Difference between revisions of "Omorfi"

From Apertium
Jump to navigation Jump to search
 
(9 intermediate revisions by 2 users not shown)
Line 1: Line 1:
{{TOCD}}
{{TOCD}}
'''OMorFi''' (Open Morphology of Finnish) is a computational morphology of Finnish written using [[HFST]].
'''Omorfi''' (Open Morphology of Finnish) is a computational morphology of Finnish written using [[HFST]].


==Requirements==
==Requirements==
Line 11: Line 11:


<pre>
<pre>
$ svn co http://svn.gna.org/svn/omorfi/trunk omorfi
$ git clone https://github.com/flammie/omorfi
$ cd omorfi/
$ cd omorfi/
$ ./autogen.sh
$ autoreconf -i
$ ./configure --prefix=/home/fran/local
$ ./configure
$ cd src/
</pre>
</pre>

In case autogen.sh does not work, do report a bug (autoreconf -i should work just as well in the meantime).


==Compilation==
==Compilation==

You need at least 1.5Gb RAM to compile Omorfi, or be willing to let your machine sit around trashing for some hours.


<pre>
<pre>
Line 24: Line 27:
</pre>
</pre>


This will compile everything. If your machine has less than 2Gb RAM you might want to just compile the analyser:
This will compile everything.


To prepare source code for new apertium language pair, use src/scripts/omor2apertium.sh... or just copy one from an existing pair, such as apertium-fin-eng.
<pre>
$ make mor-omorfi.hwfst
</pre>

This could take 10--30 minutes.


==Usage==
==Usage==


After compiling, you can test it with the <code>hfst-lookup</code> program.
After compiling, you can test it with the <code>hfst-lookup</code> program.

<pre>

$ echo "kaikki ihmiset syntyvät vapaina ja tasavertaisina arvoltaan ja oikeuksiltaan ." |\
sed 's/ /\n/g' | hfst-lookup src/mor-omorfi.hwfst

kaikki [##]kaikki[POS=PRONOUN][NUM=SG][CASE=NOM][##]

ihmiset [##]ihminen[POS=NOUN][KTN=38][NUM=PL][CASE=NOM,ACC][##]

syntyvät [##]syntyä[POS=VERB][KTN=52][KAV=J][GEN=ACT][MOOD=INDV][TENSE=PRES][PRS=PL3][##]
syntyvät [##]syntyä[POS=VERB][KTN=52][KAV=J][GEN=ACT][PCP=VA][CMP=POS][NUM=PL][CASE=NOM,ACC][##]

vapaina [##]vapaa[POS=ADJECTIVE][KTN=17][CMP=POS][NUM=PL][CASE=ESS][##]

ja [##]ja[POS=PARTICLE][##]
ja [##]ja[POS=CONJUNCTION][##]

tasavertaisina [##]tasavertainen[POS=ADJECTIVE][KTN=38][CMP=POS][NUM=PL][CASE=ESS][##]
tasavertaisina [##]tasa[POS=NOUN][KTN=9][NUM=SG][CASE=NOM][#][?]vertainen[POS=ADJECTIVE][KTN=38][CMP=POS]
[NUM=PL][CASE=ESS][##]

arvoltaan [##]arvo[POS=NOUN][KTN=1][NUM=SG][CASE=ABL][POSS=SG3,PL3][##]

ja [##]ja[POS=PARTICLE][##]
ja [##]ja[POS=CONJUNCTION][##]

oikeuksiltaan [##]oikeus[POS=NOUN][KTN=40][NUM=PL][CASE=ABL][POSS=SG3,PL3][##]

. [##].[POS=PUNCTUATION][##]

</pre>


==See also==
==See also==
Line 74: Line 41:


==External links==
==External links==

* [https://kitwiki.csc.fi/twiki/bin/view/KitWiki/OMorFiSFSTVersion#Installation OMorFi: Installation]
* [https://gna.org/projects/omorfi/ Gna!: Omorfi]
* [http://code.google.com/p/omorfi Omorfi project site at google code]
* [http://langtech.jrc.it/FSMNLP2008/m/Koskenniemi_invited_talk.pdf Overview of the HFST project (pdf)], esp. in relation to other FST technology
* [http://langtech.jrc.it/FSMNLP2008/m/Koskenniemi_invited_talk.pdf Overview of the HFST project (pdf)], esp. in relation to other FST technology



Latest revision as of 14:53, 2 June 2016

Omorfi (Open Morphology of Finnish) is a computational morphology of Finnish written using HFST.

Requirements[edit]

You will need HFST installed, you can follow the instructions on the HFST page.

Download[edit]

The following commands will download and prepare the build for OMorFi.

$ git clone https://github.com/flammie/omorfi
$ cd omorfi/
$ ./autogen.sh
$ ./configure

In case autogen.sh does not work, do report a bug (autoreconf -i should work just as well in the meantime).

Compilation[edit]

You need at least 1.5Gb RAM to compile Omorfi, or be willing to let your machine sit around trashing for some hours.

$ make

This will compile everything.

To prepare source code for new apertium language pair, use src/scripts/omor2apertium.sh... or just copy one from an existing pair, such as apertium-fin-eng.

Usage[edit]

After compiling, you can test it with the hfst-lookup program.

See also[edit]

External links[edit]