A trie data structure created for the sole purpose of easily reducing a large list of strings into a small list of strings.
More specifically, this library was created with only one use case in mind. A certain CDN only allows a small number of concurrent purge requests. If you have a thousands of purgeable files with frequent purges of various combinations of these files things get a little more complicated. You don't want to purge the contents of the entire folder every time you need to purge something but you also don't want to wait a long time for files to be purged. Fortunately, the CDN supports file globbering or the use of the '*' as a wildcard at the end of filename. So it is possible to purge these files:
somefile.txt
someotherfile.txt
some.txtwith thise purge request:
some*This library makes it easy to generate a handful of purge strings from large lists of resources to be purged. Try it out by
running the test file or experimenting on the list of files given in testset.txt in this repo.
test
testify
testament
testimony
tesseract
tess
If one were to insert every string from the list above into a trie a
structure that resembles the following diagram would be made.
['t']
|
['es']
/ \
['s'] ['t']
/ / \
['eract'] ['ament'] ['i']
/ \
['fy'] ['mony']
Then if you were to DeleteWords(6, '*') then you would end up with something like this:
['t']
| test
['es'] tess
/ \ testament
['s'] ['t'] tesseract
/ / \ testi*
['eract'] ['ament'] ['i*']
Or if you wer to run DeleteWords(3, '*') you would get something like this:
['t'] tess*
| test*
['es']
/ \
['s*'] ['t*']