A command line tool to cluster html pages based on structural and style similarity. This tool was based on Page Compare.
The quick way:
pip install html-cluster
How it works
- Download HTML form a list of files.
html-cluster download-html urls.txt
- Create a similarity score file.
html-cluster make-score-similarity-file --structural-weight=0.3
- Create the graph dot file
html-cluster make-graph similarity_score.json > graph.dot
- Render the image
neato -O -Tpng graph.dot
We use splash to take screenshots. You can use the following bash script to run it using docker
#!/usr/bin/env bash echo 'Splash is running on http://127.0.0.1:8050' docker run -p 5023:5023 -p 8050:8050 -p 8051:8051 scrapinghub/splash