def to_R(self, ds, dates_as_factor, suppress_all_value=True, **kwargs):
if not self.colname:
return
col = ds[self.colname]
if col.is_datetimetype():
rdata = rconv.Missing_Date_to_R(col)
if not dates_as_factor:
return rdata
elif MA.isMaskedArray(col.data):
if self.missing_as_none:
rdata = MA.filled(col.data, sys.maxint)
else:
if Numeric.alltrue(col.data.mask()):
raise PlotError('%r: all data-points masked' % col.name)
rdata = rconv.MA_to_R(col.data)
else:
rdata = col.data
if col.is_discrete():
if 0 and suppress_all_value:
row_vecs = [
vec for value, vec in col.inverted.items()
if value != col.all_value
]
row_vec = union(*row_vecs)
rdata = MA.take(rdata, row_vec)
return self.discrete_col_to_R(col, rdata, suppress_all_value)
else:
return self.continuous_col_to_R(col, rdata)
def from_summset(cls, ds, shaped_like=None):
self = cls(ds.name)
st = time.time()
cols = ds.get_columns()
if shaped_like is not None:
for axis in xtab_axes(shaped_like):
try:
col = ds[axis.name]
except KeyError:
pass
else:
self.axes.append(CrossTabAxis.from_col(col, axis.values))
cols.remove(col)
for col in cols:
if col.is_discrete() and not col.name.startswith('_'):
self.axes.append(CrossTabAxis.from_col(col))
if not self.axes:
raise Error('dataset %r must have at least one discrete column' %
(ds.name,))
indices = [axis.indices.filled() for axis in self.axes]
masks = [axis.indices.mask() for axis in self.axes]
map = MA.transpose(MA.array(indices, mask=masks))
shape = self.get_shape()
for col in ds.get_columns():
if col.is_scalar():
self.add_table(col.name,
data=self.from_vector(map, col.data, shape),
label=col.label)
elapsed = time.time() - st
soom.info('%r crosstab generation took %.3f, %.1f rows/s' %
(self.name, elapsed, len(map) / elapsed))
return self
def testMultipleDimensionCreation (self):
'Test creation of multidimensional arrays.'
alist = [[2,3],[4,5]]
x = MA.array(alist)
assert x.shape == (2,2)
y = MA.array([x, x + 1, x + 2])
assert y.shape == (3,2,2)
def accumulate24Hourly(data):
"""Returns 12-hourly data accumulated to 24-hours."""
newTimeValues=[]
taxis=data.getTime()
tunits=data.units
print len(data.getTime())
newarray=[]
for i in range((tlen/2)):
p1=data(time=slice(i,i+1))
p2=data(time=slice(i+1,i+2))
accum=p1+p2
newarray.append(accum)
newTimeValues.append(p2.getTime()[0])
array=MA.concatenate(newarray)
array=MA.array(array, 'f', fill_value=data.getMissing())
axes=data.getAxisList()
newTimeAxis=cdms.createAxis(newTimeValues)
newTimeAxis.units=tunits
newTimeAxis.designateTime()
newTimeAxis.id=newTimeAxis.long_name=newTimeAxis.title="time"
newaxes=[newTimeAxis]+axes[1:]
var=cdms.createVariable(array, axes=newaxes, id=data.id)
for att in ("units", "long_name"):
setattr(var, att, getattr(data, att))
return var
def testConcatenate (self):
"test concatenate"
assert eq (MA.concatenate((self.a[:3], self.a[3:])), [0,1,2,3,4,5])
m = self.m
assert eq (MA.concatenate((m, m)), [[1,2,3], [11,12,13], [1,2,3], [11,12,13]])
assert eq (MA.concatenate((m, m), axis = 1),
[[1,2,3,1,2,3],[11,12,13,11,12,13]])
def testDiagonal (self):
"Test the diagonal function."
b=MA.array([[1,2,3,4],[5,6,7,8]]*2)
assert eq(MA.diagonal(b), [1,6,3,8])
assert eq(MA.diagonal(b, -1), [5,2,7])
c = MA.array([b,b])
assert eq(MA.diagonal(c,1), [[2,7,4], [2,7,4]])
def _subSampleByAveraging(self, var, timeVar, flagVar, samplingRate, flagsToUse):
"""
Returns a new variable which is 'var' sub-sampled by averaging
at the given samplingRate with data selected according to flagVar
including the flag values specified in flagsToUse (defaults are 0 and 1).
"""
maskedArray=self._getMaskedArray(var, flagVar, flagsToUse)
shape=var.shape
if shape[1]==1:
newNumArray=MV.ravel(maskedArray)
newArrayMask=MV.ravel(maskedArray.mask())
newArray=MA.masked_array(newNumArray, mask=newArrayMask, fill_value=maskedArray.fill_value())
else:
newArray=Numeric.zeros(shape[0], 'f')
for t0 in range(shape[0]):
# Set as missing if less than half are valid
t1Array=maskedArray[t0]
if samplingRate==1:
# If half or more are good values then calculate the mean
if t1Array.count()>=(shape[1]/2.):
newArray[t0]=MA.average(t1Array)
# otherwise set as missing value
else:
newArray[t0]=maskedArray.fill_value()
else:
raise "Averaging for non 1Hz sampling rates not yet supported!"
