Changes the overlap behavior so that it is an error to write data which
would have two companies holding the same ticker. Other than one test
around which company would win in that case, all the other tests are
passing. That single test has been changed to check the write-time
error.
- Added test coverage for grouped and masked top/bottom.
- Added test coverage for grouped rank on datetime factors.
- Fixed an issue where grouped rank would fail on datetime inputs
because unary-negative isn't defined for datetimes. We now instead
directly invoke a function from rank.pyx that does the normalizations
as neeeded.
- Fixed an issue where GroupedRowTransform assumed that it produced the
same dtype as its input. This isn't true for rank() of a
datetime-dtype factor. GroupedRowTransform now takes a required dtype
parameter.
- Similarly, fixed an issue where GroupedRowTransform assumed that its
missing_value was the same as its parent's, which isn't true for
rank() of a datetime-dtype factor. GroupedRowTransform now takes a
required dtype parameter.
- Fixed an issue where Factor.demean() and Factor.zscore() weren't
properly cached because their static_identity included a closure that
was dynamically generated on each invocation. They both now always
use a function defined at module scope.
* BUG: Further corrections for days_at_time
- Revert to using DateOffset, as Timedelta doesn't handle offsetting by
one day over a tz change properly:
In [12]: pd.Timestamp('2004-04-05', tz='America/Chicago') + pd.Timedelta(days=-1)
Out[12]: Timestamp('2004-04-03 23:00:00-0600', tz='America/Chicago')
In [13]: pd.Timestamp('2004-04-05', tz='America/Chicago') + pd.DateOffset(days=-1)
Out[13]: Timestamp('2004-04-04 00:00:00-0600', tz='America/Chicago')
By creating a DateOffset using the `days` kwarg, the issue previously
fixed in bcc867b is addressed.
- To preempt any other pandas issues around day offsets, changes to
performing these with no timezone, then localizing to the local
timezone when shifting the time.
- Adds unit test for days_at_time
* STY: Remove unused import
is backwards-compatible with the previous format.
In USEquityLoader, use dailyreader's trading_calendar.
This is backwards compatible and will fall back to the NYSE calendar if
the reader doesn’t have a calendar specified.
Instead of having separate ExchangeCalendar and TradingSchedule objects, we
now just have TradingCalendar. The TradingCalendar keeps track of each
session (defined as a contiguous set of minutes between an open and a close).
It's also responsible for handling the grouping logic of any given minute
to its containing session, or the next/previous session if it's not a market
minute for the given calendar.
Limit the date range to one day, instead of a year.
On my local machine the following goes from taking 12 seconds to 0.6 seconds.
`$ nosetests -x tests/test_algorithm.py:TestMiscellaneousAPI`
In preparation of adding futures, add equity to the names of both the
classes and methods for writing bcolz data. Futures data will use a
different minutes per day with a separate reader. This change will allow
both equity and futures fixtures to be side by side.
Also, break out the method which generates the dataframes and trading
days member into fixtures (`EquityMinuteBarData` and
`EquityDailyBarData`) on which the `*BarReader` fixture depends. This
fixture is separated out to enable reader/writers in different formats
to use the same data setup. (There is internal code which needs to write
minute and daily bar data in a database format.)
Nathan Wolfe