tatoeba2 icon indicating copy to clipboard operation
tatoeba2 copied to clipboard

Kabyle / Berber - Problem with Search Results

Open ckjpn opened this issue 3 years ago • 2 comments

Note that this search for sentences in Berber with audio files shows a lot of Kabyle sentences.

https://tatoeba.org/en/sentences/search?orphans=any&sort=created&has_audio=yes&from=ber (1,000 results out of 4,659 occurrences)

This search for Kabyle sentences with audio seems to work.

https://tatoeba.org/en/sentences/search?orphans=any&sort=created&has_audio=yes&from=kab (1,000 results out of 38,404 occurrences)

Perhaps this is related to the fact the apparently Kabyle is down 41,236 sentences since this time last year.

http://tatoeba.ueuo.com/stats-2022-08-04-languages.html

Screen Shot 2022-08-04 at 15 54 46

ckjpn avatar Aug 04 '22 06:08 ckjpn

This is likely a related bug, so I'll add it here. If it is unrelated, please start a new issue.

This search to find all the "orphan" sentence that have audio on Igider's "audio" list (8407) gets 0 results, even though there are sentences on his list that are orphans.

&list=8407 &orphans=yes

https://tatoeba.org/en/sentences/search?from=&has_audio=&list=8407&native=&orphans=yes&query=&sort=modified&sort_reverse=&tags=&to=none&trans_filter=limit&trans_has_audio=&trans_link=&trans_orphan=&trans_to=&trans_unapproved=&trans_user=&unapproved=no&user=

The same search using samir_t's lists gets the same results

&list=8328 &orphans=yes

https://tatoeba.org/en/sentences/search?from=&has_audio=&list=8328&native=&orphans=yes&query=&sort=modified&sort_reverse=&tags=&to=none&trans_filter=limit&trans_has_audio=&trans_link=&trans_orphan=&trans_to=&trans_unapproved=&trans_user=&unapproved=no&user=

However, this search looking for Berber sentences with audio, which now mistakenly shows Kabyle sentences, shows a number of these sentences with audio that are orphans.

from=ber &has_audio=yes &orphans=any

https://tatoeba.org/en/sentences/search?from=ber&has_audio=yes&native=&orphans=any&query=&sort=created&sort_reverse=&tags=&to=none&trans_filter=limit&trans_has_audio=&trans_link=&trans_orphan=&trans_to=&trans_unapproved=&trans_user=&unapproved=no&user=

Here is a screenshot, just in case these orphans get adopted before this matter is looked into.

Screen Shot 2022-08-06 at 7 20 08

Changing "orphans=any" to "orphans=yes" gets "Advanced search (0 results)"

from=ber &has_audio=yes &orphans=yes

https://tatoeba.org/en/sentences/search?from=ber&has_audio=yes&native=&orphans=yes&query=&sort=created&sort_reverse=&tags=&to=none&trans_filter=limit&trans_has_audio=&trans_link=&trans_orphan=&trans_to=&trans_unapproved=&trans_user=&unapproved=no&user=

Changing "from=ber" to "from=kab" finds the sentences

from=kab &has_audio=yes &orphans=yes

https://tatoeba.org/en/sentences/search?from=kab&has_audio=yes&native=&orphans=yes&query=&sort=created&sort_reverse=&tags=&to=none&trans_filter=limit&trans_has_audio=&trans_link=&trans_orphan=&trans_to=&trans_unapproved=&trans_user=&unapproved=no&user=

Advanced search (21 results)

I've set these lists to public

I've temporarily changed all the following lists from "listed" to "public" so you can easily see that the sentences are on these lists when visiting the sentences' pages. Please let me know when this has been taken care of so I can reset these back to "listed".

Igider's audio list https://tatoeba.org/en/sentences_lists/show/8407

samir_t's audio list https://tatoeba.org/en/sentences_lists/show/8328

Yazid_Bouhamam's audio list https://tatoeba.org/en/sentences_lists/show/8051

Selyan's audio list https://tatoeba.org/en/sentences_lists/show/8046

ckjpn avatar Aug 05 '22 22:08 ckjpn

Strange. 😮 My preliminary guess is that a bunch of Kabyle sentences could have been changed into Berber, and this massively triggered an existing bug (a bug that was rare enough to go unnoticed until now) in the reindexation of sentences having audio when their flag is changed.

jiru avatar Aug 28 '22 07:08 jiru