# Now re-construct variable axes etc
newTimeAxis=self._flatten2DTimeAxis(timeVar, samplingRate)
newVar=self._recreateVariable(var, newArray, newTimeAxis, flagVar, max(flagsToUse), missingValue=maskedArray.fill_value(), sampleBy="averaging")
return newVar
def Init_diag(Ain,Aout,fignum): global Ptot,RatioM,Tmincrit,Cpa,Cpv,Lv,namincrit,Hmincrit,Tmaxcrit,Tminopt global Ratiofig,fig_hauteur,fig_longueur,RatioAxes,lineprops,lineprops_HR,x1,y1,x3,y3 Ptot=101325.0 MO=15.9994 MN=14.0067 MAr=39.948 Mair=0.78084*2*MN+0.20946*2*MO+0.009340*MAr MH=1.00794 MH2O=2*MH+MO RatioM=MH2O/Mair Tmincrit=-30.0 Cpa=Cpas(C2K(Tmincrit)) Cpv=Cpvap(C2K(0.1)) Lv=ChLv(C2K(0.1)) namincrit=RatioM*Pvap(Tmincrit)/(Ptot-Pvap(Tmincrit)) Hmincrit=Cpa*Tmincrit+namincrit*(Lv+Cpv*Tmincrit) Tmaxcrit=260.0 Tminopt=-30.0 Ratiofig=21.0/29.7 fig_hauteur=6 #inches fig_longueur=fig_hauteur/Ratiofig figure(fignum,figsize=(fig_longueur,fig_hauteur)) lineprops = dict(linewidth=0.5, color='gray', linestyle='-',antialiased=True) lineprops_HR = dict(linewidth=1, color='blue', linestyle='-',antialiased=True) Tinit=20.0 Cpa=Cpas(C2K(Tinit)) Cpv=Cpvap(C2K(Tinit)) Lv=ChLv(C2K(Tinit)) maskin=Ma.getmask(Ain.car) maskout=Ma.getmask(Aout.car) # Il s'agit ici de fixer les valeurs moyennes de Cpa,Cpv et Lv # en faisant le calcul de l'air moyen (i.e. au centre du diagramme # le calcul s'arrête à 1e-6 de différence sur Cpa,Cpv et Lv crit=True while crit: Ain.car=Ma.array(Ain.car,copy=0,mask=maskin) Aout.car=Ma.array(Aout.car,copy=0,mask=maskout) Fill_Air(Ain) Fill_Air(Aout) Amoy=air() Amoy.definir(["H","na"],[mean([Ain.car[0],Aout.car[0]]), mean([Ain.car[1],Aout.car[1]])]) Fill_Air(Amoy) Tmoy=Amoy.car[2] Cpa_new=Cpas(C2K(Tmoy)) Cpv_new=Cpvap(C2K(Tmoy)) Lv_new=ChLv(C2K(Tmoy)) crit=((Cpa-Cpa_new)**2+(Cpv-Cpv_new)**2+(Lv-Lv_new)**2)>1e-6 Cpa=Cpa_new Cpv=Cpv_new Lv=Lv_new LH,Lna,LT,LHR,Lair=Return_levels_HnaTHR(Ain,Aout) x1,y1=Transf_xy(Lna[0],Lair[0].car[0]) x3,y3=Transf_xy(Lna[-1],Lair[2].car[0]) RatioAxes=(y3-y1)/(x3-x1) return LH,Lna,LT,LHR,Lair
def testReshape (self):
"test reshape"
w = MA.reshape(self.m, (-1,2))
assert eq(w, [[1,2],[3,11],[12,13]])
w = MA.reshape (self.a, (2,3))
assert eq(w, [[0, 1, 2], [3, 4, 5]])
self.failUnlessRaises(ValueError, MA.reshape, MA.arange(5), (3,2))
self.failUnlessRaises(ValueError, MA.reshape, self.a, (4,-1))
self.failUnlessRaises(ValueError, MA.reshape, self.a, (5,6))
def _check(a):
self.assertEqual(a['na'], na)
self.assertEqual(a['maa'], maa)
self.assertEqual(a['na'][500:-500], na[500:-500])
self.assertEqual(a['maa'][500:-500], maa[500:-500])
scatter = Numeric.array(xrange(0, len(na), 3))
self.assertEqual(a['na'].take(scatter), Numeric.take(na, scatter))
self.assertEqual(Numeric.take(a['na'],scatter),
Numeric.take(na, scatter))
self.assertEqual(MA.take(a['maa'], scatter), MA.take(maa, scatter))
def testLogical (self):
"Test logical_and, logical_or, sometrue, alltrue"
x = MA.array([1,1,0,0])
y = MA.array([1,0,1,0])
assert eq(MA.logical_and (x,y), [1,0,0,0])
assert eq(MA.logical_or (x,y), [1,1,1,0])
assert MA.sometrue(x)
assert not MA.sometrue(MA.zeros((3,)))
assert MA.alltrue(MA.ones((3,)))
assert not MA.alltrue(x)
def testZeros(self):
"Test zeros"
y = MA.zeros((2,3))
assert y.shape == (2,3)
assert y.typecode() == MA.Int
assert eq(y, 0)
z = MA.zeros((2,3), MA.Float)
assert z.shape == (2,3)
assert eq (y, z)
self.failUnlessRaises(ValueError, MA.zeros, (-5,))
def testArange(self):
"Test arange"
x = MA.arange(5)
assert eq(x, [0,1,2,3,4])
x = MA.arange(4, 6)
assert eq(x, [4,5])
x = MA.arange(0, 1.5, .5)
assert eq (x, [0.0, 0.5, 1.0])
assert len(MA.arange(5,3,2)) == 0
x = MA.arange(3L)
assert x[1] == 1L
assert x.typecode() == MA.PyObject
def testTypecodeSpecifying(self):
'Test construction using the type codes.'
from Precision import typecodes
thetypes = typecodes['Integer'] \
+ typecodes['UnsignedInteger'] \
+ typecodes['Float'] \
+ typecodes['Complex']
for t in thetypes:
x = MA.array([1,2,3], t)
assert x.typecode() == t
x = MA.array(['hi', 'hoo'], 'c')
assert x.typecode() == 'c'
def testOnes(self):
"Test ones"
y = MA.ones((2,3))
assert y.shape == (2,3)
assert y.typecode() == MA.Int
assert eq(y.flat, 1)
z = MA.ones((2,3), MA.Float)
assert z.shape == (2,3)
assert eq(y, z)
w = MA.ones((2,3), MA.Int16)
assert eq(w, MA.array([[1,1,1],[1,1,1]],'s'))
self.failUnlessRaises(ValueError, MA.ones, (-5,))
def getMAMid(MA_Short, MA_Mid, symbol, K_MIN, indexoffset=0):
#计算ma以及ma_cross
bardata = pd.read_csv(DC.BAR_DATA_PATH + symbol + '\\' + symbol + ' ' +
str(K_MIN) + '.csv')
madf = pd.DataFrame({'close': bardata.close})
madf['utc_time'] = bardata['utc_time']
madf['strtime'] = bardata['strtime']
madf['index'] = bardata['Unnamed: 0'] - indexoffset
#madf['MA_Short'] = MA.calMA(madf['close'], MA_Short)
madf['MA_Mid'] = MA.calMA(madf['close'], MA_Mid)
madf['MA_True'], madf['MA_Cross'] = MA.dfCross(madf, 'close', 'MA_Mid')
#madf.to_csv("mamid.csv")
return madf
def testIndexing (self):
'Test indexing operations.'
x = MA.array([0,1,2,3,4,5])
for i in range(len(x)):
assert i == x[i]
x[2] = 20
assert eq(x, x[...])
