Welsh to English/Regression tests

To run these tests, run the script regression-tests.sh from the apertium-cy-en language pair in SVN.

Treatment of 'is' in present tense.

• (en) The boy is in the garden. → mae y bachgen yn yr ardd. (note: yr → 'r is an open bug)
• (cy) mae'r bachgen yn yr ardd. → the boy is in the garden.

These are both correct (apart from the 'r), but I thought "regressions" were when you fix something and in the process break something else? Re 'r:

In Welsh pattern "aeiouwy + space + y[r]"
output "aeiouwy + 'r"

Donnek

Yep, so these should be 'regression tests' :) --

Yep, I know the pattern, the problem is that the post-generator insists on having a ~ before anything that it deals with -- This would mean that we have to have '~' before every vowel, which would be quite difficult. There is another possibility though, if we can't fix that and it would be to just use a plain transliterator to replace:

"aeiouwy + space + ~yr + space" with "aeiouwy'r + space"

Can you think of anything this might catch by accident? or is it a fairly safe search/replace? - Francis Tyers

I would go with this in the meantime - I think it's pretty safe. Note that your rule can act on both 'y' and 'yr'. The system is:

consonant + space + y + space + consonant

consonant + space + yr + space + vowel

vowel + 'r + space + consonant-or-vowel

Donnek

No subject shift with imperative

• (cy) gwasgwch y botwm! → squeeze the button!
• (en) squeeze the button! → gwasgu y botwm! (note: infinitive for imperative is an open bug)

"yn" as stative

• (cy) yn falch → proud
• (cy) yn hapus → happy
• (cy) tyfodd fo yn fawr → he grew big

Subject shift for pronouns

• (cy) roedden nhw'n hapus → they were happy

Number agreement of verb

• (cy) roedd y bechgyn yn hapus → the boys were happy
• (cy) roedd y cwningod yn hapus → the rabbits were happy
• (cy) gwelodd y dyn y llyfr → the man saw the book

"yn" as "-ing"

• (cy) yn mynd → going
• (cy) yn gweld → seeing

Conjugation of 'mynd'

• (cy) aeth fo → he went

Comparative adjectives

• (cy) tyfodd y twnnel yn fwy → the tunnel grew bigger
• (cy) tyfodd y twnnel yn llai → the tunnel grew smaller
• (cy) tyfodd y twnnel yn hirach → the tunnel grew longer
• (cy) tyfodd y twnnel yn uwch → the tunnel grew higher

Insertion of indefinite determiner

• (cy) daeth Taid â lamp → Grandfather came with a lamp
• (cy) dychwelodd y rheolwr gyda gŵr tew → the manager returned with a fat man

Verbal nouns / Infinitives

• (cy) roedd y dyn yn gwerthu pethau rhad → the man was selling cheap things
• (cy) roedd fo yn palu → he was digging
• (cy) ceisiodd y dyn agor y bocs → the man tried to open the box

Synthetic to non-synthetic adjectives

• (en) the nice house → y tŷ neis
• (en) the nicer house → y tŷ neisiach
• (en) the nicest house → y tŷ neisiaf

• (en) the dirty cat → y gath budr
• (en) the dirtier cat → y gath fwy budr
• (en) the dirtiest cat → y gath fwyaf budr

Yes, the synthetic endings would be nice to sort out, but I don't think there would be a mutation here, so that's not a bug. - Donnek

The synthetic endings there are valid then? - Francis Tyers

Yes, in this case you would tend to say "y tŷ neisiaf". - Donnek

Hmm. Something odd about these. "y gath fudr" - fem adj has SM. Because the adj is short, you would also tend to use "butrach" and "butraf" in the second two. "peryglus" (dangerous) might be a better adj to use.- Donnek

There are a few adjectives which need to be marked as synthetic in the monodix. Is there likely to be an easy way of doing this? - Francis Tyers

As a rough rule for the time being, I would say you could mark any adj of 2 syllables or less as synthetic, but of course that is not scientific. - Donnek

Subordinate ("reported speech") clauses with "bod" + noun

• (cy) clywodd y dyn bod y trên am gyrraedd yn hwyr. → the man heard that the train will arrive late.
• (cy) yr honiadau bod yr AS wedi methu datgan £103,000 o roddion. → the allegations that the MP has failed declare £103,000 of gifts.
• (cy) clywodd y dyn bod y trên ar gyrraedd yn hwyr → the man heard that the train was about to arrive late.

Infinitive after "yn lle"

• (cy) yn lle mynd dros y ffordd → instead of going over the road
• (en) instead of going over the road → yn lle mynd dros y f/ffordd (note: initial mutation error here needs to be fixed)