bowersd / otw Goto Github PK
View Code? Open in Web Editor NEWSource files for Odawa/Nishnaabemwin morphological analysis (xfst/hfst/foma)
License: Other
Source files for Odawa/Nishnaabemwin morphological analysis (xfst/hfst/foma)
License: Other
Per the OPD, izhi (aka naa in the red dictionary) has the reciprocal idiwag and the reflexive idizo, which means the special treatment for naa needs to extend to derivational contexts.
There was an error on de-dgoshin (perhaps should be spelled dedgoshin), which is a reduplicated form with initial change. Make sure the grammar can enforce initial change on a reduplicated form.
I think I remember URs being listed with two options on the same line in the original lexicon exports. These were converted into separate entries in our lexicon, but I don't think that was necessarily a good idea.
The base model currently has been revised to allow a hyphen between a n/g/w- prefix and the stem, following usage by A. Corbiere. I think this should be shifted to a variant spelling system module.
M. A. Corbiere often places a hyphen before the negative suffix. The model does not currently support this, but a variant spelling module could certainly be added.
Nichols (1980:221-227) discusses relational paradigms. These are VAIs/VTIs that get a suffix -w plus the -aa theme sign plus number information IIRC. The model does not handle this at all.
VAI and VTI cnj 3pl suffix -waad differs in the grammar, I believe. In VAI it is given as just -waad, but for VTI it is -(o|a)waad. I think this was due to how I originally interpreted Valentine's grammar, but I encountered an analyzer failure on a VAI where just -waad was expected but -owaad was written.
Analyzer expects,Written,Comment ("" = empty string, ","=separator)
h,'
n#,""
nyag#,nhig#,plurals/obviatives ... the h is non-standard in double vowel, but the vowel change should be accomodated
yan#,yin#,a->i needs to be phonological, since this is VAI (2nd person)
yag#,yig
#mm,m
iiwii,ii'ii
n's,ns,nitpicky
gnd,ngnd
ewaa,e'aa
ihii,ii
yan#,'in#,VAI (2nd person)
zh#,sh#
hii,yii
g',g
bt,pt
o#,a#
(From Gnahmaading, Mchigiing Mskwaabiimzhiins story)
there is a hack to make oog be an ok suffix after non-nasals. wag should be restored in the properly just version of the analyzers.
Immediate imperatives have +Imm tags. This is superflous. It should just be prohibitives getting +Neg, and delayed imperatives getting +Del, and immediate imperatives getting nothing
VAIOs seem to be listed in two places, one of which has no sublexicon, and so gets dumped under QNT when compiling.
Pre(verb/nouns/adverbs) have user facing entries without syncope, which is inconsistent with the general behavior of the FST.
Secondary derivation (building new stems from pre-existing stems) is not fully active. Major suffixes that should be made active are in src/morphology/gamig.lexc. These include -gamig (building) -win (nominalizer), -gaazo/-gaade (lexical passive), -nini (man who X's, X-man) -kwe (woman who X's, X-woman). I believe there are some VTA-deriving suffixes that could also be added.
I think I encountered a problem with the representation of the extended prefixes, which was causing unexpected deletions in *VCV stems (others?)
The phonology module currently has rules for variant spellings in the phonology module (because they reference a particular special character w5 IIRC). These should not be in phonology, because they are specific to a single collection of texts.
Nicknames are often formed by truncation of the last syllable of a VAI: gaashkjaane (have a split nose) -> gaashkjaan.
An example I encountered was gaashkjaanzh ... which is surprising because jaanzh is the independent form of "nose", not what you would expect in a verb. So it would have been /gaashkjaanzh/ -> [gaashkjaanzh] (no truncation ... not exactly sure what morphological process would produce this combination), not /gaashkjaanzhe/ -> [gaashkjaanzh] (truncation).
Keeping track of words that are missing
Verbs of possession are derived via:
3+NOUN+PosThm+i
I think I omitted VTA Preterite Dubitatives
There was a major lack of data in VTI paradigms, so the grammar may not be complete there
Analyzer Expects, Written, Comment
epiichi-,epichi-
ggishkaajge,ggishkaaye
gnimaa gye,gmaage,unsure if this is expected or gmaa gye
gmaa gye,gmaage,unsure if this is expected or gnimaa gye
niwi yahiing,niiyiing
giiyenh,ge'enh,this is related to the glide/glottal fluidity that we have seen in the trivial spelling changes
(From Gnahmaading, Mchigiing Mskwaabiimzhiins)
โ (right/left curly single quote)
*
these are not managed by the punctuation portions of the FST
A lot of words are listed twice, with only a minor deviation in spelling. Often the gloss for the variant is just a redirection
VTAs ending with glottal stops are not managed very well in the grammar
Noun-Noun compounding is not currently supported. Also numeral-noun compounding is not supported.
A declarative, efficient, and flexible JavaScript library for building user interfaces.
๐ Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
An Open Source Machine Learning Framework for Everyone
The Web framework for perfectionists with deadlines.
A PHP framework for web artisans
Bring data to life with SVG, Canvas and HTML. ๐๐๐
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
Some thing interesting about web. New door for the world.
A server is a program made to process requests and deliver data to clients.
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
Some thing interesting about visualization, use data art
Some thing interesting about game, make everyone happy.
We are working to build community through open source technology. NB: members must have two-factor auth.
Open source projects and samples from Microsoft.
Google โค๏ธ Open Source for everyone.
Alibaba Open Source for everyone
Data-Driven Documents codes.
China tencent open source team.