w = MA.array([None])
assert w.typecode() == MA.PyObject
assert w[0] is None
assert isinstance(x[2], types.IntType)
assert x[2] == 20
x = MA.array([0,1,2,3,4,5,6])
assert eq (x[4:1:-1], [4,3,2])
assert eq(x[4:1:-2], [4,2])
assert eq(x[::-1], [6, 5,4,3,2,1,0])
assert eq(x[2:-1], [2,3,4,5])
m = MA.array([[1,2,3],[11,12,13]])
assert m[0,2] == 3
assert isinstance(m[0,2], types.IntType)
assert eq(m[...,1], [2,12])
assert eq(MA.arange(6)[..., MA.NewAxis], [[0],[1],[2],[3],[4],[5]])
x = MA.array([1,2,3])
y = MA.array(x)
x[0] == 66
assert y[0] != 66
b=MA.array([[1,2,3,4],[5,6,7,8]]*2)
# assert b[1:1].shape == (0,4)
# assert b[1:1, :].shape == (0,4)
# assert b[10:].shape == (0,4)
assert eq(b[2:10], [[1,2,3,4],[5,6,7,8]])
assert eq(b[2:10, ...], [[1,2,3,4],[5,6,7,8]])
def copyTest (self):
"Test how MA works with the copy module."
import copy
x = MA.array([1,2,3])
y = [1, x, 3]
c1 = copy.copy(x)
assert MA.allclose(x,c1)
x[1] = 4
assert not MA.allclose(x,c1)
c2 = copy.copy(y)
assert id(c2[1]) == id(x)
c3 = copy.deepcopy(y)
assert id(c3[1]) != id(x)
assert MA.allclose(c3[1], x)
def derive_format_str(self, data):
if len(data) == 0:
return self.default_format_str
width = 10
max = MA.maximum(MA.absolute(data))
if max > 0:
decimals = width - math.log(max, 10) - 2
if decimals < 0:
decimals = 0
elif max >= 0.00001:
decimals = 8
else:
return '%10.10g'
return '%%%d.%df' % (width, decimals)
def _getMaskedArray(self, var, flagVar=None, flagValues=(0,1), missingValuesToTest=(-9999., -32767.)):
"""
Returns a masked array that has the values only present where the flag
values are acceptable and the data itself is not missing.
"""
if flagVar:
flagFillValue=-1
flagGTValue=max(flagValues)
flagMaskWhereUnacceptable=MA.masked_greater(flagVar, flagGTValue).mask()
flagMaskWhereMissing=MA.masked_equal(flagVar, flagFillValue).mask()
flagMask=flagMaskWhereUnacceptable+flagMaskWhereMissing
for fv in missingValuesToTest:
if fv in MV.ravel(var):
print "Setting missing value for '%s' as: %s" % (var.id, fv)
varFillValue=fv
else:
varFillValue=missingValuesToTest[0]
if flagVar:
varMask=MA.masked_array(var, mask=MA.equal(var, varFillValue), fill_value=varFillValue).mask()
fullmask=MA.bitwise_or(varMask, flagMask)
maskedArray=MA.masked_array(var, mask=fullmask, fill_value=varFillValue)
else:
maskedArray=MA.masked_array(var, mask=MA.equal(var, varFillValue), fill_value=varFillValue)
#maskedArray=cdms.createVariable(maskedArray, id=var.id, fill_value=varFillValue)
#maskedArray.missing_value=varFillValue
return maskedArray
def testPickle (self):
"Test pickling of MA arrays."
import pickle
x = MA.arange(10)
fpik = open('foo.pik', 'wb')
pickle.dump(x, fpik, 0)
fpik.close()
fpik = open('foo.pik', 'rb')
y = pickle.load(open('foo.pik', 'rb'))
fpik.close()
assert eq(y,[0, 1, 2, 3, 4, 5, 6, 7, 8, 9])
assert MA.alltrue(MA.equal(x,y))
assert MA.sometrue(MA.equal(x,3))
assert y.shape == (10,)
def _getMaskedArray(self, var, flagVar=None, flagValues=(0, 1), missingValuesToTest=(-9999.0, -32767.0)):
"""
Returns a masked array that has the values only present where the flag
values are acceptable and the data itself is not missing.
"""
if flagVar:
flagFillValue = -1
flagGTValue = max(flagValues)
flagMaskWhereUnacceptable = MA.masked_greater(flagVar, flagGTValue).mask()
flagMaskWhereMissing = MA.masked_equal(flagVar, flagFillValue).mask()
flagMask = flagMaskWhereUnacceptable + flagMaskWhereMissing
for fv in missingValuesToTest:
if fv in MV.ravel(var):
print "Setting missing value for '%s' as: %s" % (var.id, fv)
varFillValue = fv
else:
varFillValue = missingValuesToTest[0]
if flagVar:
varMask = MA.masked_array(var, mask=MA.equal(var, varFillValue), fill_value=varFillValue).mask()
fullmask = MA.bitwise_or(varMask, flagMask)
maskedArray = MA.masked_array(var, mask=fullmask, fill_value=varFillValue)
else:
maskedArray = MA.masked_array(var, mask=MA.equal(var, varFillValue), fill_value=varFillValue)
# maskedArray=cdms.createVariable(maskedArray, id=var.id, fill_value=varFillValue)
# maskedArray.missing_value=varFillValue
return maskedArray
def derive_format_str(self, data):
if len(data) == 0:
return self.default_format_str
width = 10
max = MA.maximum(MA.absolute(data))
if max > 0:
decimals = width - math.log(max, 10) - 2
if decimals < 0:
decimals = 0
elif max >= 0.00001:
decimals = 8
else:
return '%10.10g'
return '%%%d.%df' % (width, decimals)
def testStandard (self):
"Test sqrt, log, log10, exp"
assert eq (MA.exp(1.0), MA.e)
assert eq (MA.sqrt(self.a), map(math.sqrt, self.a))
assert eq (MA.log(self.a), map(math.log, self.a))
assert eq (MA.log10(self.a), map(math.log10, self.a))
assert eq (MA.exp(self.a), map(math.exp, self.a))
assert eq (MA.sqrt(self.m).flat, map(math.sqrt, self.m.flat))
assert eq (MA.log(self.m).flat, map(math.log, self.m.flat))
assert eq (MA.log10(self.m).flat, map(math.log10, self.m.flat))
assert eq (MA.exp(self.m).flat, map(math.exp, self.m.flat))
def testReductions (self):
"Tests of reduce attribute."
