design-clone 2.1.0 → 2.3.0

package/src/ai/analyze-structure.py

@@ -1,375 +0,0 @@
-#!/usr/bin/env python3
-"""
-Analyze website screenshot structure using Gemini Vision API.
-
-Usage:
-  python analyze-structure.py --screenshot ./analysis/desktop.png --output ./analysis
-  python analyze-structure.py -s desktop.png -o ./out --html source.html --css source.css
-
-Options:
-  --screenshot   Path to desktop screenshot
-  --output       Output directory for structure.md
-  --html         Path to source HTML file (optional, improves accuracy)
-  --css          Path to filtered CSS file (optional, improves accuracy)
-  --model        Gemini model (default: gemini-2.5-flash)
-  --verbose      Enable verbose output
-
-Output:
-  - structure.md: Markdown description of page structure
-
-When HTML/CSS provided, extracts EXACT values from source instead of estimating.
-"""
-
-import argparse
-import json
-import os
-import sys
-from pathlib import Path
-
-# Add src directory to path for local imports
-SCRIPT_DIR = Path(__file__).parent.resolve()
-SRC_DIR = SCRIPT_DIR.parent
-sys.path.insert(0, str(SRC_DIR))
-
-# Import local env resolver (portable)
-try:
-    from utils.env import resolve_env, load_env
-    load_env()  # Load .env files on startup
-except ImportError:
-    # Fallback: simple env getter
-    def resolve_env(key, default=None):
-        return os.environ.get(key, default)
-
-# Check for google-genai dependency
-try:
-    from google import genai
-    from google.genai import types
-except ImportError:
-    print(json.dumps({
-        "success": False,
-        "error": "google-genai not installed",
-        "hint": "Run: pip install google-genai"
-    }, indent=2))
-    sys.exit(1)
-
-# Import prompts from extracted module
-from prompts.structure_analysis import build_structure_prompt
-
-
-# Fallback structure when analysis fails
-DEFAULT_STRUCTURE = """# Page Structure Analysis
-
-## 1. Header Section
-- Logo: Left-aligned, text-based
-- Navigation: Horizontal, 4-5 items
-- CTA Button: Right-aligned, primary color
-- Mobile menu: Hamburger icon for small screens
-
-## 2. Hero Section
-- Layout: Centered
-- Headline: Large (36-48px), bold, dark text
-- Subheadline: Medium (18-20px), lighter text
-- Primary CTA: Prominent button, primary color
-- Background: Solid light color
-
-## 3. Content Sections
-### Features Section
-- Layout: 3-column grid
-- Items: 3 feature cards with icons
-- Components: Icon + heading + description
-
-### About/Info Section
-- Layout: 2-column (text + image)
-- Alternating left-right pattern
-
-## 4. Footer Section
-- Layout: 4-column grid
-- Content: Logo, nav links, contact, social icons
-- Copyright: Centered at bottom
-
-## 5. Global Patterns
-- Container max-width: 1200px
-- Section padding: 64px vertical
-- Card style: Subtle shadows, 8px border-radius
-- Color scheme: Light mode
-- Typography: Sans-serif (modern)
-
-## 6. Responsive Hints
-- Navigation collapses to hamburger on mobile
-- Grid sections stack vertically
-- Padding reduces on smaller screens
-
-## 7. BEM Class Suggestions
-- header, header__container, header__logo, header__nav, header__cta
-- hero, hero__container, hero__title, hero__subtitle, hero__cta
-- features, features__grid, feature-card, feature-card__icon, feature-card__title
-- footer, footer__container, footer__column, footer__links
-"""
-
-
-def get_api_key():
-    """Get Gemini API key from environment (supports GEMINI_API_KEY or GOOGLE_API_KEY)."""
-    return resolve_env('GEMINI_API_KEY') or resolve_env('GOOGLE_API_KEY')
-
-
-def load_dimensions(output_dir: str) -> dict:
-    """Load extracted dimensions summary if available.
