readbeyond / aeneas

aeneas is a Python/C library and a set of tools to automagically synchronize audio and text (aka forced alignment)
http://www.readbeyond.it/aeneas/
GNU Affero General Public License v3.0
2.49k stars 228 forks source link

Regional Settings (i.e French) cause Errors in Aeneas #162

Closed MattGyverLee closed 7 years ago

MattGyverLee commented 7 years ago

I'm using Aeneas through SIL's Scripture App Builder 3.2. In Aeneas 1.7, and presumably 1.5, there is an error thrown when Windows has a regional setting that uses comma as a whole/part separator instead of period.

English: $1,000.00 French : 1.000,00 $ or 1 000,00 $

regional

It seems that a process in Aeneas uses the regional settings to create a timing value u'3,00', and then passes this back into a process that can't handle the comma.

Steps to reproduce: -Run Aeneas using an English regional setting to set a baseline. It should work. -Change the Regional Settings from English to something that uses a different decimal and group separator, -French (France), for example. -Reboot. Run Aeneas again, and watch it crash.

sab error french regional error

~Matthew

readbeyond commented 7 years ago

Hi,

thank you for letting me know about this issue.

aeneas itself does not handle localization settings. aeneas just expects values in seconds with a dot "." as the decimal separator.

I do not plan to add "locale support" in aeneas in the foreseeable future --- maybe in v2, but I think that such a feature should be dealt with by the user interface/client program that uses aeneas as a "library" for computing the audio/text alignment. After all, ISO mandates "." as the decimal separator, and "second" as time unit, etc.

I think this issue should be reported to SIL, since it is the Reading/Scripture App Builder program that produces the configuration string and invokes aeneas with it.

In fact, in your second screenshot you can see that the configuration string has:

is_audio_file_head_detect_max=3,00

(and also is_audio_file_head_detect_min=1,00)

with "," and not "." as aeneas expects.

I will close this issue as soon as I get back home, please report it to SIL: http://software.sil.org/readingappbuilder/support/

Does it make sense? Or do I miss something from your issue report?

AP

On 02/25/2017 11:21 AM, erros84 wrote:

I'm using Aeneas through SIL's Scripture App Builder 3.2. In Aeneas 1.7, and presumably 1.5, there is an error thrown when Windows has a regional setting that uses comma as a thousands separator instead of period.

English: $1,000.00 French : 1.000,00 $ or 1 000,00 $

regional https://cloud.githubusercontent.com/assets/986339/23330350/0d8ae57a-fb4c-11e6-91a0-d86fed24c91f.PNG

It seems that a process in Aeneas uses the regional settings to create a timing value u'3,00', and then passes this back into a process that can't handle the comma.

sab error french regional error https://cloud.githubusercontent.com/assets/986339/23330316/1659efd0-fb4b-11e6-8586-37c0a9ae54c8.PNG

Steps to reproduce: -Run Aeneas using an English regional setting to set a baseline. It should work. -Change the Regional Settings from English to something that uses a different decimal and group separator, -French (France), for example. -Reboot. Run Aneas again, and watch it crash.

sab error french regional error https://cloud.githubusercontent.com/assets/986339/23330316/1659efd0-fb4b-11e6-8586-37c0a9ae54c8.PNG

~Matthew

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/readbeyond/aeneas/issues/162, or mute the thread https://github.com/notifications/unsubscribe-auth/AFEodqxQPSYgMhterrwtnm7Z7OPGzDFYks5rgACfgaJpZM4ML-Qh.

MattGyverLee commented 7 years ago

Your analysis seems correct, Thank you for your response, I have re-submitted on the Scripture App Builder JIRA.

pettarin commented 7 years ago

Alright, thank you.