a = MA.arange(6)
m = MA.array([[1,2,3],[11,12,13]])
assert MA.add.reduce(a) == 15
assert MA.multiply.reduce(m.shape) == len(m.flat)
assert eq(MA.add.reduce (m, 0), [12,14,16])
assert eq(MA.add.reduce (m, -1), [6,36])
assert eq(MA.multiply.reduce (m, 0), [11,24,39])
assert eq(MA.multiply.reduce (m, -1), [6,11*12*13])
assert MA.add.reduce([1]) == 1
assert MA.add.reduce([]) == 0
assert MA.multiply.reduce([]) == 1
assert MA.minimum.reduce(a) == 0
assert MA.maximum.reduce(a) == 5
def testSpacesaver (self):
"Test the spacesaver property (Travis Oliphant)"
# Test of savespace property: Travis Oliphant
a = MA.array([1,2,3,4],savespace=1)
assert a.spacesaver()
self.assertEqual(a.typecode(), 's')
b = MA.array(a,'f')
self.assertEqual(b.typecode(), 'f')
assert b.spacesaver()
a.savespace(0)
assert not a.spacesaver()
assert b.spacesaver()
d = 4 * b
assert b.typecode() == d.typecode()
self.failUnlessRaises, TypeError, MA.arccos, (b/10.0)
def take(self, data, want):
if type(data) is MA.MaskedArray:
return MA.take(data, want)
elif type(data) is Numeric.ArrayType:
return Numeric.take(data, want)
else:
return _BaseDataType.take(self, data, want)
def getNewLonAxis(oldLonAxis, startLon=-180):
assert (oldLonAxis.isLongitude())
oldLonVals = oldLonAxis.getData()
newLonVals = (oldLonVals - startLon
) % 360 + startLon # range "startLon" to "startLon + 360"
lonValShift = newLonVals - oldLonVals
nlon = len(newLonVals)
##
## identify shift by jump in longitude values
for i in range(1, nlon):
if newLonVals[i] < newLonVals[i - 1]:
nShift = i
break
else:
nShift = 0
## create array with new longitude values
newLonVals = MA.concatenate((newLonVals[nShift:], newLonVals[:nShift]))
## print oldLonVals
## print newLonVals
newLonAxis = cdms.createAxis(newLonVals)
newLonAxis.designateLongitude()
newLonAxis.replace_external_attributes(oldLonAxis.attributes)
#
return newLonAxis, nShift, lonValShift
def get_store_chain(self, data, mask=None):
"""
Return a chain of generators for processing the src data,
which can either be a list or an iterable.
"""
multisourcecols = self.multisourcecols or self.heterosourcecols
if multisourcecols:
src = self._multisrc_gen(multisourcecols)
else:
src = iter(data)
if MA.isMaskedArray(data) and mask is None:
mask = data.mask()
if mask is not None:
src = self._mask_gen(src, mask)
if self.is_multivalue():
src = self._tuple_gen(src, self.ignorenone)
if self.missingvalues:
src = self._tuplemissing_gen(src, self.missingvalues)
else:
if self.calculatedby:
src = self._calcby_gen(src, self.calculatedby,
self.calculatedargs)
if self.missingvalues:
src = self._missing_gen(src, self.missingvalues)
src = self._storedata_gen(src)
return src
def get_store_chain(self, data, mask=None):
"""
Return a chain of generators for processing the src data,
which can either be a list or an iterable.
"""
multisourcecols = self.multisourcecols or self.heterosourcecols
if multisourcecols:
src = self._multisrc_gen(multisourcecols)
else:
src = iter(data)
if MA.isMaskedArray(data) and mask is None:
mask = data.mask()
if mask is not None:
src = self._mask_gen(src, mask)
if self.is_multivalue():
src = self._tuple_gen(src, self.ignorenone)
if self.missingvalues:
src = self._tuplemissing_gen(src, self.missingvalues)
else:
if self.calculatedby:
src = self._calcby_gen(src, self.calculatedby,
self.calculatedargs)
if self.missingvalues:
src = self._missing_gen(src, self.missingvalues)
src = self._storedata_gen(src)
return src
def testComplex (self):
"Test complex numbers"
y = MA.arange(5) * 1.0
x = y + y[::-1] * 1.0j
assert x.typecode() == MA.Complex
assert eq(x.real, y)
assert eq(x.imag, y[::-1])
def take(self, data, want):
if type(data) is MA.MaskedArray:
return MA.take(data, want)
elif type(data) is Numeric.ArrayType:
return Numeric.take(data, want)
else:
return _BaseDataType.take(self, data, want)
def to_R(self, ds, dates_as_factor, **kwargs):
col = ds[self.colname]
if MA.isMaskedArray(col.data):
if Numeric.alltrue(col.data.mask()):
raise PlotError('%r: all data-points masked' % col.name)
return rconv.MA_to_R(col.data)
elif col.is_datetimetype():
return rconv.Missing_Date_to_R(col.data)
else:
return col.data
def _subSampleByAveraging(self, var, timeVar, flagVar, samplingRate, flagsToUse):
"""
Returns a new variable which is 'var' sub-sampled by averaging
at the given samplingRate with data selected according to flagVar
including the flag values specified in flagsToUse (defaults are 0 and 1).
"""
maskedArray = self._getMaskedArray(var, flagVar, flagsToUse)
shape = var.shape
if shape[1] == 1:
newNumArray = MV.ravel(maskedArray)
newArrayMask = MV.ravel(maskedArray.mask())
newArray = MA.masked_array(newNumArray, mask=newArrayMask, fill_value=maskedArray.fill_value())
else:
newArray = Numeric.zeros(shape[0], "f")
for t0 in range(shape[0]):
# Set as missing if less than half are valid
t1Array = maskedArray[t0]
if samplingRate == 1:
# If half or more are good values then calculate the mean
if t1Array.count() >= (shape[1] / 2.0):
newArray[t0] = MA.average(t1Array)
# otherwise set as missing value
else:
newArray[t0] = maskedArray.fill_value()
else:
raise "Averaging for non 1Hz sampling rates not yet supported!"
# Now re-construct variable axes etc
newTimeAxis = self._flatten2DTimeAxis(timeVar, samplingRate)
newVar = self._recreateVariable(
var,
newArray,
newTimeAxis,
flagVar,
max(flagsToUse),
missingValue=maskedArray.fill_value(),
sampleBy="averaging",
)
return newVar
def to_frame(self, frame, ds, **kwargs):
limit_cols = self.limit_cols(ds)
if limit_cols:
self.max = 0
self.min = 0
for colname in (self.colname, ) + limit_cols:
col = ds[colname]
if MA.isMaskedArray(col.data):
if Numeric.alltrue(col.data.mask()):
raise PlotError('%r: all data-points masked' %
col.name)
data = rconv.MA_to_R(col.data)
self.max = max(MA.maximum(col.data), self.max)
self.min = min(MA.minimum(col.data), self.min)
else:
data = col.data
self.max = max(max(col.data), self.max)
self.min = min(min(col.data), self.min)
frame[rconv.rname(colname)] = data
else:
super(RMeasureCondCol, self).to_frame(frame, ds, **kwargs)
def from_col(cls, col, values=None):
self = cls(col.name, col.label, col=col)
if values is not None:
# this_values = sets.Set(col.inverted.keys())
# wanted_values = sets.Set(values)
# if not wanted_values.issubset(this_values):
# missing_vals = [repr(v) for v in (wanted_values - this_values)]
# raise Error('column %r values mismatch - %r' %
# (self.name, ', '.join(missing_vals)))
self.values = values
else:
self.values = col.inverted.keys()
self.values.sort()
self.indices = fully_masked(len(col))
if 0:
