Fix arrow.Arrow.interval()

[dwrpayne]

arrow.Arrow.interval() produces unexpected results if the interval doesn't evenly divide the range. It leaves off the last chunk of time.

$> start = datetime(2013, 5, 5, 12, 30)
$> end = datetime(2013, 5, 5, 17, 15)
$> for r in arrow.Arrow.interval('hour', start, end, 4): print (r)
(<Arrow [2013-05-05T12:00:00+00:00]>, <Arrow [2013-05-05T15:59:59.999999+00:00]>)

Expected result:

$> for r in arrow.Arrow.interval('hour', start, end, 4): print (r)
(<Arrow [2013-05-05T12:00:00+00:00]>, <Arrow [2013-05-05T15:59:59.999999+00:00]>)
(<Arrow [2013-05-05T16:00:00+00:00]>, <Arrow [2013-05-05T17:59:59.999999+00:00]>)

[andrewelkins]

Can you add a test case for this?

[dwrpayne]

Sure, I'll do that tomorrow. Can you explain why it failed the CI build? I can't find a full log that details exactly which text didn't pass.

…

On Oct 10, 2017 17:11, "Andrew Elkins" ***@***.***> wrote: Can you add a test case for this? — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <#466 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AFmGt-VI3O7C_HYx8uMWBl024XhR_Fdkks5srAfNgaJpZM4PySCV> .

[codecov-io]

Codecov Report

Merging #466 into master will not change coverage.
The diff coverage is 100%.

@@          Coverage Diff          @@
##           master   #466   +/-   ##
=====================================
  Coverage     100%   100%           
=====================================
  Files          14     14           
  Lines        3189   3195    +6     
  Branches      229    230    +1     
=====================================
+ Hits         3189   3195    +6

Impacted Files	Coverage Δ
arrow/arrow.py	100% <100%> (ø)	⬆️
tests/arrow_tests.py	100% <100%> (ø)	⬆️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update b33925e...ad63085. Read the comment docs.

[richard-engineering]

Imho it would be better the last chunk of time be in a separate interval rather than appended to the last time's end date. That way you are guaranteed the maximum time span between any two dates is the specified interval rather than the last one potentially being higher. That seems to be how you implemented things (different from the original post)

[richard-engineering]

can also fix some pep8 issues along the way with this PR ;)

    if interval < 1:
        raise ValueError("interval has to be a positive integer")
    range_span = arrow.Arrow.span_range(frame, start, end, tz)
    bound = (len(range_span) // interval) * interval
    _range = [(range_span[i][0], range_span[i + interval - 1][1]) for i in range(0, bound, interval)]
    if bound < len(range_span):
        _range.append((range_span[bound][0], range_span[-1][1]))
    return _range

[dwrpayne]

I'm confused. I wrote it so that if, for example, you ask for intervals of 2 hours over a 5 hour range, you are returned a list of length 3. The first two elements have a 2 hour interval and the 3rd has the extra hour.

The previous implementation would have dropped the 5th hour.

[dwrpayne]

@rhan-mentad I think you misunderstood. I'm not appending the extra chunk of time to the last pair, I'm appending it to the return list.

Interval() returns a list of (start,end) tuples. I'm just appending a final (start,end) tuple in cases where there is a bit of time in the range left over.