site-content-catalog（网站内容目录）

Site Content Catalog

Crawl a website's sitemap and blog to build a complete content inventory — every page cataloged with URL, title, date, content type, and topic cluster. Groups content by category, identifies publishing patterns, and optionally deep-analyzes top pages.

Quick Start

# Basic content inventory
python3 scripts/catalog_content.py --domain "example.com"

# With deep analysis of top 20 pages
python3 scripts/catalog_content.py --domain "example.com" --deep-analyze 20

# Output to specific file
python3 scripts/catalog_content.py --domain "example.com" --output content-inventory.json

Inputs

Cost

• Sitemap/RSS crawling: Free (direct HTTP requests)
• Apify sitemap extractor (fallback): ~$0.50 per site
• Deep analysis: Free (WebFetch on individual pages)

Process

Phase 1: Discover All Pages

The script attempts multiple methods to find all pages on a site, in order:

A) Sitemap.xml

1. Fetch https://[domain]/sitemap.xml
2. If it's a sitemap index, recursively fetch all child sitemaps
3. Common alternate locations: /sitemap_index.xml, /sitemap-index.xml, /wp-sitemap.xml
4. Check robots.txt for Sitemap: directives

B) RSS/Atom Feeds

1. Check /feed, /rss, /atom.xml, /blog/feed, etc.
2. Extract posts with titles, dates, and URLs
3. RSS typically only surfaces recent content (last 10-50 posts)

C) Blog Index Crawl

1. Fetch /blog, /resources, /insights, /news, /articles
2. Extract links from the page
3. Follow pagination if present (/blog/page/2, ?page=2, etc.)

D) Site: Search (fallback)

1. WebSearch: site:[domain] to estimate total indexed pages
2. WebSearch: site:[domain]/blog to find blog content
3. WebSearch: site:[domain] intitle: to discover page title patterns

E) Apify Sitemap Extractor (fallback for JS-heavy sites)

• Actor: onescales/sitemap-url-extractor
• Use when sitemap.xml is missing and the site is JS-rendered

Phase 2: Classify Each Page

For each discovered URL, classify by:

Content Type

Classify based on URL patterns and page titles:

Topic Cluster

Group by extracting topic signals from URL slugs and titles:

• Extract keywords from URL path segments
• Group similar keywords into clusters (e.g., "aws-cost", "cloud-spending", "finops" → "Cloud Cost Management")
• Use simple keyword co-occurrence for clustering

Phase 3: Analyze Publishing Patterns

From the dated content (primarily blog posts):

• Total content pieces by type
• Publishing frequency: Posts per month over last 12 months
• Trend: Increasing, decreasing, or stable output
• Recency: Date of most recent publish
• Author diversity: Unique authors (if extractable from RSS)

Phase 4: Deep Analysis (Optional)

If --deep-analyze N is specified, fetch the top N pages (prioritizing blog posts) and extract:

• Word count (approximate)
• Target keyword (inferred from title + H1 + URL)
• Funnel stage: TOFU (awareness), MOFU (consideration), BOFU (decision)
• Content depth: Shallow (<500 words), Medium (500-1500), Deep (1500+)
• Has images/video: Boolean
• Has CTA: Boolean (detected by common CTA patterns)
• Internal links count

Phase 5: Output

JSON Output (default)

{
  "domain": "example.com",
  "crawl_date": "2026-02-25",
  "total_pages": 347,
  "discovery_methods": ["sitemap.xml", "rss"],
  "pages": [
    {
      "url": "https://example.com/blog/reduce-aws-costs",
      "title": "How to Reduce Your AWS Bill by 40%",
      "date": "2025-11-15",
      "type": "blog-post",
      "topic_cluster": "Cloud Cost Optimization",
      "deep_analysis": {
        "word_count": 2100,
        "target_keyword": "reduce aws costs",
        "funnel_stage": "TOFU",
        "content_depth": "deep",
        "has_images": true,
        "has_cta": true
      }
    }
  ],
  "summary": {
    "by_type": {"blog-post": 89, "landing-page": 23, "case-study": 12, ...},
    "by_topic": {"Cloud Cost Optimization": 34, "FinOps": 18, ...},
    "publishing_cadence": {
      "posts_per_month_avg": 4.2,
      "trend": "increasing",
      "most_recent": "2026-02-20"
    }
  }
}

Markdown Summary (also generated)

# Content Inventory: example.com
**Crawled:** 2026-02-25 | **Total pages:** 347

## Content by Type
| Type | Count | % |
|------|-------|---|
| Blog Posts | 89 | 25.6% |
| Landing Pages | 23 | 6.6% |
| ...

## Content by Topic Cluster
| Topic | Posts | Most Recent |
|-------|-------|-------------|
| Cloud Cost Optimization | 34 | 2026-02-20 |
| ...

## Publishing Cadence
- Average: 4.2 posts/month
- Trend: Increasing (3.1 → 5.4 over last 6 months)
- Most recent: 2026-02-20

## Full Catalog
| # | Date | Type | Topic | Title | URL |
|---|------|------|-------|-------|-----|
| 1 | 2026-02-20 | blog-post | Cloud Cost | How to Reduce... | https://... |

Tips

• Sitemap.xml is the best source. Most well-maintained sites have one. If missing, it's itself an SEO signal (negative).
• RSS only shows recent content. If you need the full catalog, sitemap is essential. RSS is supplementary.
• Deep analysis is optional but valuable. Use it when feeding into brand-voice-extractor or when you need funnel stage mapping.
• JS-rendered sites may need the Apify fallback. Signs: sitemap.xml returns HTML, or blog page returns mostly JavaScript.
• Combine with seo-domain-analyzer to overlay traffic data on the content inventory — see which content actually performs.

Dependencies

• Python 3.8+
• requests library (pip install requests)
• APIFY_API_TOKEN env var (only for Apify fallback mode)