Difference between revisions of "Northern Sámi and Norwegian/smemorf"
Jump to navigation
Jump to search
| (50 intermediate revisions by 3 users not shown) | |||
| Line 1: | Line 1: | ||
apertium-sme-nob |
Documentation and TODO's for apertium-sme-nob's sme morphological analyser from Giellatekno. |
||
== |
==Description== |
||
* [[Ideas for Google Summer of Code/Morphology with HFST|HFST tokenisation]] |
|||
** Split up lexicon into standard and inconditional (punctuation) for hfst-proc |
|||
The sme morphological analyser is a trimmed and tag-changed version of the one in Giellatekno. |
|||
===Trimming=== |
|||
| ⚫ | |||
Trimming happens using the same HFST method as in the Turkic pairs etc. [[Automatically_trimming_a_monodix#Compounds_vs_trimming_in_HFST|Compounds are not handled correctly by this method]]. |
|||
===Tagset changes=== |
|||
* Proper casing support in the sme lexicon. (Mánát vs. mánát) |
|||
** In the xerox software, this is done by a separate fst m (->) M || .#. _ ; |
|||
The tagset in apertium-sme-nob is closer to Apertium's tagset, thus a bit different from the one in Giellatekno. |
|||
* We remove Err/Orth/Usage tags (+Use/Sub, etc) – |
|||
* We remove any derivational analyses that aren't yet handled by transfer/bidix |
|||
** https://victorio.uit.no/langtech/trunk/langs/sme/tools/mt/apertium/filters/remove-derivation-strings-modifications.nob.regex |
|||
** https://victorio.uit.no/langtech/trunk/langs/sme/src/filters/remove-illegal-derivation-strings.regex |
|||
** [[Northern Sámi and Norwegian/Derivations|More on derivations in sme-nob]] |
|||
* We reorder some tags |
|||
** https://victorio.uit.no/langtech/trunk/langs/sme/src/filters/reorder-subpos-tags.sme.regex |
|||
* We change the format of tags, so +N becomes <n>, +Der/PassL becomes <der_passl>, etc. |
|||
* We change certain tags |
|||
** https://victorio.uit.no/langtech/trunk/langs/sme/tools/mt/apertium/tagsets/apertium.nob.relabel |
|||
** https://victorio.uit.no/langtech/trunk/langs/sme/tools/mt/apertium/tagsets/apertium.postproc.relabel |
|||
** https://victorio.uit.no/langtech/trunk/langs/sme/tools/mt/apertium/tagsets/modify-tags.nob.regex |
|||
==TODO== |
|||
===Misc=== |
|||
* add entries from bidix that are missing from the analyser |
|||
* regex for URL's (don't want telefonkatalogen.no => *telefonkatalogen.nå) |
|||
| ⚫ | |||
* <code>8632: SUBSTITUTE:TV_IV (V TV) (V IV) FAUXV (0 ("lávet"));</code> -- this should be analysed as both, and disambiguated |
* <code>8632: SUBSTITUTE:TV_IV (V TV) (V IV) FAUXV (0 ("lávet"));</code> -- this should be analysed as both, and disambiguated |
||
===Typos=== |
|||
| ⚫ | |||
* a list of [http://paste.pocoo.org/raw/498998/ high-frequency typos] where the correction has an analysis |
|||
* Suddenly some nouns have the +Actor tag after +N... remove? |
|||
* geafivuohta gets one Der/vuohta-analysis that's got no A tag, and one with an A tag, fix: |
|||
** geafivuohta geafi+SgNomCmp+SgGenCmp+PlGenCmp+A+Der3+Der/vuohta+N+Sg+Nom |
|||
** geafivuohta geafi+SgNomCmp+SgGenCmp+PlGenCmp+Der3+Der/vuohta+N+Sg+Nom |
|||
** geafivuohta geafivuohta+SgGenCmp+DefPlGenCmp+N+Sg+Nom |
|||
** http://giellatekno.uit.no/bugzilla/show_bug.cgi?id=852 |
|||
| ⚫ | |||
* oktan seems to be used as a preposition, can we have that? |
|||
** Juo cuoŋománu 10. beaivvi ija vuostá mátkkoštii Ruvdnaprinseassa Märtha badjel ráji Ruŧŧii oktan Ruvdnaprinsabára golmmain mánáin. |
|||
** áhčči oktan mánáinis |
|||
** http://giellatekno.uit.no/bugzilla/show_bug.cgi?id=853 |
|||
==Compounding== |
|||
===ensure compounds are tagged as such=== |
|||
http://giellatekno.uit.no/bugzilla/show_bug.cgi?id=851 |
|||
* ''allaskuvla'', ''Riikaráđi'' and ''1100-logu'' get no compound tags in the compound analysis, the first can be fixed by adding +Cmpnd in LEXICON ATTR, the RReal border, similarly for the second in LEXICON ACCRA-NE, both RReal borders. |
|||
* -prográmma gives <code>^-prográmma/+Cmpnd+prográmma<N><Sg><Nom>/+Cmpnd+prográmma<N><Sg><Acc>/+Cmpnd+prográmma<N><Sg><Gen>$</code>. We currently remove the +Cmpnd+ in dev/xfst2apertium.hashtags.twol, but then we have no indication that the word started with a dash (-). In other language pairs, the - is output separately (either an inconditional lexical unit, like punctuation, or without any analysis, like unknown symbols). |
