API/CLN: more common ops to integrate with Series/index OpsMixin

[jreback]

after #6380, ability to pretty easily move ops over to the mix-in

• date/time (added in API/CLN: add in common operations to Series/Index, refactored as a OpsMixin #6380)
• more periods ops? ENH: Period ops NaT & timedelta ops #5202
• additional field accessor (is_beg_year), API: add is_beg_month/quarter/year, is_end_month/quarter/year accessors (#4565) #4823
• validate argmin/max, (PR INT: followup to Index not a sub-class of ndarray #7904)
• StringMethods, e.g. str for Index (not for sub-classes though), ENH: add StringMethods (e.g. .str) to Index #9068

(this turns out to be quite tricky, as sometimes construction depends on the existence of a tz attribute)

• freq,frestr,inferred_freq
• is_monotonic ENH: Series has gained the properties .is_monotonic* #13336
• pd.infer_freq(), related infer_freq broken in 0.13.1 #6407,API: pd.infer_freq interacts correctly with a Series #6408
• is_unique,lexsort_depth,nlevels
• nunique(),unique(),value_counts(), ENH: added nunique function to Index #6734
• factorize() ENH/CLN: Add factorize to IndexOpsMixin #7090
• get_loc/find index where, first_index_where, first_value_where, first_valid_value #4616
• __invert__ API: Index should support __inverse__ ops #8875

These are related in that they are generic.py functions as they have to compute something then set the axis with the result.

• normalize ENH: Add Index.normalize as a Series or DataFrame method #5502
• tz_convert/tz_localize (though this requires a bit of wrapping as they have an axis parm,
  but should be possible), see also here: ENH: New level argument for DataFrame.tz_localize and DataFrame.tz_convert (GH7846) #7915 (currently implemented only on DatetimeIndex, PeriodIndex has these as not implemented (and delegating from NDFrame is in core/generic)
• duplicated/drop_duplicates Add duplicated method to Index classes #4060, (PR ENH: Add duplicated/drop_duplicates to Index #7979)
• has_duplicates BUG: Bug in MultiIndex.has_duplicates when having many levels causes an indexer overflow (GH9075) #9077

[jreback]

these are mostly done / shared already. @jbrockmendel if anything looks interesting can open a specific issue.