TODO for suggestions module

[mathematicalmichael]

not yet implemented:

• [ ] argparse
• [ ] not part of setup.py
• [ ] testing

[mathematicalmichael]

low priority because it does work.

arguments:

• threshold for similarity (set at 75% now because 80% wasn't automating shorter words)
• n-gram length

[mathematicalmichael]

make sure that the program can run even without the presence of matches. think of this as an "extra feature", however it is generated.

right now, the package doesn't require pandas or anything like that, which is pretty convenient, although inefficient since it's reading files.

one possible solution that "meets in the middle" is to use pandas within suggest to create a new dataframe, with each column name being what is currently used as the file name, and the entries for each are below. Then you can save it as a JSON and read that in as an extra data-source once (on startup). this minimizes I/O operations to disk, but does increase the RAM that the program needs.

TL;DR make dictionary, save and load once, instead of saving and reading lots of files.