Samplers len() might need to be removed

[pzelasko]

When working on #372, I noticed that there is an issue in calling len(sampler). We currently compute len by iterating the sampler and caching the number of batches. However, the number of batches (i.e., len) may change when we shuffle the cuts. I don't think there is a really elegant way to provide len (we can't recompute it at the start of an iteration, because len itself calls iter(sampler)...). Unless anybody has some idea, I'll remove len from samplers.

CC @janvainer @danpovey @csukuangfj

[danpovey]

I agree that removing len() is probably the best option. If an approximate length is really needed for something, perhaps we could add a function that returns an approximated length.

…

On Sat, Aug 14, 2021 at 11:42 AM Piotr Żelasko ***@***.***> wrote: When working on #372 <#372>, I noticed that there is an issue in calling len(sampler). We currently compute len by iterating the sampler and caching the number of batches. However, the number of batches (i.e., len) may change when we shuffle the cuts. I don't think there is a really elegant way to provide len (we can't recompute it at the start of an iteration, because len itself calls iter(sampler)...). Unless anybody has some idea, I'll remove len from samplers. CC @janvainer <https://github.com/janvainer> @danpovey <https://github.com/danpovey> @csukuangfj <https://github.com/csukuangfj> — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#373>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAZFLOZM3DZGIFYYCOOM4STT4XQ2FANCNFSM5CEU6LDQ> . Triage notifications on the go with GitHub Mobile for iOS <https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675> or Android <https://play.google.com/store/apps/details?id=com.github.android&utm_campaign=notification-email> .