# AM - the .inverted attribute of filtered datasets is returning
# invalid row indices, and the summ code relies on these indices,
# so the following code cannot be used until they are fixed.
self.axis_map = {}
for i, v in enumerate(self.values):
self.axis_map[v] = i
vec = col.inverted.get(v)
if vec is not None:
for idx in vec:
try:
self.indices[idx] = i
except IndexError:
print col.name, idx
raise
MA.put(self.indices, vec, i)
else:
axis_map = dict([(v, i) for i, v in enumerate(self.values)])
for i, v in enumerate(col.data):
try:
self.indices[i] = axis_map[v]
except KeyError:
pass
self.axis_map = axis_map
return self
def __setitem__(self, key, a):
index = self.dict.get(key)
if index is None:
if MA.isMaskedArray(a):
index = self._save_new_masked(a)
elif type(a) == Numeric.ArrayType:
index = self._save_new_array(a)
elif type(a) == types.StringType:
index = self._save_new_str(a)
else:
index = self._save_new_str(repr(a))
self.dict[key] = index
self.dict_dirty = 1
else:
if MA.isMaskedArray(a):
self._save_masked(index, a)
elif type(a) == Numeric.ArrayType:
self._save_array(index, a)
elif type(a) == types.StringType:
self._save_str(index, a)
else:
index = self._save_str(index, repr(a))
def __setitem__(self, key, a):
index = self.dict.get(key)
if index is None:
if MA.isMaskedArray(a):
index = self._save_new_masked(a)
elif type(a) == Numeric.ArrayType:
index = self._save_new_array(a)
elif type(a) == types.StringType:
index = self._save_new_str(a)
else:
index = self._save_new_str(repr(a))
self.dict[key] = index
self.dict_dirty = 1
else:
if MA.isMaskedArray(a):
self._save_masked(index, a)
elif type(a) == Numeric.ArrayType:
self._save_array(index, a)
elif type(a) == types.StringType:
self._save_str(index, a)
else:
index = self._save_str(index, repr(a))
def replicate_axis(data, axis, size):
"""
Returns a new array with an additional axis of the specified size.
The new axis is formed by replicating along the other axes.
"""
assert axis > 0
repeat_map = size * Numeric.ones(data.shape[axis-1])
shape = list(data.shape)
data = MA.repeat(data, repeat_map, axis=axis-1)
shape.insert(axis, size)
data.shape = tuple(shape)
return data
def __getitem__(self, key):
index = self.dict[key]
blob = self.store[index]
if blob.type == BLOB_ARRAY:
return MmapArray(blob)
elif blob.type == BLOB_FILLED:
data = MmapArray(blob)
blob = self.store[blob.other]
mask = MmapArray(blob)
return MA.array(data, mask=mask)
elif blob.type == BLOB_STRING:
return blob.as_str()
else:
raise Error('bad BLOB type %s in index' % blob.type)
def BOLL(df, N=26, M=26, P=2):
print 'start computing DMI'
mid = MA.calMA(df['close'], N)
tmp2 = pd.rolling_std(df['close'], M)
TOP = mid + P * tmp2
BOTTOM = mid - P * tmp2
df1 = pd.DataFrame({
'close': df['close'],
'MID': mid,
'TOP': TOP,
'BOTTOM': BOTTOM
})
print 'DMI compute finished'
return df1
def affiche(self) :
l=[]
for i in range(6) :
if Ma.getmask(self.car)[i] :
l.append("--")
else :
if i==0 :
s="%0.f" %(self.car[i])
elif i==1 :
s="%0.3f" %(self.car[i])
else :
s="%0.2f" %(self.car[i])
l.append(s)
return "[ "+join(l," , ")+" ]"
def __getitem__(self, key):
index = self.dict[key]
blob = self.store[index]
if blob.type == BLOB_ARRAY:
return MmapArray(blob)
elif blob.type == BLOB_FILLED:
data = MmapArray(blob)
blob = self.store[blob.other]
mask = MmapArray(blob)
return MA.array(data, mask = mask)
elif blob.type == BLOB_STRING:
return blob.as_str()
else:
raise Error('bad BLOB type %s in index' % blob.type)
def dataPrepare(self):
startTime = datetime.time(self.tradeEndHour, self.tradeStartMin,
0).strftime("%H:%M:%S")
if self.mode == 4:
d, t = self.start_time.split(' ', 1)
y, m, d = d.split('-', 2)
d = datetime.date(int(y), int(m), int(d))
startDate = (
d -
datetime.timedelta(days=self.backwardDay)).strftime("%Y-%m-%d")
endTime = self.start_time
else:
startDate = (
datetime.date.today() -
datetime.timedelta(days=self.backwardDay)).strftime("%Y-%m-%d")
endTime = datetime.date.today().strftime(
"%Y-%m-%d") + ' ' + startTime
sT = startDate + ' ' + startTime
bars = self.get_bars(self.exchange_id + '.' + self.sec_id,
self.K_min * 60, sT, endTime)
#这里数据只用来计算MA
rownum = 0
for bar in bars:
rownum = self.update_dailyBar(bar)
if rownum % self.K_min == 0 and rownum >= self.K_min:
self.update_dailyBarMin(bar)
self.MAdf = MA.MA(self.dailyBarMin.close, self.MA_Short, self.MA_Long)
self.DMIdf = DMI.DMI(self.dailyBarMin, self.DMI_N, self.DMI_M)
kdjdf = KDJ.KDJ(self.dailyBarMin, self.KDJ_N, self.KDJ_M)
kdjdf['KDJ_True'] = 0
kdjdf.loc[(self.KDJ_HLim > kdjdf['KDJ_K']) &
(kdjdf['KDJ_K'] > kdjdf['KDJ_D']), 'KDJ_True'] = 1
kdjdf.loc[(self.KDJ_LLim < kdjdf['KDJ_K']) &
(kdjdf['KDJ_K'] < kdjdf['KDJ_D']), 'KDJ_True'] = -1
self.KDJdf = kdjdf
#下面要再做实盘下当天数据的处理
if self.mode == 2:
pass
if rownum > 0:
self.last_update_time = datetime.datetime.fromtimestamp(
self.dailyBar.iloc[-1].utctime)
print(
"------------------------data prepared-----------------------------"
)
pass
def store_data(self, data, mask, filename=None):
if mask is None:
data = Numeric.array(data, typecode=self.numeric_type)
else:
data = MA.array(data, typecode=self.numeric_type, mask=mask)
if filename:
try:
os.unlink(filename)
except OSError:
pass
data_blob = ArrayDict(filename, 'w+')
