d52a7b6d8e
DOC: Added doc string for instant_fill.
2013-12-04 07:16:18 -05:00
5f4d8817e8
ENH: Add len() functionality to BarData.
2013-12-03 16:29:17 -05:00
bbad5b386a
MAINT: Removed unused test_utils.check functions.
...
The test_utils were the only references to the blist package,
removing so that blist can be removed as a dependency.
2013-12-02 13:19:00 -05:00
860a340e55
BUG: Erroneous call to log.
2013-11-25 19:52:05 -05:00
37ef8f6a93
STY: Removed unused import.
2013-11-25 18:07:36 -05:00
c199a0d956
STY: Move order_value from blotter to TradingAlgorithm.
2013-11-25 17:38:30 -05:00
579cb56663
STY: Long line break for pep8.
2013-11-25 12:16:41 -05:00
9cb9831c08
STY: Prepend order_ to all target methods.
2013-11-25 11:28:27 -05:00
571e07f89c
MAINT: Add data_frequency to SimulationParameters repr.
2013-11-22 14:10:27 -05:00
a07c94665b
BUG: Need to set portfolio_needs_update on the algorithm
...
not on the simulator
2013-11-19 16:56:09 -05:00
cfe6296a3e
Merge lazy loading of portfolio and related values.
2013-11-19 14:42:36 -05:00
ccb7f493f7
PERF: Only update the portfolio once per dt.
...
So that each reference to `.portfolio` in the algoscript,
cache the value of the portfolio, and mark the need for a new
value at the end of each dt in the tradesimulation loop.
2013-11-19 14:39:19 -05:00
6d46eb71ea
PERF: moved performance calculation out of inner loop
...
lazy loading for portfolio
less repeating in performance period updates
2013-11-19 10:39:57 -05:00
50800a9863
BUG: Fix data cache filepath on Windows.
...
Prevent the ':' char, generated by converting a datetime to a string,
from creating on incompatible filepath for Windows.
2013-11-18 20:37:45 -05:00
e9dbffbbaa
MAINT: Upgrade Logbook to 0.6.0.
2013-11-18 16:16:51 -05:00
1f1be35734
ENH: Added commission model PerDollar
...
Commissions will be calculated based on total dollar traded.
2013-11-15 11:49:58 -05:00
395ce67c53
PERF: using comparison operation instead of tolerant_equals
2013-11-13 21:08:13 -05:00
923d3e4cbc
PERF: re-ordered if statements, made into exclusive checks
2013-11-13 20:47:49 -05:00
9b22b86703
BLD: Peg pytz to latest.
...
2013b -> 2013.8
2013-11-13 10:37:41 -05:00
4a317762d1
BLD: Peg tornado to latest version.
...
3.1 -> 3.1.1
2013-11-13 10:35:20 -05:00
f33193069c
PERF: Remove in operator from inner loop.
...
Ask forgiveness via exception handling instead.
2013-11-12 22:32:30 -05:00
f8755a5602
STY: Remove unused imports.
2013-11-12 13:26:08 -05:00
b555af157b
PERF: Use stored values for open and close.
...
The market_open_and_close method was a performance bottleneck,
since it was creating new dates on each query for open and close.
Instead use the pre-rendered frame of open and closes values
from the trading environment.
2013-11-12 13:12:12 -05:00
a2a56f7c63
MAINT: Remove noop_environment.
...
The code that was consuming noop_environment now uses a
real trading environment.
As more behavior relies on an accurate trading calendar, maintaining
the noop environment was a constraint that was more overhead than it
is worth.
2013-11-12 13:11:52 -05:00
102cfcbe5b
BUG: Fix bad dates from test factory.
...
Passing the exchange time timestamp to is_market_hours was ending
up with odd behavior due to conversion back to UTC when checking
the is_trading_day boolean.
2013-11-12 12:45:05 -05:00
20e0ad191b
TST: Disable test covering London stock exchange.
...
Changes to trading calendar and environments for supporting market
minutes, etc. have made the non-NYSE stock exchange support lag.
Disabling the test, with the intent of bringing support back up to
parity with NYSE.
2013-11-12 12:12:35 -05:00
aff8311f2f
MAINT: Change name of column in open and close frame.
...
Use more natural sounding 'market_open' and 'market_close'.
2013-11-11 16:05:10 -05:00
48be898a13
ENH: Add open and closes to trading calendar.
...
