Introduction provides a simple interface for python applications to interact with installed omorfi lexicons. Consider it as a standard python entry-point to omorfi. N.B. python module provides slightly different analyses and more functionality than the raw automaton.

Current interface

There will be apidoc generated.

  • module omorfi:
    • class Omorfi:
      • init(self, verbosity)
      • load_filename(self, path)
      • load_from_dir(self, path=None)
      • tokenise(self, line)
      • analyse(self, token)
      • analysers: dict of analysis automata

load_from_dir(self, path=None)

Loads omorfi automata into memory. If path is given it should point to the automaton, otherwise standard installation paths are tried, currently standard linux install paths are tried in order:

  1. /usr/local/share/hfst/fi/morphology.omor.hfst
  2. /usr/share/hfst/fi/morphology.omor.hfst
  3. /usr/local/share/omorfi/morphology.omor.hfst
  4. /usr/share/omorfi/morphology.hfst
  5. $HOME/.hfst/fi/morphology.omor.hfst
  6. $HOME/.omorfi/morphology.omor.hfst

Last two paths require an accessible HOME env.var. Returned value is a handle to the analyser, currently a swig object pointing to HFST automaton.

analyse(self, token)

Look up token from morphological analyser(s) loaded. If self.can_titlecase does not evaluate to False, and the token cannot be analysed, the analysis will be retried with first letter’s case variants. If self.can_uppercase evaluates to not False and the token cannot be analysed, the words casing will be swapped and retried. The analyses with case mangling will have an additional element to them identifying the casing.


If invoked from command-line, can be used as simple lookup tool. Invoke with --help for more info.