data_blob['data'] = data
del data_blob # this writes the data to disc -
# we really need a .sync() method...
return None # Flag for load on demand
else:
return data
def calDMI(df, N=14, M=6):
high = df.high
low = df.low
close = df.close
closeshift1 = close.shift(1).fillna(0)
open = df.open
c = high - low
d = high - closeshift1
df1 = pd.DataFrame({'c': c, 'd': d})
df1['A'] = df1.max(axis=1)
df1.drop('c', axis=1, inplace=True)
df1.drop('d', axis=1, inplace=True)
df1['B'] = np.abs(low - closeshift1)
df1['C'] = df1.max(axis=1)
# df1.drop('A',axis=1,inplace=True)
# df1.drop('B',axis=1,inplace=True)
df1['TR'] = df1['C'].rolling(N).sum()
# 2、HD=最高价-昨日最高价
# 3、LD=昨日最低价-最低价
HD = high - high.shift(1).fillna(0)
LD = low.shift(1).fillna(0) - low
df1['HD'] = HD
df1['LD'] = LD
# DMP:= SUM(IFELSE(HD>0 && HD>LD,HD,0),N);
# DMM:= SUM(IFELSE(LD>0 && LD>HD,LD,0),N);
df2 = pd.DataFrame({'HD': HD, 'LD': LD})
df2['DMP_1'] = df2[(df2['HD'] > df2['LD']) & (df2['HD'] > 0)]['HD']
df2['DMM_1'] = df2[(df2['LD'] > df2['HD']) & (df2['LD'] > 0)]['LD']
df2 = df2.fillna(0)
df1['DMP'] = df2['DMP_1'].rolling(N).sum()
df1['DMM'] = df2['DMM_1'].rolling(N).sum()
del df2
# 6、PDI=DMP*100/TR
# 7、MDI=DMM*100/TR
df1['PDI'] = df1['DMP'] * 100 / df1['TR']
df1['MDI'] = df1['DMM'] * 100 / df1['TR']
# ADX: MA(ABS(MDI-PDI)/(MDI+PDI)*100,M);
# ADXR:(ADX+REF(ADX,M))/2;
df1['ADX'] = MA.calMA(
np.abs(df1['MDI'] - df1['PDI']) / (df1['MDI'] + df1['PDI']) * 100, M)
df1['ADXR'] = (df1['ADX'] + df1['ADX'].shift(M).fillna(0)) / 2
return df1['PDI'], df1['MDI'], df1['ADX'], df1['ADXR']
def train(args, model):
model.train()
loader = DataLoader(MA(
'/Users/zhangweidong03/Code/dl/pytorch/github/piwise/MAdata',
input_transform, target_transform),
num_workers=1,
batch_size=1,
shuffle=True)
weight = torch.ones(2)
weight[0] = 0
use_cuda = False
if use_cuda:
criterion = CrossEntropyLoss2d(weight.cuda())
else:
criterion = CrossEntropyLoss2d(weight)
criterion = CrossEntropyLoss2d()
optimizer = SGD(model.parameters(), 1e-4, .9, 2e-5)
for epoch in range(1, 51):
epoch_loss = []
for step, (images, labels) in enumerate(loader):
if use_cuda:
images = images.cuda()
labels = labels.cuda()
inputs = Variable(images)
targets = Variable(labels)
outputs = model(inputs)
optimizer.zero_grad()
loss = criterion(outputs, targets[:, 0])
loss.backward()
optimizer.step()
epoch_loss.append(loss.data[0])
average = sum(epoch_loss) / len(epoch_loss)
print(f'loss: {average} (epoch: {epoch}, step: {step})')
def updateParadata(self):
#更新MA数据
self.MAdf.loc[self.MAdf.shape[0]] = MA.newMA(self.dailyBarMin.close,
self.MAdf, self.MA_Short,
self.MA_Long)
#更新DMI数据
self.DMIdf.loc[self.DMIdf.shape[0]] = DMI.newDMI(
self.dailyBarMin, self.DMIdf, self.DMI_N, self.DMI_M)
#更新KDJ数据
kdj = KDJ.newKDJ(self.dailyBarMin, self.KDJdf, self.KDJ_N, self.KDJ_M)
newk = kdj[0]
newd = kdj[1]
KDJ_True = 0
if self.KDJ_HLim > newk and newk > newd: KDJ_True = 1
elif self.KDJ_LLim < newk and newk < newd: KDJ_True = -1
kdj.append(KDJ_True)
self.KDJdf.loc[self.KDJdf.shape[0]] = kdj
pass
def summ_to_array(ds, measure, debug=False):
from SOOMv0.CrossTab import CrossTab
ct = CrossTab.from_summset(ds)
data = ct[measure].data
if MA.isMaskedArray(data):
data = data.filled() # Dealing with frequencies, so this is valid.
