hlotze/lichess_puzzles_to_pdf

Chess Puzzles to PDF booklet

Chess Puzzles to PDF booklet¹

What this site provides

• a collection of PDF files (A4), evtl. for your personal printout as a booklet¹,
• some scripts that has been used for its generation, and
• this readme with a description about the PDF compilation.

My intention

... was to support myself learning chess.

Starting chess learning at lichess.org which has a good entry point for online-training, I like to have a couple (e.g. ten) of these puzzles as a printout, to take them with me for later offline-walkthrough.

After some googling I found the lichess puzzle database giving us access to lots of the puzzle as a free download of the CSV-formated data.

Steps

... to prepare the PDF-collection

1. Install a MySQL or MariaDB database, for later accessing the puzzles' data, etvl. plus an apache Webserver, a PHP and the phpMyAdmin - alternatively you may use XAMPP, which has that environment in one package.
2. Download the puzzle's CSV-file from lichess puzzle database
3. (optionally) split the large file into multiple smaller files with split command, e.g. split -l 1000 <input-filename>
4. Import the CSV-file into your database
5. add some columns e.g. for tuning the access to a group of puzzles by its theme. The CSV-file has a column that lists for each puzzle the themes belonging to that puzzle. There are less than 64 themes, so I used the set type and made a view for each theme with the same columns
6. use the Python script from this repository generate_theme_puzzles.py to generate for each theme (that are the views in your database) a dedicated subdirectory with ten puzzles' HTML-files
   • The Python script uses the python-chess to parse the FEN² and UCI moves³ to generate the SVG⁴ chess board diagrams.
7. use google-chrome --headless to generate for each HTML-file a PDF-file
   • e.g.
     • for one file: google-chrome --headless --disable-gpu --print-to-pdf-no-header --print-to-pdf=underPromotion_00001.pdf underPromotion_00001.html
     • for multiple files: ls *.html | awk '{html = $1; gsub(/html/, "pdf", $1); pdf = $0; print "google-chrome --headless --disable-gpu --print-to-pdf-no-header --print-to-pdf=",pdf, html}' | sed -e 's/= /=/g' | while read -r line; do $line; done
     • wkhtmltopdf and python-pdfkit corrupts my A4 layout; although HTML-files has the CSS @page { size: A4 }
8. use gs ghostscript for merging / concatination ten puzzle's PDF-files to one puzzle's PDF-file
   • e.g.
     • for all PDF-files of one puzzle: gs -dNOPAUSE -sDEVICE=pdfwrite -sOUTPUTFILE=../zugzwang.pdf -dBATCH $(ls -a *.pdf)

Open item

• alternatively to python-chess SVG drawing the chessboards, use free chess fonts (chess TTF examples here). Just B/W diagrams may give more contrast; anyway seems that Latex will be best choise for this (see that paper).
• documentation of the used database structure and used import and alter scripts
• add some manual comments or automatically generated comments to each puzzle's move; e.g. by means of python-chess accessing the Stockfish engine
• scripts for tasks' automation (with 1.) HTML-->PDF, 2.) merging PDFs, etc.)
• (evtl. for another project) similar script(s) for PGN⁵ files or a script that takes the FEN² plus the PGN⁵ of a match generating such a booklet¹.

Contact

@hlotze

Footnotes

1. booklet - one A4 sheet folded takes four pages ↩ ↩² ↩³

2. FEN - see Wikipedia Forsyth–Edwards Notation ↩ ↩²

3. UCI - see Wikipedia Universal Chess Interface ↩

4. SVG - see Wikipedia Scalable Vector Graphics ↩

5. PGN - see Wikipedia Portable Game Notation ↩ ↩²