INT: mixin for PeriodIndex / Datetimeindex for common methods

[jreback]

related #6779

• asobject (BUG: DatetimeIndex.asobject raises ValueError when contains NaT #7544)
• __unicode__ (API: DatetimeIndex and PeriodIndex have same representation #7602)
• min/max (BUG: PeriodIndex.min/max returns int #7609)
• summary (API: Datetime-like indexes summary should output the same format #9116)
• equals (BUG/CLN: datetimelike Index.equals may return True with non-Index #13986)
• isin
• unique (BUG/API: PeriodIndex.unique returns ndarray #7540, BUG: PeriodIndex.unique results in Int64Index #7843)
• value_counts (BUG: DTI.value_counts doesnt preserve tz #7735)
• resolution (API: Add PeriodIndex.resolution #7708)
• __iter__ (BUG: DatetimeIndex.__iter__ creates a temp array of Timestamp (GH7683) #7702, BUG: DatetimeIndex.__iter__ creates a temp array of Timestamp (GH7683) #7709, PERF: improve perf of index iteration (GH7683) #7720)
• __add__, __iadd__, __sub__, __isub__ (ENH/BUG: DatetimeIndex and PeriodIndex in-place ops behaves incorrectly #7741)
• __getitem__ (BUG: TimedeltaIndex slicing may reset freq #10292, for DatetimeIndex and TimedeltaIndex) (CLN: move PeriodIndex.__getitem__ to DatetimeIndexOpsMixin #13987, for PeriodIndex)
• repeat (CLN: Move PeriodIndex.repeat to DatetimeIndexOpsMixin #13995)
• NaT related methods / property. hasnans, _isnan, _na_value (CLN: move _na_value to DatetimeIndexOpsMixin #13997)

next priorities:

• get_duplicates
• repeat
• get_value / get_value_maybe_box
• searchsorted
• argmin / argmax

[jreback]

cc @sinhrks you might find this interesting

[sinhrks]

Seems to be important.

• list was moved to top

[jreback]

@sinhrks comment here and I'll check/update boxes up top

[sinhrks]

Added __iter__