Three Dollar Quill

I've got some stray thoughts on the synthesis stuff (like, it turns out that both volume and root frequency affect the spectra of piano notes in qualitative ways, so I need to spend some pondering time on how the heck do I wire in this additive synthesis code in a useful way), but I'm right now trying to rebuild my fundamentals for working with code. My plan there is to design a simpler system, and work through building it the way that I think is right.

Here's the idea:

I'm interested in whether I can use the cut-up technique to deal with the blank page problem in writing. There are various apps and websites out there, but I thought it would be nice to roll my own. My goal is to work my way up to something like the following set of features:

• Input via text box or file chooser.
• Extract text streams from various file formats.
• Choose between different strategies for chunking text, with different controls for different strategies.
• Choose the number of samples to use in the output.
• Generate aggregate statistics to chunked text.
• Assign different weights to different sources?

I've put together a minimal set of features and defaults to get things rolling, so what I need to do now is focus on writing tests, and to get that working, I need to figure out what kind of data structures I'll be working with. A lot of them are going to be built in, like Sequence[str] and collections.Counter. The main new stuff is going to be attrs classes defining the chunking strategies; chunking should be done using a method, possibly __call__, that takes an iterator that yields "words" and returns an iterator that yields "chunks".

I think this is enough planning for me to get something together. Let's see if there are any really obvious things I've missed thus far.

...

Nothing seems wrong just yet. I don't want to push myself, so I'm going to try to wind down. I can hopefully get this somewhere interesting over the next few days.

Good night.