Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

The French to Arpitan translator is not working properly on the website #108

Closed
hectoralos opened this issue Oct 10, 2020 · 4 comments
Closed

Comments

@hectoralos
Copy link
Member

The recently released French to Arpitan translator is not working properly on the website (www.apertium.org). If one puts a text and tries to translate it, most of the time he/she receives a blank page or a previous translation into Arpitan. Instead, if the same text is translated into, for instance, Spanish or Occitan, one quickly gets the correct translation into the given language. The translation from Arpitan into French seems to work properly.

The result of this issue is that, for some unknown reason, the translator is not reliable and very hard to use.

@TinoDidriksen
Copy link
Member

apertium/apertium-apy#17 would work around the issue, though fixing the actual problem would be better. Your null-flush vs. non-null-flush testcase showed differences already in morphology, so that's the place to start.

@hectoralos
Copy link
Member Author

hectoralos commented Oct 10, 2020

The problem with null-flush vs. non-null-flush testcase is that we have the same odd results with null-flush for e.g. fra-oci, but the fra-oci pair is working well. In fact, it seems logical to have the same results in the morph mode for fra-oci and fra-frp because if there is something odd in the French morphological analyser, the problem very probably should be found in any translator from French. But they are all working, except fra-frp. So I can't understand your conclusion.

To be more precise, in fra-oci I am executing this:

make && cat ../apertium-fra/corpus/wikipedia.20000.txt | head -n 1000 | apertium -z -d . fra-oci-morph > yyy_morph
make && cat ../apertium-fra/corpus/wikipedia.20000.txt | head -n 1000 | perl -wpne 's/\n/\n\0/g;' | apertium -z -d . fra-oci-morph > xxx_morph
diff yyy_morph xxx_morph | more

And I am getting this:

