Rather than specifying only the package "zipline" in setup.py, use
`find_packages` to find all the subpackages as well, so they (or, most
specifically, their `__init__.py` files) are properly packaged in the
egg file.
Modify setup.py to defer the use of Cython and numpy until
`setup_requires` has already been processed, so that Cython and numpy
are available when they are needed.
- Parse our requirements.txt to keep install_requires up to date.
- Create extras builds for talib and dev.
- Use pip install -e .[dev] on Travis to install Zipline before testing.
If lookup_future_chain was provided with an as_of_date or knowledge date that was pandas.NaT, the query we were forming wasn't what we want. Instead, as_of_date, if not NaT, is used for knowledge_date, and if both are NaT, no date filtering is done in the query.
`period_end` can be outside the range of data for which we have dates.
`last_close` properly gets pulled back to the last date for which we
actually have data.
We should consider whether or not we need to be storing period_end at
all.
- Previously it was returning a DataFrame because of how we applied an &
with a DataFrame mask. The error was masked by the fact that
`np.assert_array_equal` coerces inputs to arrays before comparing.
- Added `zp.utils.test_utils.check_arrays`, which checks type equality
before calling `np.assert_array_equal`.
This patch lays the groundwork for a compute engine designed to
facilitate construction of factor-based universe screening and portfolio
allocation. It contains:
A new module, `zipline.modelling`, containing entities that can be used
to express computations as dependency graphs. Each node in such a graph
is an instance of the base `Term` class, defined in
`zipline.modelling.term`. Dependency graphs are executed by instances
of `FFCEngine`, defined in `zipline.modelling.engine`.
A new module, `zipline.data.ffc`, containing loaders and dataset
definitions for inputs to the modelling API.
New `TradingAlgorithm` api methods: `add_factor`, and `add_filter`.
These methods can only be called from `initialize`, and are used to
inform the algorithm that each day it should compute the given terms.
Computed factor results are made available through a new attribute of
the `data` object in `before_trading_start` and `handle_data`. Computed
filter results control which assets are available in the factor matrix
on each day.
As of logbook 0.10.0, logbook no longer installs a default handler,
which means that if the application doesn't install one, log messages
disappear into the ether.
Therefore, all of our scripts with `__main__` endpoints need to push a
`logbook.StderrHandler` if they're not already pushing some other
handler.
Instead of creating a new cursor with each query, use the same cursor
throughout the lifetime of the finder instance to remove any overhead
from creating a new cursor in tight loops.
Scott Sanderson