Skip to content

wader/ansisvg

Repository files navigation

ansisvg

Convert ANSI to SVG.

Pipe output from a program thru ansisvg and it will output a SVG file on stdout.

Can be used to produce nice looking example output for presentations, markdown files etc. Note that it does not support programs that do cursor movements like ncurses programs etc.

./colortest | ansisvg > colortest.svg

Produces colortest.svg:

ansisvg output for colortest

$ ansisvg -h
ansisvg - Convert ANSI to SVG
Usage: ansisvg [FLAGS]

Example usage:
  program | ansisvg > file.svg

--charboxsize WxH   Character box size (use pixel units instead of font units)
--colorscheme NAME  Color scheme
--fontfile PATH     Font file to use and embed
--fontname NAME     Font name
--fontref URL       External font URL to use
--fontsize NUMBER   Font size
--grid              Grid mode (sets position for each character)
--help, -h          Show help
--listcolorschemes  List color schemes
--marginsize WxH    Margin size (in either pixel or font units)
--transparent       Transparent background
--version, -v       Show version
--width, -w NUMBER  Terminal width (auto if not set)

Color themes are the ones from https://github.com/mbadolato/iTerm2-Color-Schemes

Install

Pre-built binaries for Linux, macOS and Windows can be downloaded from releases.

macOS

For macOS you might have to allow to run the binary in security preferences. Alternatively run the below command:

xattr -d com.apple.quarantine ansisvg && spctl --add ansisvg

Build

To build you will need at least go 1.18 or later.

Install latest master and copy it to /usr/local/bin:

go install github.com/wader/ansisvg@master
cp $(go env GOPATH)/bin/ansisvg /usr/local/bin

Build from cloned repo:

go build -o ansisvg .

Fonts

ansisvg can either use system-installed fonts (-fontname), link to a webfont on a HTTP server (-fontref) or embed a webfont from the local filesystem (-fontfile).

Compatibility issues

  • Embedded and/or linked fonts might not be supported by some SVG viewers. At time of writing this is not supported by Inkscape.

  • For SVGs that are intended to be included in websites via <img>, the only way to make a custom font work is embedding it in the SVG.

Variations of custom fonts (regular/bold/italic)

  • System wide fonts (-fontname) get correctly rendered with variations, but when using external fonts with -fontref or -fontfile the SVG viewer knows only the regular variant and will try to render italic/bold text 'extrapolated' from it which may look different than the actual font variation. To use the actual bold/italic font variants, different woff2 files have to be used for the respective text styles which needs additional CSS code (currently not supported by ansisvg).

  • Bold style 'extrapolated' from the regular font may even break monospace alignment. Use -grid option to mitigate that.

Font-relative vs. pixel coordinates

By default, ansisvg uses font-relative ch/em coordinates. This should make SVG dimensions and line/character spacing consistent with font family/size. When SVG dimensions and/or text coordinates are off, it is possible to force explicit pixel units for coordinates by specifying -charboxsize in X/Y pixel units, e.g. 8x16.

Margin size

With --marginsize a margin can be defined, so there is a bit of empty space (or "border") around the image. Default is zero margin size, i.e. the terminal characters are touching the edge of the image. --marginsize is interpreted as X/Y in the currently selected units, i.e. ch/em by default, and px if --charboxsize is used.

Consolidated text vs. grid mode

By default, ansisvg consolidates text to <tspan> chunks, leaving the X positioning of characters to the SVG renderer. This usually works well for monospace fonts. However if not all glyphs involved are monospace (e.g. when exotic characters are used, making the SVG renderer fall back to a different font for those characters) then the alignment will be off; this can be worked around with -grid mode which will make ansisvg put each character to explicit positions, making the SVG bigger and less readable but ensuring proper positioning/alignment for all characters.

Tricks

ANSI to PDF or PNG

... | ansisvg | inkscape --pipe --export-type=pdf -o file.pdf
... | ansisvg | inkscape --pipe --export-type=png -o file.png

Use bat to produce source code highlighting

bat --color=always -p main.go | ansisvg

Use script to run with a pty

script -q /dev/null <command> | ansisvg

ffmpeg

TERM=a AV_LOG_FORCE_COLOR=1 ffmpeg ... 2>&1 | ansisvg

jq

jq -C | ansisvg

Make screenshots from a terminal

tmux

# <prefix>-H: Create a SVG screenshot of the current pane
bind H capture-pane -e \; run "tmux save-buffer - | $HOME/go/bin/ansisvg > $HOME/Pictures/tmux-$(date +%F_%T).svg"; delete-buffer

kitty

# F3: Create a SVG screenshot of the current selection
map f3 combine : copy_ansi_to_clipboard : launch sh -c 'kitty +kitten clipboard -g | $HOME/go/bin/ansisvg > $HOME/Pictures/kitty-$(date +%F_%T).svg'

Development and release build

Run all tests and write new test output:

go test ./... -update

Manual release build with version can be done with:

go build -ldflags "-X main.version=1.2.3" -o ansisvg .

Visual inspect test output in browser:

for i in cli/testdata/*.svg; do echo "$i<br><img src=\"$i\"/><br>" ; done  > all.html
open all.html

Using ffcat:

for i in cli/testdata/*.ansi; do echo $i ; cat $i | go run . | ffcat ; done

Thanks

  • Patrick Huesmann @patrislav1 for better ANSI support and lots SVG output improvements.

Licenses and thanks

Color themes from https://github.com/mbadolato/iTerm2-Color-Schemes, license https://github.com/mbadolato/iTerm2-Color-Schemes/blob/master/LICENSE

Uses colortest from https://github.com/pablopunk/colortest and terminal-colors from https://github.com/eikenb/terminal-colors.

UbuntuMonoNerdFontMono-Regular.woff2 from https://github.com/ryanoasis/nerd-fonts license https://github.com/ryanoasis/nerd-fonts/blob/master/LICENSE

TODO and ideas

  • Underline overlaps a bit, sometimes causing weird blending
  • Handle vertical tab and form feed (normalize into spaces?)
  • Handle overdrawing
  • More CSI, keep track of cursor?
  • PNG output (embed nice fonts?)