dohliam / html-table2text

HTML Table to Text - Extract and convert HTML tables to plain text formats

HTML Table to Text - Extract and convert HTML tables to plain text formats

This is a simple (optionally interactive) script that can extract any or all tables from a given HTML file or URL. The data can be output to CSV (comma-separated values), TSV (tab-separated values), Markdown, Asciidoc, or raw HTML.

Requirements

This script relies on Nokogiri to parse HTML. You can install it with:

gem install nokogiri

Markdown conversion uses reverse_markdown which can be installed the same way:

gem install reverse_markdown

Asciidoc conversion uses the reverse_adoc gem:

gem install reverse_adoc

Usage

To extract tables from an arbitrary URL, just run the webtable_to_text script with the -u option followed by the URL:

./webtable_to_text.rb -u [URL]

For example:

./webtable_to_text.rb -u "https://en.wikipedia.org/wiki/Gabon"

This will print out all the tables found on the specified page.

To output a specific table only, use the n option, followed by the number of the table:

./webtable_to_text.rb -u "https://en.wikipedia.org/wiki/Gabon" -n 3

The script also works with local files, using the f option, e.g.:

./webtable_to_text.rb -f some_file.html

Interactive mode

To use interactive mode, add the -i option to the command and specify a URL or file as normal. For example:

./webtable_to_text.rb -u "https://en.wikipedia.org/wiki/Gabon" -i

This will print a message with the total number of tables found in the document. If you enter a number at the prompt, it will print the corresponding table. Otherwise, pressing ENTER or RETURN will print all tables found.

For example, pressing 3 will print something like the following:

Population in Gabon
Year	Million 
1950	0.5 
2000	1.2 
2016	2

To run tests, just enter the following command:

ruby tests.rb

Options

The following options are available:

• -A, --all: Print all tables found on the specified page
• -a, --asciidoc: Output in asciidoc/asciidoctor format
• -c, --csv: Output in CSV / comma separated values format
• -f, --file FILE: Specify HTML input file as source for extracting tables
• -h, --help: Print help text
• -i, --interactive: Interactive mode
• -m, --markdown: Output in markdown format
• -n, --number NUM: Print specific table number only; separate multiple numbers with commas
• -o, --output FILE: Specify output file (default: output to STDOUT)
• -r, --raw: Output raw table HTML
• -t, --tsv: Output in TSV / tab separated values format (default)
• -u, --url URL: Specify URL as source for extracting tables

To do

• add options and non-interactive mode
• output to raw HTML
• output to Markdown
• output to AsciiDoc

Credits

• Table extraction to CSV based on this Stack Overflow answer by user audiodude.

License

MIT.