame extraction helpers
+# ─────────────────────────────────────────────────────────────────────────────
+
+def _frames_cv2(video_path: str, max_frames: int = 4) -> List[str]:
+    """Extract evenly-spaced frames using OpenCV → base64 JPEG list."""
+    import cv2
+    cap = cv2.VideoCapture(video_path)
+    total = int(cap.get(cv2.CAP_PROP_FRAME_COUNT))
+    if total <= 0:
+        cap.release()
+        return []
+    indices = [int(i * total / max_frames) for i in range(max_frames)]
+    out = []
+    for idx in indices:
+        cap.set(cv2.CAP_PROP_POS_FRAMES, idx)
+        ret, frame = cap.read()
+        if not ret:
+            continue
+        ok, buf = cv2.imencode(".jpg", frame, [cv2.IMWRITE_JPEG_QUALITY, 75])
+        if ok:
+            out.append(base64.b64encode(buf.tobytes()).decode())
+    cap.release()
+    return out
+
+
+def _get_duration_s(video_path: str) -> int:
+    import subprocess
+    r = subprocess.run(
+        ["ffprobe", "-v", "error", "-select_streams", "v:0",
+         "-show_entries", "format=duration", "-of", "csv=p=0", video_path],
+        capture_output=True, text=True,
+    )
+    try:
+        return max(1, int(float(r.stdout.strip())))
+    except Exception:
+        return 60
+
+
+def _frames_ffmpeg(video_path: str, max_frames: int = 4) -> List[str]:
+    """Fallback frame extraction via ffmpeg."""
+    import subprocess, tempfile, os
+    out = []
+    with tempfile.TemporaryDirectory() as tmpdir:
+        dur = _get_duration_s(video_path)
+        interval = max(1, dur // max_frames)
+        pattern = os.path.join(tmpdir, "frame_%02d.jpg")
+        subprocess.run(
+            ["ffmpeg", "-i", video_path, "-vf", f"fps=1/{interval}",
+             "-vframes", str(max_frames), "-q:v", "3", pattern],
+            capture_output=True,
+        )
+        for fname in sorted(os.listdir(tmpdir)):
+            with open(os.path.join(tmpdir, fname), "rb") as f:
+                out.append(base64.b64encode(f.read()).decode())
+    return out
+
+
+def extract_frames(video_path: str, max_frames: int = 4) -> List[str]:
+    """Extract frames from a video, trying OpenCV first then ffmpeg."""
+    try:
+        import cv2  # noqa: F401
+        frames = _frames_cv2(video_path, max_frames)
+        if frames:
+            print(f"  📸 Extracted {len(frames)} frame(s) via OpenCV")
+            return frames
+    except ImportError:
+        pass
+    except Exception as e:
+        print(f"  ⚠️  OpenCV failed: {e}")
+
+    try:
+        frames = _frames_ffmpeg(video_path, max_frames)
+        if frames:
+            print(f"  📸 Extracted {len(frames)} frame(s) via ffmpeg")
+            return frames
+    except Exception as e:
+        print(f"  ⚠️  ffmpeg failed: {e}")
+
+    return []
+
+
+def image_to_b64(image_path: str) -> str:
+    """Read an image file and return a base64 JPEG string."""
+    try:
+        from PIL import Image
+        import io
+        img = Image.open(image_path).convert("RGB")
+        buf = io.BytesIO()
+        img.save(buf, format="JPEG", quality=80)
+        return base64.b64encode(buf.getvalue()).decode()
+    except ImportError:
+        with open(image_path, "rb") as f:
+            return base64.b64encode(f.read()).decode()
+
+
+# ─────────────────────────────────────────────────────────────────────────────
+#  Analyze
+# ─────────────────────────────────────────────────────────────────────────────
+
+VISION_PROMPT = """Analyze this image/video frame from an Instagram post. Provide:
+1. Main subject and what is happening
+2. Any visible text, captions, or overlays
+3. Setting / location clues
+4. Products, brands, or notable items visible
+5. Overall content theme
+
+Be concise and factual."""
+
+
+def analyze(file_path: str) -> None:
+    """Analyze a video or image file and print structured results."""
+
+    print("=" * 70)
+    print("🎬 VIDEO & IMAGE ANALYZER")
+    print("=" * 70)
+    print()
+
+    path = Path(file_path)
+    if not path.exists():
+        print(f"❌ File not found: {file_path}")
+        return
+    if path.is_dir():
+        print("❌ Provide a file, not a folder")
+        return
+
+    suffix = path.suffix.lower()
+    if suffix not in (".mp4", ".avi", ".mov", ".mkv", ".jpg", ".jpeg", ".png", ".gif"):
+        print(f"❌ Unsupported file type: {suffix}")
+        return
+
+    print(f"📹 File: {path.name}")
+    print()
+
+    # Build image list
+    images_b64: List[str] = []
+
+    if suffix in (".mp4", ".avi", ".mov", ".mkv"):
+        print("🎥 Extracting video frames...")
+        images_b64 = extract_frames(str(path), max_frames=4)
+        if not images_b64:
+            print("❌ Could not extract any frames from video")
+            return
+    else:
+        print("🖼️  Loading image...")
+        try:
+            images_b64 = [image_to_b64(str(path))]
+        except Exception as e:
+            print(f"❌ Could not load image: {e}")
+            return
+
+    # Send to vision model via router
+    print()
+    print("🤖 Analyzing with vision AI...")
+    print()
+
+    router = get_router()
+    try:
+        result = router.analyze_images(VISION_PROMPT, images_b64)
+    except RuntimeError as e:
+        print(f"❌ Vision analysis failed: {e}")
+        return
+
+    print("=" * 70)
+    print("📊 VISUAL ANALYSIS RESULTS")
+    print("=" * 70)
+    print()
+    print("📝 ANALYSIS:")
+    print("-" * 70)
+    print(result)
+    print("-" * 70)
+    print()
+    print("=" * 70)
+    print("✅ Done!")
+    print("=" * 70)
+
+
+def main():
+    if len(sys.argv) > 1:
+        file_path = sys.argv[1]
+    else:
+        print("=" * 70)
+        print("🎬 VIDEO & IMAGE ANALYZER")
+        print("=" * 70)
+        print()
+        file_path = input("📂 File path: ").strip()
+
+    file_path = file_path.strip('"').strip("'").strip()
+
+    if file_path:
+        analyze(file_path)
+    else:
+        print("❌ No path provided!")
+
+
+if __name__ == "__main__":
+    main()