A simple sitemap builder

The sitemap builder traverses links from a website and constrains itself to the given domain name. The final result will be a simple sitemap deduced from the links visited. The crawler will accept & process only URLs with http or https schemes.

Installation and usage

To run the following command to install the tool:

pip install -U sitemapbuilder

To run the sitemap builder:

sitemapbuilder -u 'https://monzo.com' -o test_monzo.dot

Some websites have strong protection and the tool will not work for them:

sitemapbuilder -u 'https://bloomberg.com' -o test_bloomberg.dot

Highlights

1. Generate Graphviz .dot file showing directed links between pages. One can generate PNG/PDF and other image/document formats.

2. Have configurable decay (maximum depth) to avoid abuse.

3. Visit web link within the same hostname by default.

4. Use 5 threads by default and times out after 10 seconds.

5. Timeout after 5 seconds when fetching a URL.

6. Handle timeout exceptions when querying a website.

7. Send a HTTP HEAD request and verify that Content-Type is text/html and charset is either UTF-8 or US-ASCII.

8. Have a map of visited URLs to avoid revisiting them.

9. Follow HTTP redirects.

Upcoming features

• Configure the number of threads and timeout via cmd args.

• Allow web links from all subdomains.

• Allow web links from a list of domains.

• Allow web links matching a pattern.

• Add an option for hierarchical sitemap instead of directed graph.

• Use PriorityQueue instead of Queue to process links with higher decay first.

• Fine-graned info, warn and error logging.

• Pass seed links from a file.

• Save to and resume from a DB/persistent data source.

• Faster concurrency and better performance with asyncio.