-
-    Args:
-        output_dir: Directory containing dimensions-summary.json
-
-    Returns:
-        Dimensions dict or None if not found
-    """
-    summary_path = Path(output_dir) / "dimensions-summary.json"
-    if summary_path.exists():
-        try:
-            with open(summary_path, 'r', encoding='utf-8') as f:
-                return json.load(f)
-        except (json.JSONDecodeError, IOError) as e:
-            print(f"Warning: Failed to load dimensions: {e}", file=sys.stderr)
-    return None
-
-
-def load_dom_hierarchy(output_dir: str) -> dict:
-    """Load extracted DOM hierarchy if available.
-
-    Args:
-        output_dir: Directory containing dom-hierarchy.json
-
-    Returns:
-        Hierarchy dict or None if not found
-    """
-    hierarchy_path = Path(output_dir) / "dom-hierarchy.json"
-    if hierarchy_path.exists():
-        try:
-            with open(hierarchy_path, 'r', encoding='utf-8') as f:
-                return json.load(f)
-        except (json.JSONDecodeError, IOError) as e:
-            print(f"Warning: Failed to load DOM hierarchy: {e}", file=sys.stderr)
-    return None
-
-
-def load_content_counts(output_dir: str) -> tuple:
-    """Load content counts and summary if available.
-
-    Args:
-        output_dir: Directory containing content-counts.json and content-summary.md
-
-    Returns:
-        Tuple of (counts_dict, summary_text) or (None, None) if not found
-    """
-    counts_path = Path(output_dir) / "content-counts.json"
-    summary_path = Path(output_dir) / "content-summary.md"
-
-    counts = None
-    summary = None
-
-    if counts_path.exists():
-        try:
-            with open(counts_path, 'r', encoding='utf-8') as f:
-                counts = json.load(f)
-        except (json.JSONDecodeError, IOError) as e:
-            print(f"Warning: Failed to load content counts: {e}", file=sys.stderr)
-
-    if summary_path.exists():
-        try:
-            with open(summary_path, 'r', encoding='utf-8') as f:
-                summary = f.read()
-        except IOError as e:
-            print(f"Warning: Failed to load content summary: {e}", file=sys.stderr)
-
-    return counts, summary
-
-
-def analyze_structure(
-    screenshot_path: str,
-    output_dir: str = None,
-    html_path: str = None,
-    css_path: str = None,
-    model: str = "gemini-2.5-flash",
-    verbose: bool = False
-) -> str:
-    """Analyze screenshot structure using Gemini Vision.
-
-    Args:
-        screenshot_path: Path to desktop screenshot
-        output_dir: Output directory (also reads dimensions-summary.json if present)
-        html_path: Optional path to source HTML (improves accuracy)
-        css_path: Optional path to filtered CSS (improves accuracy)
-        model: Gemini model to use
-        verbose: Enable verbose output
-
-    Returns:
-        Markdown structure analysis
-    """
-
-    api_key = get_api_key()
-    if not api_key:
-        if verbose:
-            print("Warning: GEMINI_API_KEY not found, using default structure")
-        return DEFAULT_STRUCTURE
-
-    screenshot = Path(screenshot_path)
-    if not screenshot.exists():
-        if verbose:
-            print(f"Warning: Screenshot not found: {screenshot_path}")
-        return DEFAULT_STRUCTURE
-
-    # Load extracted dimensions if available (highest priority)
-    dimensions = None
-    if output_dir:
-        dimensions = load_dimensions(output_dir)
-        if dimensions and verbose:
-            print(f"Loaded extracted dimensions from {output_dir}/dimensions-summary.json")
-
-    # Load DOM hierarchy if available (enhances dimensions)
-    hierarchy = None
-    if output_dir:
-        hierarchy = load_dom_hierarchy(output_dir)
-        if hierarchy and verbose:
-            stats = hierarchy.get('stats', {})
-            print(f"Loaded DOM hierarchy: {stats.get('totalNodes', 0)} nodes, depth {stats.get('maxDepth', 0)}")
-
-    # Load content counts if available
-    content_counts, content_summary = None, None
-    if output_dir:
-        content_counts, content_summary = load_content_counts(output_dir)
-        if content_counts and verbose:
