Add method iterRecords_range

[lguez]

I have tested this by reading about one million record:

#!/usr/bin/env python3

import time

import shapefile

r = shapefile.Reader("extremum")
t0 = time.perf_counter()

for i in range(10, 1000000):
    x = r.record(i)

t1 = time.perf_counter()
print(t1 - t0)
t0 = time.perf_counter()

for x in r.iterRecords_range(10, 1000000):
    pass

t1 = time.perf_counter()
print(t1 - t0)

I get:

$ test_iterRecords_range.py 
21.27043527108617
7.878919951850548

So iterRecords_range is somewhat faster.

[JamesParrott]

Great stuff - thanks Lionel :). This appears to be good from a quick look - I'll review it fully over the next few days and get back to you.

[lguez]

OK. Thanks. I have written a separate method because you said you prefered so, but I still think it would make simpler code and it would be clearer to the user to have additional optional arguments to the iterRecords method:

def iterRecords(self, fields=None, start=0, stop=None)

[JamesParrott]

Yeah, adjusting iterRecords is more sensible if the range iterator is not customisable, just start and stop (in future, step could be added too). It avoids duplication.

How's this look?
lguez/pyshp@master...GeospatialPython:pyshp:combine_iterRecords_range_into_iterRecords

[edit] I added one test, and the branch above passes it (and also passes all the previous ones).

I think "yields the same _Records, as Reader.record" covers the claim and the main contract with the user.

Can you, or anyone, think of anything else that should be tested?

[lguez]

No, it seems fine. Thanks.

[JamesParrott]

A branch has been made from this PR, which will be merged: https://github.com/GeospatialPython/pyshp/tree/combine_iterRecords_range_into_iterRecords