Merge branch 'master' of https://github.com/kanghua309/strategy

# Conflicts:
#	campaign/deeplearn/training.py
#	me/helper/research_env.py

campaign/checkmode/factors.py

@@ -36,7 +36,7 @@
 
 
 start = '2015-8-5'  # 必须在国内交易日
-end   = '2015-9-30'  # 必须在国内交易日
+end   = '2015-9-28'  # 必须在国内交易日
 
 c,_ = get_sector_class()
 ONEHOTCLASS = tuple(c)
@@ -53,31 +53,35 @@ def compute(self, today, assets, out, close,volume):
 
 def make_pipeline(asset_finder):
     universe = make_china_equity_universe(
-        target_size=2000,
-        #mask=default_china_equity_universe_mask([]),
+        target_size=300,
+        #mask=default_china_equity_universe_mask(['000001']),
         mask=None,
         max_group_weight=0.01,
-        smoothing_func=lambda f: f.downsample('month_start'),
-
+        smoothing_func=lambda f: f.downsample('week_start'),
+        asset_finder = asset_finder,
     )
 
-    #private_universe = private_universe_mask(['000001','000002','000005'],asset_finder=asset_finder)
+
+
+    #private_universe = private_universe_mask(['000001'],asset_finder=asset_finder)
     private_universe = universe
+    print ("---universe:",private_universe)
+    #print private_universe_mask(['000001','000002','000005'],asset_finder=asset_finder)
     ######################################################################################################
     returns = Returns(inputs=[USEquityPricing.close], window_length=5)  # 预测一周数据
     ######################################################################################################
-    ep = 1/Fundamental(asset_finder).pe
-    bp = 1/Fundamental(asset_finder).pb
-    bvps = Fundamental(asset_finder).bvps
-    market = Fundamental(asset_finder).outstanding
+    ep = 1/Fundamental(mask = private_universe,asset_finder=asset_finder).pe
+    bp = 1/Fundamental(mask = private_universe,asset_finder=asset_finder).pb
+    bvps = Fundamental(mask = private_universe,asset_finder=asset_finder).bvps
+    market = Fundamental(mask = private_universe,asset_finder=asset_finder).outstanding
 
 
     rev20 = Returns(inputs=[USEquityPricing.close], window_length=20,mask = private_universe)
     vol20 = AverageDollarVolume(window_length=20,mask = private_universe)
 
-    illiq = ILLIQ(window_length=20,mask = private_universe)
-    rsi = RSI(window_length=20,mask = private_universe)
-    mom = Momentum()
+    illiq = ILLIQ(window_length=22,mask = private_universe)
+    rsi = RSI(window_length=22,mask = private_universe)
+    mom = Momentum(window_length=252,mask = private_universe)
 
     sector = get_sector(asset_finder=asset_finder,mask=private_universe)
     ONEHOTCLASS,sector_indict_keys = get_sector_by_onehot(asset_finder=asset_finder,mask=private_universe)
@@ -112,13 +116,14 @@ def make_pipeline(asset_finder):
 
 
 
-pd.set_option('display.width', 800)
+pd.set_option('display.width', 8000)
 research = Research()
 #print(research.get_engine()._finder)
 my_pipe = make_pipeline(research.get_engine()._finder)
 result = research.run_pipeline(my_pipe,
                                Date(tz='utc', as_timestamp=True).parser(start),
                                Date(tz='utc', as_timestamp=True).parser(end))
+
 print result
 print type(result)
 print result.reset_index()

me/pipeline/classifiers/tushare/sector.py

@@ -84,7 +84,6 @@ def get_sector(sector_dict=None,mask = None,asset_finder = None):
         sector_dict,_ = get_sector_class()
     #print("++enter getSector++",len(sector_dict))
     basic=load_tushare_df("basic")
-    #print(asset_finder)
     def _sid(sid):
         return asset_finder.retrieve_asset(sid)
     class Sector(CustomClassifier):  #CustomClassifier 是int , factor 是float
@@ -114,7 +113,9 @@ def findSector(self,assets):
         def compute(self, today, assets, out, *inputs):
             rs = self.findSector(assets)
             out[:] = rs
-    return Sector(mask=mask)
+    if mask != None:
+        return Sector(mask=mask)
+    return Sector()
 
 def get_sector_by_onehot(sector_dict=None,mask = None,asset_finder = None):
     if sector_dict is None:
@@ -183,7 +184,10 @@ def compute(self,today, assets, out):
                     i += 1
                 idx += 1
             #print out
-    return OneHotSector(mask=mask),sector_indict_keys
+    if mask != None:
+        return OneHotSector(mask=mask),sector_indict_keys
+    else:
+        return OneHotSector(),sector_indict_keys
 
 
 

me/pipeline/factors/tsfactor.py

@@ -43,7 +43,7 @@ def compute(self, today, assets, out, close):
     return MarketCap()
 '''
 # can be used outside algo scope - give your asset_finder
-def Fundamental(asset_finder = None):
+def Fundamental(mask = None,asset_finder = None):
     def _sid(sid):
         return asset_finder.retrieve_asset(sid)
     columns = ['pe',  # 市盈率
@@ -101,7 +101,8 @@ def compute(self, today, assets, out,close):
             out.gpr[:] = df.gpr
             out.npr[:] = df.npr
             out.holders[:] = df.holders
-
+    if mask != None:
+        return Fundamental(mask = mask)
     return Fundamental()
 
