POST: intro dev blog / segment workflow changes after NMAC 2024 GRC

[NickleDave]

• Decide we definitely want to be able to save Segments in some flat file format after discussion with Nick Jourjine: to avoid having a file full of audio clips produced by segmenting, when you already have the audio and what you really need is some way to programatically work with your segmentations
  • one might ask, why bother being able to save? Because we also often want to do some analysis on the segmentation itself. For bonus points give an example of this

[NickleDave]

• talk about removing path attribute from Sound, making the library more functional (in the programming sense)
  • because Sound is just a container for data, the same way a numpy.array is, and we are most likely going to want to do some operations on that data, not just have it stay static. So we are more likely than not to break the contract that "this Sound is exactly the data from this file"

[NickleDave]

• Note that we remove path and audio_path from Spectrogram by similar reasoning: it's a strong assumption that a Spectrogram can always be mapped to a path, and then we add an additional assumption that we have a 1-to-1 mapping from an audio file to the spectrogram
• I think we want to be able to declare this mapping when we build a FilesDataset for purposes of tracking ... but we don't want to enforce it all the time when we're doing some interactive analysis

[NickleDave]

I should talk about the workshop, thank Nick and Diana, talk about the need for computational specific behavioral conferences, at the top of the blog post!!! Link to NMAC GRC site!!!

[NickleDave]

In the spirit of making blog posts shorter:
only talk about the main API change here, write separate posts about adding vignettes from workshop, and possibly about re-writing examples API? (maybe that last one should be my own personal blog post since it's not really about how VocalPy per se)