kylebgorman
commented
5 years ago I assume it's going to be the case that any of the rational algorithms applied to a totally empty FST are going to produce strange output.
If however you have an FST with a valid start and final state but no transitions things work as expected.
On Wed, Jan 23, 2019 at 9:22 AM Kay-Michael Würzner < notifications@github.com> wrote:
Most likely intended but I thought noteworthy nonetheless: Whenever I concatenate an fst to an empty fst (e.g. empty result of a composition) the result is also empty. From a formal point of view, this seems to be perfectly fine (cf. e.g. https://courses.engr.illinois.edu/cs373/fa2013/Lectures/lec06.pdf). It is however not explicitly mentioned in the documentation ( http://www.opengrm.org/twiki/bin/view/FST/ConcatDoc).
— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/kylebgorman/Pynini/issues/12, or mute the thread https://github.com/notifications/unsubscribe-auth/AAJuOfnBg2JxlxVt3__tVK-5h0Uw2Wp8ks5vGHBDgaJpZM4aO4qf .
wrznr
Most likely intended but I thought noteworthy nonetheless: Whenever I concatenate an fst to an empty fst (e.g. empty result of a composition) the result is also empty. From a formal point of view, this seems to be perfectly fine (cf. e.g. https://courses.engr.illinois.edu/cs373/fa2013/Lectures/lec06.pdf). It is however not explicitly mentioned in the documentation (http://www.opengrm.org/twiki/bin/view/FST/ConcatDoc).