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

1. Download HTML form a list of files.

html-cluster download-html urls.txt

2. Create a similarity score file.

html-cluster make-score-similarity-file --structural-weight=0.3

3. Create the graph dot file

html-cluster make-graph similarity_score.json > graph.dot

4. 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 https://127.0.0.1:8050'
docker run -p 5023:5023 -p 8050:8050 -p 8051:8051 scrapinghub/splash