Another (minor) bug in the PEG morphology

16 views
Skip to first unread message

Jorge Llambías

unread,
Oct 25, 2014, 9:11:14 PM10/25/14
to bpfk...@googlegroups.com
I found a new minor bug in the morphology, It only affects certain lujvo with fu'ivla rafsi which are not used very much so it's not a big surprise it wasn't detected before.

The problem occurs when fu'ivla rafsi are followed by a short one-syllable rafsi and nothing more. So for example: "spageti'yprenu" works fine, but "spageti'ypre", which should also parse, doesn't.

I have changed the stress rule, which I think is what's causing the problem, to:

stress <- consonant* h? y? syllable pause

I added "h?", which is needed in order to detect the stress in stressed extended rafsi.

Could someone check whether this fixes the problem? Thanks!

mu'o mi'e xorxes

Alex Burka

unread,
Oct 25, 2014, 9:13:02 PM10/25/14
to bpfk...@googlegroups.com

Heh, I had noticed this problem when playing with fu'ivla rafsi, but I just assumed I didn't understand how to use them correctly!

--
You received this message because you are subscribed to the Google Groups "BPFK" group.
To unsubscribe from this group and stop receiving emails from it, send an email to bpfk-list+...@googlegroups.com.
To post to this group, send email to bpfk...@googlegroups.com.
Visit this group at http://groups.google.com/group/bpfk-list.
For more options, visit https://groups.google.com/d/optout.

Jorge Llambías

unread,
Oct 25, 2014, 9:17:01 PM10/25/14
to bpfk...@googlegroups.com
On Sat, Oct 25, 2014 at 10:13 PM, Alex Burka <dur...@gmail.com> wrote:

Heh, I had noticed this problem when playing with fu'ivla rafsi, but I just assumed I didn't understand how to use them correctly!


In fact spageTI''ypre parses ok, so it's the detection of the stress by syllable count which is failing.

Alex Burka

unread,
Oct 26, 2014, 1:17:23 AM10/26/14
to bpfk...@googlegroups.com
ja'aku si'au lo javni nunstika be do cu tolyspofu'ygau[1] lo lujvo poi se fanmo pa tenrafsi ce'o pa cmarafsi

[1] $ ./vlatai.py "tolyspofu'ygau"
tolyspofu'ygau: lujvo/tol-spofu-gau

mu'o mi'e la durkavore

mukti

unread,
Oct 26, 2014, 3:08:17 AM10/26/14
to bpfk...@googlegroups.com
On Saturday, October 25, 2014 10:11:14 PM UTC-3, xorxes wrote:
I have changed the stress rule, which I think is what's causing the problem, to:

stress <- consonant* h? y? syllable pause

I re-evaluated the camxes test corpus and the current contents of jbovlaste with this modification, and there are no changes beyond what has been discussed in the other thread for v122 of the morphology.

mi'e la mukti mu'o  

Jorge Llambías

unread,
Oct 26, 2014, 10:03:14 AM10/26/14
to bpfk...@googlegroups.com

On Sun, Oct 26, 2014 at 2:17 AM, Alex Burka <dur...@gmail.com> wrote:
ja'aku si'au lo javni nunstika be do cu tolyspofu'ygau[1] lo lujvo poi se fanmo pa tenrafsi ce'o pa cmarafsi

[1] $ ./vlatai.py "tolyspofu'ygau"
tolyspofu'ygau: lujvo/tol-spofu-gau
 
On Sun, Oct 26, 2014 at 4:08 AM, mukti <shun...@gmail.com> wrote:

I re-evaluated the camxes test corpus and the current contents of jbovlaste with this modification, and there are no changes beyond what has been discussed in the other thread for v122 of the morphology.

ki'e sai doi lo re do
Reply all
Reply to author
Forward
0 new messages