data = data.astype(Numeric.Float64)
if debug: print type(data)
if debug: print data
array = r.as_array(data)
dimnames = []
labels = []
for axis in ct.axes:
dimnames.append([axis.col.do_outtrans(v) for v in axis.values])
labels.append(axis.col.label or axis.col.name)
dimnames = r.list(*dimnames)
dimnames = r.names__(dimnames, labels)
array = r.dimnames__(array, dimnames)
if debug: r.print_(array)
return array
def test_arraydict(self):
def _check(a):
self.assertEqual(a['na'], na)
self.assertEqual(a['maa'], maa)
self.assertEqual(a['na'][500:-500], na[500:-500])
self.assertEqual(a['maa'][500:-500], maa[500:-500])
scatter = Numeric.array(xrange(0, len(na), 3))
self.assertEqual(a['na'].take(scatter), Numeric.take(na, scatter))
self.assertEqual(Numeric.take(a['na'],scatter),
Numeric.take(na, scatter))
self.assertEqual(MA.take(a['maa'], scatter), MA.take(maa, scatter))
tmpfile = TempFile('soomtestarray_tmpfile')
try:
a = soomarray.ArrayDict(tmpfile.fn(), 'w')
self.assertEqual(len(a), 0)
self.assertRaises(KeyError, a.__getitem__, 'not_found')
na = Numeric.arrayrange(10000)
a['na'] = na
na[1000] = -2222
# MmapArray __setitem__
a['na'][1000] = -2222
maa = MA.arrayrange(10000)
for i in xrange(0, len(maa), 5):
maa[i] = MA.masked
a['maa'] = maa
maa[1000] = -2222
a['maa'][1000] = -2222
_check(a)
del a
a = soomarray.ArrayDict(tmpfile.fn(), 'r')
_check(a)
finally:
try: del a
except UnboundLocalError: pass
tmpfile.done()
def whos(module=globals().keys(), what='A', sstr='', casesen=0, \
out=0):
#=====================================
'''
# Before using whos, do:
execfile('whos.py')
#=== EXAMPLES ===
# to see a list of all objects in the global scope
whos()
# to see a list of functions
whos(what='f')
# to see a list of variables (also their values)
whos(what='v')
# to see a list of Numeric array
whos(what='a')
# to see a list of variables in module os
import os
whos(os, what='v')
# to see a list of all variables of a dictionary
whos({}, what='v')
# to see a list of all functions of a list
whos([], what='f')
# to see a list of variables whose name contains 'path'
whos(what='v', sstr='path')
# to see a list of variables whose name contains 'path' (case insensitive)
whos(what='v', sstr='path', casesen=0)
Argument what can be one of the following:
A -- all
a -- array
f -- function
m -- module
v -- variable
'''
whosList = 0
if type(module) == type([]):
if len(module) > 0:
dirs0 = module
whosList = 1
if not whosList:
temp2 = dir(module)
dirs0 = map(lambda a: 'module.' + a, temp2)
type2 = {}
arrayOk = 1
if what == 'a' or what == 'A':
try:
import Numeric as Num
temp1 = type(Num.zeros((1, )))
type2['array'] = temp1
except:
arrayOk = 0
if arrayOk:
import MA
temp1 = type(MA.zeros((1, )))
type2['MA'] = temp1
if what == 'a' and not arrayOk:
print 'Cannot import Numeric'
return []
if what == 'v' or what == 'A':
temp1 = type(1)
type2['int'] = temp1
temp1 = type(1L)
type2['long'] = temp1
temp1 = type(1.1)
type2['float'] = temp1
temp1 = type('a')
type2['str'] = temp1
if what == 'f' or what == 'A':
temp1 = type(lambda a: a)
type2['function'] = temp1
if what == 'm' or what == 'A':
import os
temp1 = type(os)
type2['module'] = temp1
type1 = type2.values()
dirs0.sort()
if sstr:
if casesen:
p1 = re.compile(sstr)
else:
p1 = re.compile(sstr, re.I)
dirs1 = []
for d in dirs0:
if p1.search(d):
dirs1.append(d)
dirs0 = dirs1
collect1 = []
dirsType = []
if what == 'A':
for i in range(len(dirs0)):
t1 = eval('type(' + dirs0[i] + ')')
collect1.append(i)
dirsType.append(t1)
else:
for i in range(len(dirs0)):
t1 = eval('type(' + dirs0[i] + ')')
#print t1, dirs0[i]
if t1 in type1:
collect1.append(i)
dirsType.append(t1)
for i in range(len(collect1)):
ii = collect1[i]
iType = dirsType[i]
iType1 = string.split(str(iType), "'")[1]
if iType1[:2] == "MA":
iType1 = iType1[:2]
if iType1[:7] == "builtin":
iType1 = iType1[:7]
print '%15s %15s' % (iType1, string.split(dirs0[ii], '.')[-1]),
if (what == 'a' or what == 'A') and arrayOk:
if iType == type2['MA'] or iType == type2['array']:
print '%s %20s ' % (eval(dirs0[ii] + '.typecode()'),
str(eval(dirs0[ii] + '.shape'))),
if iType == type2['MA']:
print 'unmask=%d of %d' % (eval(dirs0[ii] + '.count()'),
eval(dirs0[ii] + '.size()')),
if what == 'v' or what == 'A':
if iType == type2['int'] or iType == type2[
'long'] or iType == type2['float']:
print '= ', eval(dirs0[ii]),
if iType == type2['str']:
temp3 = eval(dirs0[ii])
temp3 = string.replace(temp3, '\n', '\\')
if len(temp3) < 30:
print '= ', temp3,
else:
print '= ', temp3[:30] + ' ...',
print ''
if out:
return [dirs0[i] for i in collect1]
def train(args, model):
model.train()
# loader = DataLoader(MA('/Users/zhangweidong03/Code/dl/pytorch/github/piwise/MAdata', input_transform, target_transform),
# num_workers=1, batch_size=1, shuffle=True)
loader = DataLoader(MA(args.datadir, input_transform, target_transform),
num_workers=args.num_workers,
batch_size=args.batch_size,
shuffle=True)
weight = torch.ones(2)
weight[0] = 0
use_cuda = False
if args.cuda:
criterion = CrossEntropyLoss2d(weight.cuda())
else:
criterion = CrossEntropyLoss2d(weight)
criterion = CrossEntropyLoss2d()
# optimizer = SGD(model.parameters(), 1e-4, .9, 2e-5)
optimizer = Adam(model.parameters())
if args.model.startswith('FCN'):
optimizer = SGD(model.parameters(), 1e-4, .9, 2e-5)
if args.model.startswith('PSP'):
optimizer = SGD(model.parameters(), 1e-2, .9, 1e-4)
if args.model.startswith('Seg'):
optimizer = SGD(model.parameters(), 1e-3, .9)
