Some more getters

[ralsei]

This adds a few utility functions, and therefore closes #8.

Changes to existing functions:

• bsearch now has a #:rightmost? argument, which determines whether to get the index of the rightmost element in the vector, rather than the leftmost one. necessary for series-index-range

New functions:

• series-index-range: binary searches on a series, and returns both the left and rightmost index of an element. useful for when a series has duplicates
• df-index-range: same for data-frames
• series-is-sorted?: determines if a series has a comparison function. useful for knowing whether series-index-of is possible, or sorting/linear search is required
• df-is-sorted?: same for data-frames
• series-shallow-copy: copies a series's vector, retaining all of its properties
• df-duplicate-series: returns a copy of the given series from the data-frame, retaining all of its properties, using series-shallow-copy

Nit-picks:

• added df-shallow-copy to docs
• fixed for/data-frame syntax in docs (call that cleaning up my own mess)

All functions have tests, but do feel free to let me know if they're insufficient.

[alex-hhh]

Another consideration you might want to give to df-index-range is whether it should return two values or a cons cell. If you find that most of the uses of df-index-range are like so:

(match-define (cons start end) (df-index-range ...))

You might want to make df-index-range return two values instead, as this would be more efficient.

[alex-hhh]

Hi @ralsei , can you pull from the ah/lower-bound branch as this implements the lower-bound and upper-bound functions, plus an equal-range function which you can use in the series-index-range function? These functions have better semantics than bsearch, but require the use of strict ordering for the sort function (e.g. use < instead of <=)

https://github.com/alex-hhh/data-frame/tree/ah/lower-bound

[ralsei]

So the issue with the changes in the ah/lower-bound branch is that sorting with <= no longer produces expected behavior, but using check-valid-sort with < and duplicate items in a series fails. This is why I added an equal? check -- but I don't know if this is what is wanted.

Still going to incorporate the requested changes.

[ralsei]

I think I have resolved all requested changes. Feel free to tell me if it doesn't.