Scott Sanderson
705f87215b
DOC: Update engine docstrings.
2016-03-19 17:04:28 -04:00
dmichalowicz
990c76e9fb
BUG: Empty pipeline failed to tz_localize
2016-03-15 14:19:31 -04:00
Scott Sanderson
535d05e714
MAINT: Remove notion of "atomic" pipeline terms.
...
Replace it by distinguishing between "Loadable" and "Computable".
This is useful because it's now possible to write computable terms that
don't require any inputs (e.g. an `Always` filter or an `Everything`
classifier).
2016-03-08 13:49:45 -05:00
Scott Sanderson
f440cb73b2
MAINT: Make __slots__ a tuple.
2016-02-12 21:21:19 -05:00
Scott Sanderson
3f28a924d7
DOC: Add missing period in docstring.
2015-11-25 15:26:37 -05:00
Scott Sanderson
cf7e87a4b9
DOC: Typo in docstring.
2015-11-20 20:02:56 -05:00
Scott Sanderson
8cd4f7d100
MAINT: Make load_adjusted_array return a dict.
...
Rather than a list that's ordered the same as the received columns.
Most nontrivial loaders were constructing dicts internally and then
converting back to lists, only to have the engine convert **back again**
into a dict. This cuts out the middleman, and prevents bugs due to
incorrect ordering of the output arrays.
2015-11-03 11:16:21 -05:00
John Ricklefs
f599795d27
ENH: Allow pipelines to run with matching start/end dates
2015-10-22 14:23:00 -04:00
llllllllll
fca637a40b
ENH: Adds a better error if root mask is empty
2015-10-19 16:35:02 -04:00
llllllllll
8f661387df
ENH: allows any iterator to be the return of load_adjusted_arrays
2015-10-19 16:35:02 -04:00
Richard Frank
ee26a21855
MAINT: Renamed loader_dispatch to get_loader
...
Now it raises a KeyError instead of returning None,
if loader not found.
2015-10-12 16:13:55 -04:00
Richard Frank
99de89c817
PERF: Don't recalc similar atomic terms
2015-10-12 16:13:55 -04:00
Richard Frank
2dabda6b76
MAINT: Reworked Term atomicity
2015-10-12 16:11:19 -04:00
Richard Frank
7bd6b69a89
MAINT: Group by loader and extra_rows
...
so that the mask and dates are the same for all the columns
the loader is loading at a time.
2015-10-12 10:48:29 -04:00
Richard Frank
83bd1310d9
PERF: Using pipeline_loader_dispatch to group by loader
...
instead of dataset
2015-10-12 10:48:29 -04:00
Richard Frank
e880fa3e34
PERF: Batch load atomic terms by dataset
...
Added CompositeTerm and now we dispatch more generally on atomic
2015-10-12 10:48:28 -04:00
Scott Sanderson
2d683961bd
MAINT: More renaming.
...
s/FFCEngine/PipelineEngine/
s/FFCLoader/PipelineLoader/
2015-10-01 18:03:54 -04:00
Scott Sanderson
f82a01841b
MAINT: Rename ALL the things.
...
zipline.modelling.* -> zipline.pipeline.*
zipline.data.ffc.loaders -> zipline.pipeline.loaders
tests/modelling -> tests/pipeline
2015-10-01 18:03:53 -04:00
Scott Sanderson