|||
** A combination of the two errors above: Ánde-máhka gives <code>^Ánde<N><Prop><Mal><Sg><Nom><-máhka><N><Sg><Nom>$</code>... we should have something like <code>^Ánde<N><Prop><Mal><Sg><Nom>+<-máhka><N><Sg><Nom>$</code> |
|||
===ensure compounding is only tried if there is no other solution=== |
|||
Most general solution: Use a weighted transducer, and give the compound border (ie the dynamic compounding border, the R lexicon) a non-zero weight. |
|||
===Dashes=== |
|||
However, hfst-proc does the same by just removing any analysis with a + in it if we have one without a +, so we're OK for now… |
|||
lexc handles dashes by adding them literally (like a lemma), doing that in with a bidix pardef would be very messy (also, doesn't it give issues with lemma-matching in CG?). |
|||
==Multiwords== |
===Multiwords=== |
||
Add simple multiwords and fixed expressions to the analyser. |
Add simple multiwords and fixed expressions to the analyser. |
||
* lea go => <code>^leat<V><IV><Ind><Prs><Sg3><Qst></code> (just like "leago") |
|||
* dasa lassin => i tillegg (til det) |
* dasa lassin => i tillegg (til det) |
||
* dán áigge => for tiden |
* dán áigge => for tiden |
||
| Line 63: | Line 54: | ||
* oaidnaleapmái => 'see you' |
* oaidnaleapmái => 'see you' |
||
* ovdamearkka => for eksempel |
* ovdamearkka => for eksempel |
||
* Mo manná? => Hvordan går det? |
|||
* |
|||
* ja nu ain => og så videre |
|||
(Some of these MWE's might be very Apertium-specific, but in that case we just keep our own file and append the entries with update-lexc.sh.) |
(Some of these MWE's might be very Apertium-specific, but in that case we just keep our own file and append the entries with update-lexc.sh.) |
||
Latest revision as of 09:46, 15 April 2015
Documentation and TODO's for apertium-sme-nob's sme morphological analyser from Giellatekno.
Contents
Description[edit]
The sme morphological analyser is a trimmed and tag-changed version of the one in Giellatekno.
Trimming[edit]
Trimming happens using the same HFST method as in the Turkic pairs etc. Compounds are not handled correctly by this method.
Tagset changes[edit]
The tagset in apertium-sme-nob is closer to Apertium's tagset, thus a bit different from the one in Giellatekno.
- We remove Err/Orth/Usage tags (+Use/Sub, etc) –
- We remove any derivational analyses that aren't yet handled by transfer/bidix
- We reorder some tags
- We change the format of tags, so +N becomes <n>, +Der/PassL becomes <der_passl>, etc.
- We change certain tags
TODO[edit]
Misc[edit]
- add entries from bidix that are missing from the analyser
- regex for URL's (don't want telefonkatalogen.no => *telefonkatalogen.nå)
- regex for acronyms like "GsoC:as" (tokenisation dependent...)
8632: SUBSTITUTE:TV_IV (V TV) (V IV) FAUXV (0 ("lávet"));-- this should be analysed as both, and disambiguated
Typos[edit]
I've seen "odda" many places ("ođđa"), can we just add these to the analyser? (Would be cool to have charlifter/Diacritic Restoration, but until then…)
- a list of high-frequency typos where the correction has an analysis
Dashes[edit]
lexc handles dashes by adding them literally (like a lemma), doing that in with a bidix pardef would be very messy (also, doesn't it give issues with lemma-matching in CG?).
Multiwords[edit]
Add simple multiwords and fixed expressions to the analyser.
- dasa lassin => i tillegg (til det)
- dán áigge => for tiden
- mun ieš => meg selv
- bures boahtin => velkommen
- Buorre beaivi => God dag
- leat guollebivddus => å fiske
- maid ban dainna => hva i all verden
- jagis jahkái => fra år til år
- oaidnaleapmái => 'see you'
- ovdamearkka => for eksempel
- Mo manná? => Hvordan går det?
- ja nu ain => og så videre
(Some of these MWE's might be very Apertium-specific, but in that case we just keep our own file and append the entries with update-lexc.sh.)
Also, oktavuohta.N.Sg.Loc turns into an mwe preposition in nob:
- oktavuođas => i forbindelse med
it'd be a lot simpler for transfer to just analyse such a fixed expression as oktavuođas.Po in the first place.