209c209
< ^00209/00209<num>$^./.<sent>$ ^Économie/économie<n><f><sg>$ ^des/de<pr>+le<det><def><mf><pl>$ ^États-Unis/États-Unis<np><top><m><pl>$
---
> ^00209/00209<num>$^./.<sent>$ ^Économie/économie<n><f><sg>$ ^des/de<pr>+le<det><def><mf><pl>$ 
218c218
< ^00218/00218<num>$^./.<sent>$ - ^Louis/Louis<np><ant><m><sg>$ ^II/II<num><mf><sp>$ ^de/de<pr>$ ^France/France<np><top><f><sg>$
---
> ^00218/00218<num>$^./.<sent>$ - ^Louis/Louis<np><ant><m><sg>$ ^II/II<num><mf><sp>$ ^de/de<pr>$ 
545c545
< ^00545/00545<num>$^./.<sent>$ ^!/!<sent>$ ^scope/*scope$=^col/col<n><m><sg>$ ^width/*width$="^30/30<num>$%" ^|/|<sent>$  ^Tombe/tombe<n><f><sg>/tombe
r<vblex><pri><p3><sg>/tomber<vblex><pri><p1><sg>/tomber<vblex><prs><p3><sg>/tomber<vblex><prs><p1><sg>/tomber<vblex><imp><p2><sg>$
---
> ^00545/00545<num>$^./.<sent>$ ^!/!<sent>$ ^scope/*scope$=^col/col<n><m><sg>$ ^width/*width$="^30/30<num>$%" ^|/|<sent>$  
835c835
< ^00835/00835<num>$^./.<sent>$ ^L'/le<det><def><mf><sg>/le<prn><pro><p3><f><sg>/le<prn><pro><p3><m><sg>$ ^élimination/élimination<n><f><sg>$ ^entraîne/
entraîner<vblex><pri><p3><sg>/entraîner<vblex><pri><p1><sg>/entraîner<vblex><prs><p3><sg>/entraîner<vblex><prs><p1><sg>/entraîner<vblex><imp><p2><sg>$ ^
la/le<det><def><f><sg>/le<prn><pro><p3><f><sg>$ ^disqualification/disqualification<n><f><sg>$ ^et/et<cnjcoo>$ ^la/le<det><def><f><sg>/le<prn><pro><p3><f
><sg>$ ^sortie/sortie<n><f><sg>/sortir<vblex><pp><f><sg>$ ^immédiate/immédiat<adj><f><sg>$ ^du/de<pr>+le<det><def><m><sg>$ ^conducteur/conducteur<adj><m
><sg>/conducteur<n><m><sg>$ ^et/et<cnjcoo>$ ^de/de<pr>$
---
> ^00835/00835<num>$^./.<sent>$ ^L'/le<det><def><mf><sg>/le<prn><pro><p3><f><sg>/le<prn><pro><p3><m><sg>$ ^élimination/élimination<n><f><sg>$ ^entraîne/
entraîner<vblex><pri><p3><sg>/entraîner<vblex><pri><p1><sg>/entraîner<vblex><prs><p3><sg>/entraîner<vblex><prs><p1><sg>/entraîner<vblex><imp><p2><sg>$ ^
la/le<det><def><f><sg>/le<prn><pro><p3><f><sg>$ ^disqualification/disqualification<n><f><sg>$ ^et/et<cnjcoo>$ ^la/le<det><def><f><sg>/le<prn><pro><p3><f
><sg>$ ^sortie/sortie<n><f><sg>/sortir<vblex><pp><f><sg>$ ^immédiate/immédiat<adj><f><sg>$ ^du/de<pr>+le<det><def><m><sg>$ ^conducteur/conducteur<adj><m
><sg>/conducteur<n><m><sg>$ ^et/et<cnjcoo>$ 
1000c1000,1001
< ^01000/01000<num>$^./.<sent>$ ^En/en<pr>/en<prn><pro><adv>$ ^mathématiques/mathématique<adj><mf><pl>/mathématique<n><f><pl>$^,/,<cm>$ ^un/un<num><m><s
p>/un<det><ind><m><sg>/un<prn><tn><m><sg>$ ^diagramme/diagramme<n><m><sg>$ ^est/est<adj><mf><sp>/est<n><m><sg>/être<vbser><pri><p3><sg>$ ^un/un<num><m><
sp>/un<det><ind><m><sg>/un<prn><tn><m><sg>$ ^schéma/schéma<n><m><sg>$ ^utilisé/utiliser<vblex><pp><m><sg>$ ^pour/pour<adv>/pour<pr>$ ^représenter/représ
enter<vblex><inf>$ ^des/de<pr>+le<det><def><mf><pl>$ ^objets/objet<n><m><pl>$ ^et/et<cnjcoo>$ ^servant/servir<vblex><ger>$ ^de/de<pr>$ ^support/support<
n><m><sg>$ ^du/de<pr>+le<det><def><m><sg>$ ^raisonnement/raisonnement<n><m><sg>$^./.<sent>$^./.<sent>$
---
> ^01000/01000<num>$^./.<sent>$ ^En/en<pr>/en<prn><pro><adv>$ ^mathématiques/mathématique<adj><mf><pl>/mathématique<n><f><pl>$^,/,<cm>$ ^un/un<num><m><s
p>/un<det><ind><m><sg>/un<prn><tn><m><sg>$ ^diagramme/diagramme<n><m><sg>$ ^est/est<adj><mf><sp>/est<n><m><sg>/être<vbser><pri><p3><sg>$ ^un/un<num><m><
sp>/un<det><ind><m><sg>/un<prn><tn><m><sg>$ ^schéma/schéma<n><m><sg>$ ^utilisé/utiliser<vblex><pp><m><sg>$ ^pour/pour<adv>/pour<pr>$ ^représenter/représ
enter<vblex><inf>$ ^des/de<pr>+le<det><def><mf><pl>$ ^objets/objet<n><m><pl>$ ^et/et<cnjcoo>$ ^servant/servir<vblex><ger>$ ^de/de<pr>$ ^support/support<
n><m><sg>$ ^du/de<pr>+le<det><def><m><sg>$ ^raisonnement/raisonnement<n><m><sg>$^./.<sent>$
> ^./.<sent>$

This seemingly is not a problem for fra-oci. Why would it cause the problem in fra-frp?

@TinoDidriksen
Copy link
Member

apertium/apertium-apy#17 is now worked around and deployed on beta, and seems to be working exactly as desired.

@hectoralos
Copy link
Member Author

Yes, it is!! Hopefully, it can be deployed soon in the "real" web.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

2 participants