HFST-REGEXP2FST(1)

📦
version 3.17.1 · 2 versions
  • 3.17.1 · debian bookworm, debian trixie, nixos 26.05, opensuse-leap 16.0, ubuntu jammy, ubuntu noble, ubuntu questing, ubuntu resolute, ubuntu stonking (latest)
  • 3.17.3 · nixpkgs unstable
, in debian bookworm, debian trixie, nixos 26.05, opensuse-leap 16.0, ubuntu jammy, ubuntu noble, ubuntu questing, ubuntu resolute, ubuntu stonking

🦋 hfst-regexp2fst — HFST manual, section 1 🦋

okay so. bestie. sit down, grab a boba, because we're about to talk about the absolute moment that is hfst-regexp2fst. it's giving compiler. it's giving transducers. it's giving everything. 💅

✨ NAME ✨

hfst-regexp2fst — compile (weighted) regular expressions into transducer(s).

(the man page literally opens the description with a stray =, like =Compile, and ngl we are keeping that exactly as-is because it's iconic. no notes. 🫶)

🎀 SYNOPSIS 🎀

hfst-regexp2fst [OPTIONS...] [INFILE]

like, that's it. that's the whole vibe. options, then optionally an input file. 😌

💖 DESCRIPTION 💖

So what does this bad boy actually do? It compiles (weighted) regular expressions into transducer(s). plural, bestie, TRANSDUCER(S). You hand it regexps, it hands you transducers. that's the whole relationship. 💞

Common options, coming right up:

  • -h, --help — print the help message, i.e. the "what even is this" button
  • -V, --version — print version info, for when you need receipts
  • -v, --verbose — print verbosely while processing, so it narrates its own life
  • -q, --quiet — only print fatal erros and requested output (yes, the typo is canon, we do not fix the canon)
  • -s, --silent — alias of --quiet, same energy, different name ✨

📥 Input/Output options 📤

  • -i, --input=INFILE — read input transducer from INFILE
  • -o, --output=OUTFILE — write output transducer to OUTFILE

🍓 String and format options 🍓

  • -f, --format=FMT — write result in FMT format
  • -j, --disjunct — disjunct all regexps instead of transforming each regexp into a separate transducer (one big happy family instead of a bunch of little ones 🫂)
  • -l, --line — input is line separated (this is the default, so if you say nothing, this is what you get)
  • -S, --semicolon — input is semicolon separated
  • -e, --epsilon=EPS — map EPS as zero, i.e. epsilon
  • -x, --xerox-composition=VALUE — whether flag diacritics are treated as ordinary symbols in composition (default is false)
  • -X, --xfst=VARIABLE — toggle xfst compatibility option VARIABLE

🌸 Harmonization and optimization options 🌸

  • -H, --do-not-harmonize — do not expand '?' symbols
  • -F, --harmonize-flags — harmonize flag diacritics
  • -E, --encode-weights — encode weights when minimizing (default is false)
  • -M, --do-not-minimize — determinize result instead of minimizing it

okay and here's the fine print, but make it fashion: 💅

If OUTFILE or INFILE is missing or -, standard streams will be used. standard streams, bestie, she's flexible like that.

FMT must be one of the following: {foma, sfst, openfst-tropical, openfst-log}.

If EPS is not defined, the default representation of 0 is used.

*VALUE*s recognized are {true,ON,yes} and {false,OFF,no}.

Xfst variables are {flag-is-epsilon (default OFF)}.

🔥 EXAMPLES 🔥

this is the part where it gets real. actual commands, actual vibes:

echo " {cat}:{dog} " | hfst-regexp2fst

→ create transducer {cat}:{dog} 🐱➡️🐶

echo " {cat}:{dog}::3 " | hfst-regexp2fst

→ same but with weight 3

echo " c:d a:o::3 t:g " | hfst-regexp2fst

→ same but with weight 3 in the middle

echo " cat ; dog ; "3" " | hfst-regexp2fst -S

→ create transducers "cat" and "dog" and "3"

see how the last one uses -S? semicolon separated. the semicolons are the moment. periodt. 💅✨

😭 REPORTING BUGS 😭

found a bug? first of all, relatable. second of all, report bugs to <hfst-bugs@helsinki.fi> or directly to our bug tracker at: <https://github.com/hfst/hfst/issues>

Copyright © 2017 University of Helsinki, License GPLv3: GNU GPL version 3 <http://gnu.org/licenses/gpl.html>

This is free software: you are free to change and redistribute it.

There is NO WARRANTY, to the extent permitted by law. (no cap. literally none. zero. the warranty does not exist. 🙈)

and that's the whole page, bestie. this was brought to you by HFST, dated August 2018 — a vintage, a classic, a legend. go compile something gorgeous. 🦋💖🎀

📦 where this page comes from