def execExpression(self,data,date):
"""
date 指买入日期
>> accuracy.execExpression(data = {u'200056': {'statistics': {u'\u4e00\u65e5\u4e0a\u6da81%': 1.0,u'\u4e09\u5341\u65e5\u4e0a\u6da830%': 1.0,u'\u5341\u4e94\u65e5\u4e0a\u6da815%': 1.0000000000000002},'strategic': [u'\u77ed\u65f6\u95f4\u5185\u4e0b\u8dcc19%\u4ee5\u4e0a\u540e\u653e\u91cf\u6536\u7ea2']}}, date = '20110804')
{u'\u5341\u4e94\u65e5\u4e0a\u6da815%': {'matchCode': [], 'notMatchCode': [u'200056']}, u'\u4e00\u65e5\u4e0a\u6da81%': {'matchCode': [], 'notMatchCode': [u'200056']}}
"""
execResult = {}
result = {}
today = Date.getDate()
for code in data:
for statisticsName in data[code].get('statistics',{}):
expression = self.expression.getExpressionByName(name = statisticsName,expressions = self.expressions)
if not expression:
continue
if not code in execResult:
execResult[code]={}
offsetDate = Date.getDate(0-expression["offset"],date)
if int(offsetDate) > int(today):
continue
context = {
"result":execResult,
"code":code,
"name":statisticsName,
"stock":self.stock,
"date":offsetDate
}
self.expression.execExpression(expression = expression.get("content"), context = context)
if execResult[code][statisticsName] == True:
sub = MATCH_KEY
else:
sub = NOTMATCH_KEY
if not statisticsName in result:
result[statisticsName]={MATCH_KEY:[],NOTMATCH_KEY:[]}
result[statisticsName][sub].append(code)
return result
def __init__(self,conf={}):
self.conf = conf
self.dataHandler = Data(conf)
self.allCode = self.dataHandler.get("allCode")
self.date = conf.get('date',Date.getDate()) #抓取指定日期的detail
self.sourceName=conf.get('SOURCE_NAME')
self.threadNum = int(conf.get('THREAD_NUM',THREAD_NUM)) # 采用多线程模式抓取数据时的线程数
def variance(self,stock,date = None):
"""
方差公式:
s^2 = [(x1-x)^2+(x2-x)^2+......(xn-x)^2]/n
>> statistics.variance(stock = stock,date = '20111024')
0.0002890704988633324
"""
ex = self.expection(stock = stock,date = date)
total = 0
count = 0
if date == None:
date = [Date.getDate()]
elif date == "all":
date = stock.allDate
elif type(date) != list:
date = [date]
for date in date:
for code in stock.allCode:
change = stock[code][date].change
if type(change) == float:
total +=(change - ex)**2
count +=1
return total/count
def statistics(self,subject,stock,date = None):
"""
>> subject = [0.03,0.04]
>> statistics.statistics(subject = subject,stock = stock,date = '20111024')
0.19164835164835164
"""
if type(subject) == list and len(subject) == 2:
pass
else:
raise ValueError
if date == None:
date = [Date.getDate()]
elif date == "all":
date = stock.allDate
elif type(date) != list:
date = [date]
total = 0
count = 0
for oneDate in date:
for code in stock[oneDate].allCode:
change = stock[code].change
if type(change)!=float :
continue
minItem = subject[0]
maxItem = subject[1]
if float(minItem) <= change <float(maxItem): #Do note that left is <=
count += 1
total += 1
if not total:
p = 0
else:
p = float(count)/total
return p
def __init__(self,conf={}):
self.conf = conf
self.stock = Stock(self.conf)
self.date = conf.get('date',Date.getDate()) #抓取指定日期的detail
self.sourceName=conf.get('DETAIL_SOURCE_NAME') or DETAIL_SOURCE_NAME
self.threadMode = conf.get('THREAD_MODE',THREAD_MODE) # Detail的线程模式 single|mutil
self.threadNum = int(conf.get('THREAD_NUM',THREAD_NUM)) # 采用多线程模式抓取数据时的线程数
self.parser=Parser.ParserFactory(self.sourceName)
def __init__(self,conf={}):
self.date=conf.get("date",Date.getDate())
self.stock = conf.get("stock",Stock(conf=conf))
self.expression=Expression(conf)
self.expressions = self.expression.load(\
path = os.path.join(conf.get("SYS_HOME"),conf.get("STATISTICS_PROJECT_PATH",STATISTICS_PROJECT_PATH)))
self.report = Report(conf=conf)
self.data = {}
self.conf=conf
def collect(self,date=None):
"""
收集需要被统计的数据
>> accuracy.collect(date = "20110825")
"""
conf = self.conf
date = date or self.date or conf.get("date")
detail={}
dateRange = self.conf.get("STATISTICS_DATE_RANGE",STATISTICS_DATE_RANGE)
for i in range(0,dateRange):
for expression in self.expressions:
curDate = Date.getDate(0-i,date)
offsetDate = Date.getDate(expression['offset'],curDate)
trade = self.report.getTrade(date = offsetDate)
execDetail = self.execExpression(data = trade,date = offsetDate)
if not execDetail:
continue
detail.update({offsetDate:execDetail}) #将前一天统计数据加上今日统计数据便得到要生成的统计数据
return detail
def run(self):
date = self.conf.get("date",Date.getDate())
queue = self.conf.get('queue')
handle = self.conf.get('handle')
if queue:
while True:
code=queue.get()
try:
handle(date = date ,code = code)
queue.task_done()
except:
logging.error('Error hanpped when download data.')
queue.task_done()
def collect(self,projectPath=None):
"""
>> projectPath = os.path.join(conf.get("SYS_HOME"),case["collect"]['projectPath'])
>> train.collect(projectPath)
>> type(train.collect(projectPath)) == dict
True
"""
conf = self.conf
data={}
projectPath = projectPath or os.path.join(conf.get("SYS_HOME"),conf.get("STATISTICS_PROJECT_PATH"))
expressions = self.expression.load(path=projectPath)
for i in range(0,conf.get("MIN_DATA_NUM",MIN_DATA_NUM)):
date=Date.getDate(0-i,conf.get('date',Date.getDate()))
result={}
for code in self.stock.allCode:
result[code]={}
self.expression.execute(expression= conf.get("expressions",expressions), context = {"result":result,"date":date})
if not result:
break;
data[date]=result
return data
def expection(self,stock,date = None):
"""
期望值公式 : expection = (x1 + x2 + x3 +... +xn)/n
>> statistics.expection(stock = stock,date = '20111024')
0.02870564489608117
"""
sumValue = 0
count = 0
if date == None:
date = [Date.getDate()]
elif date == "all":
date = stock.allDate
elif type(date) != list:
date = [date]
for oneDate in date:
for code in stock[oneDate].allCode:
change = stock[code].change
if type(change) == float:
sumValue = change + sumValue
count +=1
return sumValue / count
def __init__(self,conf={}):
self.conf=conf
self.stock=conf.get("stock",Stock(conf))
self.date=conf.get("date",Date.getDate())
