fst-infl(1)

📦 version 1.4.7b, in debian bookworm, ubuntu jammy, ubuntu noble

✨ fst-infl2-daemon ✨

okay bestie, you're looking at the fst-infl(1) manual page — and specifically at the entry for fst-infl2-daemon, which is a whole mood. it's a morphological analyser that lives on a socket. yes, really. she's networked. she's serving. 💅

🌸 NAME 🌸

fst-infl2-daemon — morphological analysers 🧚

🦋 SYNOPSIS 🦋

fst-infl2-daemon [ options ] <socket-no> file

💖 OPTIONS 💖

  • -t file — read an alternative transducer from file and use it whenever the main transducer just... can't find an analysis (we've all been there 🥲). and the way that you can stack this option?? iterate it and you get a whole cascade of transducers being tried in order until something hits. icon behaviour, fr fr.
  • -b — print surface and analysis symbols. (this one's fst-infl2 only, bestie)
  • -n — print multi-character symbols without the enclosing angle brackets. (and this one's fst-infl only, so don't get it twisted)
  • -d — the analyses get symbolically disambiguated: you only get back the analyses with a minimal number of morphemes. but she has a requirement: morpheme boundaries have to be marked with the tag <X>. if there's no <X> tag anywhere in the analysis string, then the program (basically) counts up the multi-character symbols made entirely of upper-case characters and uses that count to disambiguate instead. that little heuristic? developed for the German SMOR morphology, no cap. (heads up: this option only exists with fst-infl2 and fst-infl3.)
  • -e n — if no regular analysis turns up, go do robust matching and print analyses with up to n edit errors. the set of edit operations currently includes replacement, insertion and deletion, and each operation currently has a fixed error weight of 1. (this is fst-infl2 only 🫶)
  • -% f — disambiguates the analyses statistically and prints the most likely analyses that hold at least f % of the total probability mass of the analyses. the transducer weights get read from a file obtained by appending .prob to the name of the transducer file. those weight files are created with fst-train. (also fst-infl2 only ✨)
  • -p — print the probability of each analysis. (yep, fst-infl2 only)
  • -c — use this one if the transducer was compiled on a computer with a different endianness. like, say you've got a transducer compiled on a Sparc computer and you wanna use it on a Pentium — that's exactly when you need this option. (fst-infl2 only)
  • -q — suppress status messages. silence is golden, periodt 🤫
  • -h — print usage information. your girl loves to explain herself.

🌈 DESCRIPTION 🌈

fst-infl2-daemon is basically fst-infl2's cousin, except she reads and writes from/to a socket. that's the whole vibe. socket girl summer. 🔌

💀 BUGS 💀

no bugs are known so far. and honestly? we love a flawless queen. 👑

🎀 SEE ALSO 🎀

fst-infl2, fst-compiler, fst-mor

👑 AUTHOR 👑

Helmut Schmid, Institute for Computational Linguistics, University of Stuttgart, Email: schmid@ims.uni-stuttgart.de, This software is available under the GNU Public License.

k that's the whole page bestie 💜 this was the fst-infl2-daemon entry from the fst-infl(1) manual, dated December 2012 — go forth and analyse those morphemes, mwah 🫶💖

📦 where this page comes from