-            print(f"Loaded content counts: {content_counts.get('summary', {}).get('totalRepeatedItems', 0)} items")
-
-    # Load HTML/CSS if provided
-    html_content = None
-    css_content = None
-
-    if html_path and Path(html_path).exists():
-        with open(html_path, 'r', encoding='utf-8') as f:
-            html_content = f.read()
-        if verbose:
-            print(f"Loaded HTML: {len(html_content)} chars")
-
-    if css_path and Path(css_path).exists():
-        with open(css_path, 'r', encoding='utf-8') as f:
-            css_content = f.read()
-        if verbose:
-            print(f"Loaded CSS: {len(css_content)} chars")
-
-    # Build prompt with context (hierarchy+dimensions have highest priority)
-    prompt = build_structure_prompt(html_content, css_content, dimensions, content_summary, hierarchy)
-
-    if verbose:
-        if hierarchy and dimensions:
-            print("Using HIERARCHY prompt with DOM structure + EXACT dimensions")
-        elif dimensions:
-            print("Using ENHANCED prompt with EXACT extracted dimensions")
-        if content_summary:
-            print("Using prompt with EXACT content counts")
-        elif html_content and css_content:
-            print("Using prompt with HTML/CSS context")
-
-    try:
-        client = genai.Client(api_key=api_key)
-
-        # Load image
-        with open(screenshot, 'rb') as f:
-            img_bytes = f.read()
-
-        content = [
-            prompt,
-            types.Part.from_bytes(data=img_bytes, mime_type='image/png')
-        ]
-
-        if verbose:
-            print(f"Analyzing structure with {model}...")
-
-        response = client.models.generate_content(
-            model=model,
-            contents=content
-        )
-
-        if hasattr(response, 'text') and response.text:
-            if verbose:
-                print("Structure analysis complete")
-            return response.text
-        else:
-            if verbose:
-                print("Warning: Empty response, using default structure")
-            return DEFAULT_STRUCTURE
-
-    except Exception as e:
-        if verbose:
-            print(f"Error during analysis: {e}")
-        return DEFAULT_STRUCTURE
-
-
-def main():
-    parser = argparse.ArgumentParser(
-        description="Analyze website screenshot structure using Gemini Vision"
-    )
-    parser.add_argument(
-        '--screenshot', '-s',
-        required=True,
-        help='Path to desktop screenshot'
-    )
-    parser.add_argument(
-        '--output', '-o',
-        required=True,
-        help='Output directory for structure.md (also reads dimensions-summary.json if present)'
-    )
-    parser.add_argument(
-        '--html',
-        default=None,
-        help='Path to source HTML file (optional, improves accuracy)'
-    )
-    parser.add_argument(
-        '--css',
-        default=None,
-        help='Path to filtered CSS file (optional, improves accuracy)'
-    )
-    parser.add_argument(
-        '--model', '-m',
-        default='gemini-2.5-flash',
-        help='Gemini model to use (default: gemini-2.5-flash)'
-    )
-    parser.add_argument(
-        '--verbose', '-v',
-        action='store_true',
-        help='Enable verbose output'
-    )
-
-    args = parser.parse_args()
-
-    # Create output directory
-    output_path = Path(args.output)
-    output_path.mkdir(parents=True, exist_ok=True)
-
-    # Analyze structure (dimensions loaded automatically from output_dir)
-    structure = analyze_structure(
-        screenshot_path=args.screenshot,
-        output_dir=args.output,
-        html_path=args.html,
-        css_path=args.css,
-        model=args.model,
-        verbose=args.verbose
-    )
-
-    # Save structure.md
-    md_path = output_path / "structure.md"
-    with open(md_path, 'w') as f:
-        f.write(structure)
-
-    if args.verbose:
-        print(f"Saved: {md_path}")
-
-    # Output result as JSON
-    result = {
-        "success": True,
-        "structure_file": str(md_path),
-        "model": args.model
-    }
-
-    print(json.dumps(result, indent=2))
-
-
-if __name__ == '__main__':
-    main()