 '''

me/pipeline/filters/universe.py

@@ -27,7 +27,7 @@
 ADV_ADJ_DOWNLIMIT = 1.0e+8
 
 
-def universe_filter(smoothing_func = None):
+def universe_filter(smoothing_func = None,asset_finder=None):
     """
     Create a Pipeline producing Filters implementing common acceptance criteria.
     Returns
@@ -36,9 +36,9 @@ def universe_filter(smoothing_func = None):
         Filter to control tradeablility
     """
     factors = {
-     'MarketCap': Fundamental().outstanding,
+     'MarketCap': Fundamental(asset_finder=asset_finder).outstanding,
      'ADV_adj'  : ADV_adj(),
-     'Sector'   : get_sector(),
+     'Sector'   : get_sector(asset_finder=asset_finder),
     }
 
 
@@ -69,7 +69,7 @@ def universe_filter(smoothing_func = None):
 
     return filters
 
-def sector_filter(tradeable_count,sector_exposure_limit,smoothing_func = None):
+def sector_filter(tradeable_count,sector_exposure_limit,smoothing_func = None,asset_finder=None):
 
     """
     Mask for Pipeline in create_tradeable. Limits each sector so as not to be over-exposed
@@ -88,7 +88,7 @@ def sector_filter(tradeable_count,sector_exposure_limit,smoothing_func = None):
 
     industry_class,_ = get_sector_class()
     #print("g_inds",g_inds)
-    sector_factor = get_sector(industry_class)
+    sector_factor = get_sector(industry_class,asset_finder=asset_finder)
     # set thresholds
     sector_size = len(industry_class)
     if sector_exposure_limit < ((1. / sector_size)):
@@ -178,7 +178,7 @@ def sector_filter(tradeable_count,sector_exposure_limit,smoothing_func = None):
     return filters
 
 
-def default_china_equity_universe_mask(unmask):
+def default_china_equity_universe_mask(unmask,asset_finder=None):
     #a_stocks = []
     info = load_tushare_df("basic")
     sme =  load_tushare_df("sme")
@@ -192,13 +192,18 @@ def default_china_equity_universe_mask(unmask):
         pass
     #Returns a factor indicating membership (=1) in the given iterable of securities
     #print("==enter IsInSymbolsList==")
+    def _sid(sid):
+        return asset_finder.retrieve_asset(sid)
     class IsInDefaultChinaUniverse(CustomFilter):
         inputs = [];
         window_length = 1
         def compute(self, today, asset_ids, out, *inputs):
             #print asset_ids
             #print maskset
-            assets  = [sid(id).symbol for id in asset_ids]
+            if asset_finder != None:
+                assets  = [_sid(id).symbol for id in asset_ids]
+            else:
+                assets  = [sid(id).symbol for id in asset_ids]
             #print "--------------"
             #print pd.Series(assets)
             out[:] = pd.Series(assets).isin(maskdf.index)
@@ -207,7 +212,7 @@ def compute(self, today, asset_ids, out, *inputs):
 
 def private_universe_mask(mask,asset_finder = None):
     mask = mask
-    def sid(sid):
+    def _sid(sid):
         return asset_finder.retrieve_asset(sid)
     class IsInPrivateUniverse(CustomFilter):
         inputs = [];
@@ -218,7 +223,10 @@ def compute(self, today, asset_ids, out, *inputs):
             # for id in asset_ids:
             #     print id
             #     print type(sid(id))
-            assets  = [sid(id).symbol for id in asset_ids]
+            if asset_finder != None:
+                assets  = [_sid(id).symbol for id in asset_ids]
+            else:
+                assets  = [sid(id).symbol for id in asset_ids]
             #print "--------------"
             #print pd.Series(assets)
             out[:] = pd.Series(assets).isin(mask)
@@ -232,13 +240,12 @@ def make_china_equity_universe(
         mask,
         #groupby,
         max_group_weight,
-        smoothing_func):
-    ufilters = universe_filter(smoothing_func)
-    sfilters = sector_filter(target_size,max_group_weight,smoothing_func)
-
+        smoothing_func,
+        asset_finder = None):
+    ufilters = universe_filter(smoothing_func,asset_finder)
+    sfilters = sector_filter(target_size,max_group_weight,smoothing_func,asset_finder)
     if mask != None:
         return (ufilters & sfilters) & mask # &? TODO
-
     return  (ufilters & sfilters)
 
 

simulate_research_env_example/run_algo_example.py

@@ -16,6 +16,8 @@
 from zipline.utils.calendars import get_calendar
 from zipline.utils.factory import create_simulation_parameters
 from zipline.utils.cli import Date, Timestamp
+from me.pipeline.filters.universe import make_china_equity_universe, default_china_equity_universe_mask, \
+    private_universe_mask
 from zipline.api import (
     attach_pipeline,
     date_rules,
@@ -38,6 +40,13 @@
 
 N = 10
 def make_pipeline():
+    # universe = make_china_equity_universe(
+    #     target_size = 100,
+    #     mask = default_china_equity_universe_mask(['000001']),
+    #     max_group_weight= 0.01,
+    #     smoothing_func = lambda f: f.downsample('week_start'),
+    # )
+
     dollar_volume = AverageDollarVolume(window_length=1)
     high_dollar_volume = dollar_volume.percentile_between(N, 100)
     recent_returns = Returns(window_length=N, mask=high_dollar_volume)