Use the early closes to populate a DataFrame which includes
the open and close minute for each day.
To be used by the environment instead of calculating each value
mid-backtest.
2013-11-11 15:48:44 -05:00
43b85cffb0
MAINT: Calculate tradingcalendar with days beyond the current day.
...
To make 'next open' calculations more straight ahead, calculate more
than enough days in the trading calendar.
2013-11-11 15:48:44 -05:00
797cb8ece3
BUG: Fix bad reference to benchmark timezone in loader.
2013-11-11 14:39:11 -05:00
89793e371c
MAINT: Protect loader against Series saved with no tz.
...
Checking for tz.UTC is not sufficient, since it is possible for
the index.tz value to be None.
2013-11-11 14:17:14 -05:00
408c129760
BLD: Remove Delorean as a build and setup dependency.
...
Now that we use pandas Timestamp throughout, Delorean is no longer
needed as a dependency.
2013-11-11 11:47:41 -05:00
796b9fb67a
MAINT: Use Timestamp for calculating next_trading_dt in test factory.
...
Instead of Delorean, use pandas Timestamp.
Could also consider using environment trading_days directly.
2013-11-11 11:44:40 -05:00
4d11ca131e
TST: Use Timestamp instead of Delorean for trading calendar test.
...
Removes another use of Delorean in favor of Timestamp 'today'.
2013-11-11 10:38:04 -05:00
dd95b6bfa3
MAINT: Use Timestamp instead of Delorean for tradingcalendar end.
...
Instead of Delorean, use pandas `today` behavior for Timestamp.
2013-11-11 10:13:05 -05:00
8014d9d938
BUG: Order.status now returns correct value
...
whether or not Order.open has been queried
2013-11-06 14:31:49 -05:00
4dbdf45006
MAINT: Factored process_transactions out of blotter's process_trade
2013-11-06 14:27:19 -05:00
3f89904e33
MAINT: Remove unused calculations of max_leverage, et al.
...
In the performance period the max_leverage, max_capital_used,
cumulative_capital_used were calculated but not used.
At least one of those calculations, max_leverage, was causing a
divide by zero error.
Instead of papering over that error, the entire calculation was
a bit suspect so removing, with possibility of adding it back in
later with handling the case (or raising appropriate errors) when
the algorithm has little cash on hand.
2013-11-06 14:22:20 -05:00
2492feb938
ENH: Keep track of total commissions as attribute on Order
...
Value is summed from TRANSACTION and COMMISSION events.
Defaults to None, meaning unset.
2013-11-01 17:23:50 -04:00
4b023a852a
MAINT: Use 1 as the first bitmask value instead of 0.
...
To prevent a possible future bug where values like SELL | STOP | LIMIT
would accidentally match STOP | LIMIT.
2013-11-01 16:26:48 -04:00
1575867b40
STY: Use named args for Transaction object creation.
...
Instead of creating and passing a dict of the object values,
use named args directly.
2013-11-01 16:10:19 -04:00
a192ba01a2
STY: Used named args in hardcoded test objects.
...
Instead of unrolling a dictionary, use named args.
2013-11-01 15:28:35 -04:00
73faf9133e
MAINT: Clean up imports of zipline.finance.trading
...
Use "from zipline.finance import trading" instead of "import
zipline.finance.trading as trading".
2013-10-29 13:50:14 -04:00
1794e4e716
DEV: Update flake8 to latest release.
2013-10-29 13:40:28 -04:00
f0465c5b87
STY: Tweak over-indented code.
...
flake8 recommendation
2013-10-29 13:38:31 -04:00
c45c1a22e1
BUG: Only localize benchmark index if it is naive.
...
Check for whether or not the index's timezone is UTC or not before
attempting to localize, since an already localized index throws an
error when tz_localize is called.
2013-10-29 13:17:58 -04:00
0a7539b6de
MAINT: flake8
2013-10-29 12:02:51 -04:00
c3ddbc258d
ENH: Accept simulation parameters and benchmark source args to create_test_zipline
2013-10-29 11:39:07 -04:00
e36f70f541
ENH: Allow market data loader to be specified to create_simulation_parameters
2013-10-29 11:17:41 -04:00
2d64ab8bfe
BUG: Fix naive timestamps in benchmarks.
...
Always convert the benchmarks to UTC, not just on reload.
2013-10-29 08:36:53 -04:00
Thomas Wiecki