Minimal Keys

Compute the minimal unique keys for a collection of strings or sequences.

This is intended for use in presenting data in a user interface. For example, the nbcollate command-line interface from the nbcollate package uses it to guess student names from notebook filenames, for use in the creation of the collated Jupyter notebook.

The minimal keys from ["assignments/alice/hw1.txt", "assignments/bob/hw1.txt"] are ["alice", "bob"], because the input strings share the common prefix "assignments/" and the common suffix ".txt".

The minimal keys from ["alice/assignments/hw1.txt", "bob/assignments/hw1.txt"] are also ["alice", "bob"], because the input strings share the common suffix "/assignments/hw1.txt".

Finally, the minimal keys from ["assignments/alice.txt", "assignments/bob.txt"] are—wait for it—["alice", "bob"].

There are options to ignore case, and to split the strings at different boundaries. (The default is split to on words, so that ["alice", "adam"] is not abbreviated to ["lice", "dam"].)

This is the same basic idea as a database superkey, except that the actual minimal unique keys are returned, instead of the attributes that select these keys.

The current implementation trims only the beginnings and ends of sequences, because this is all that I've needed so far. I have in my head a more sophisticated implementation that uses difflib, but it is too long to fit in the header of this README.

Install

$ pip install minimalkeys

Usage

>>> from minimalkeys import minimal_keys
>>> minimal_keys(["assignments/alice/hw1.txt", "assignments/bob/hw1.txt"])
['alice', 'bob']

For more examples, see the API documentation.

License

MIT