# for epoch in range(1, 51):
# epoch_loss = []
#
# for step, (images, labels) in enumerate(loader):
# if use_cuda:
# images = images.cuda()
# labels = labels.cuda()
#
# inputs = Variable(images)
# targets = Variable(labels)
# outputs = model(inputs)
#
# optimizer.zero_grad()
# loss = criterion(outputs, targets[:, 0])
# loss.backward()
# optimizer.step()
#
# epoch_loss.append(loss.data[0])
#
# average = sum(epoch_loss) / len(epoch_loss)
# print(f'loss: {average} (epoch: {epoch}, step: {step})')
if args.steps_plot > 0:
board = Dashboard(args.port)
for epoch in range(1, args.num_epochs + 1):
epoch_loss = []
for step, (images, labels) in enumerate(loader):
if args.cuda:
images = images.cuda()
labels = labels.cuda()
inputs = Variable(images)
targets = Variable(labels)
outputs = model(inputs)
optimizer.zero_grad()
loss = criterion(outputs, targets[:, 0])
loss.backward()
optimizer.step()
epoch_loss.append(loss.data[0])
if args.steps_plot > 0 and step % args.steps_plot == 0:
image = inputs[0].cpu().data
image[0] = image[0] * .229 + .485
image[1] = image[1] * .224 + .456
image[2] = image[2] * .225 + .406
board.image(image, f'input (epoch: {epoch}, step: {step})')
board.image(color_transform(outputs[0].cpu().max(0)[1].data),
f'output (epoch: {epoch}, step: {step})')
board.image(color_transform(targets[0].cpu().data),
f'target (epoch: {epoch}, step: {step})')
if args.steps_loss > 0 and step % args.steps_loss == 0:
average = sum(epoch_loss) / len(epoch_loss)
print(f'loss: {average} (epoch: {epoch}, step: {step})')
if args.steps_save > 0 and step % args.steps_save == 0:
filename = f'{args.model}-{epoch:03}-{step:04}.pth'
torch.save(model.state_dict(), filename)
print(f'save: {filename} (epoch: {epoch}, step: {step})')
def IntradayFryWin(symbolinfo, rawdata, paraSet):
setname = paraSet['Setname']
MACD_S = paraSet['MACD_S']
MACD_L = paraSet['MACD_L']
MACD_M = paraSet['MACD_M']
MA_N = paraSet['MA_N']
# print setname
rawdata['Unnamed: 0'] = range(rawdata.shape[0])
#beginindex = rawdata.ix[0, 'Unnamed: 0']
# 计算MACD
macd = MA.calMACD(rawdata['close'], MACD_S, MACD_L, MACD_M)
rawdata['DIF'] = macd[0]
rawdata['DEA'] = macd[1]
rawdata['MA'] = MA.calEMA(rawdata['close'], MA_N)
# 计算MACD的金叉和死叉
rawdata['MACD_True'], rawdata['MACD_Cross'] = MA.dfCross(rawdata, 'DIF', 'DEA')
# ================================ 找出买卖点================================================
# 1.先找出SAR金叉的买卖点
# 2.找到结合判决条件的买点
# 3.从MA买点中滤出真实买卖点
# 取出金叉点
goldcrosslist = pd.DataFrame({'goldcrosstime': rawdata.loc[rawdata['MACD_Cross'] == 1, 'strtime']})
goldcrosslist['goldcrossutc'] = rawdata.loc[rawdata['MACD_Cross'] == 1, 'utc_time']
goldcrosslist['goldcrossindex'] = rawdata.loc[rawdata['MACD_Cross'] == 1, 'Unnamed: 0']
goldcrosslist['goldcrossprice'] = rawdata.loc[rawdata['MACD_Cross'] == 1, 'close']
# 取出死叉点
deathcrosslist = pd.DataFrame({'deathcrosstime': rawdata.loc[rawdata['MACD_Cross'] == -1, 'strtime']})
deathcrosslist['deathcrossutc'] = rawdata.loc[rawdata['MACD_Cross'] == -1, 'utc_time']
deathcrosslist['deathcrossindex'] = rawdata.loc[rawdata['MACD_Cross'] == -1, 'Unnamed: 0']
deathcrosslist['deathcrossprice'] = rawdata.loc[rawdata['MACD_Cross'] == -1, 'close']
goldcrosslist = goldcrosslist.reset_index(drop=True)
deathcrosslist = deathcrosslist.reset_index(drop=True)
# 生成多仓序列（金叉在前，死叉在后）
if goldcrosslist.ix[0, 'goldcrossindex'] < deathcrosslist.ix[0, 'deathcrossindex']:
longcrosslist = pd.concat([goldcrosslist, deathcrosslist], axis=1)
else: # 如果第一个死叉的序号在金叉前，则要将死叉往上移1格
longcrosslist = pd.concat([goldcrosslist, deathcrosslist.shift(-1).fillna(0)], axis=1)
longcrosslist = longcrosslist.set_index(pd.Index(longcrosslist['goldcrossindex']), drop=True)
# 生成空仓序列（死叉在前，金叉在后）
if deathcrosslist.ix[0, 'deathcrossindex'] < goldcrosslist.ix[0, 'goldcrossindex']:
shortcrosslist = pd.concat([deathcrosslist, goldcrosslist], axis=1)
else: # 如果第一个金叉的序号在死叉前，则要将金叉往上移1格
shortcrosslist = pd.concat([deathcrosslist, goldcrosslist.shift(-1).fillna(0)], axis=1)
shortcrosslist = shortcrosslist.set_index(pd.Index(shortcrosslist['deathcrossindex']), drop=True)
# 取出开多序号和开空序号
openlongindex = rawdata.loc[
(rawdata['MACD_Cross'] == 1) & (rawdata['close'] > rawdata['MA']) & (rawdata['ATR'] > 6)].index
openshortindex = rawdata.loc[
(rawdata['MACD_Cross'] == -1) & (rawdata['close'] < rawdata['MA']) & (rawdata['ATR'] > 6)].index
# 从多仓序列中取出开多序号的内容，即为开多操作
longopr = longcrosslist.loc[openlongindex]
longopr['tradetype'] = 1
longopr.rename(columns={'goldcrosstime': 'opentime',
'goldcrossutc': 'openutc',
'goldcrossindex': 'openindex',
'goldcrossprice': 'openprice',
'deathcrosstime': 'closetime',
'deathcrossutc': 'closeutc',
'deathcrossindex': 'closeindex',
'deathcrossprice': 'closeprice'}, inplace=True)
# 从空仓序列中取出开空序号的内容，即为开空操作
shortopr = shortcrosslist.loc[openshortindex]
shortopr['tradetype'] = -1
shortopr.rename(columns={'deathcrosstime': 'opentime',
'deathcrossutc': 'openutc',
'deathcrossindex': 'openindex',
'deathcrossprice': 'openprice',
'goldcrosstime': 'closetime',
'goldcrossutc': 'closeutc',
'goldcrossindex': 'closeindex',
'goldcrossprice': 'closeprice'}, inplace=True)
# 结果分析
result = pd.concat([longopr, shortopr])
result = result.sort_index()
result = result.reset_index(drop=True)
# result.drop(result.shape[0] - 1, inplace=True)
result = result.dropna()
# 去掉跨合约的操作
# 使用单合约，不用再去掉跨合约
# result = removeContractSwap(result, contractswaplist)
slip = symbolinfo.getSlip()
result['ret'] = ((result['closeprice'] - result['openprice']) * result['tradetype']) - slip
result['ret_r'] = result['ret'] / result['openprice']
results = {}
'''
# 使用单合约，策略核心内不再计算结果
if calcResult:
result['commission_fee'], result['per earn'], result['own cash'], result['hands'] = RS.calcResult(result,
symbolinfo,
initialCash,
positionRatio)
endcash = result['own cash'].iloc[-1]
Annual = RS.annual_return(result)
Sharpe = RS.sharpe_ratio(result)
DrawBack = RS.max_drawback(result)[0]
SR = RS.success_rate(result)
max_single_loss_rate = abs(result['ret_r'].min())
results = {
'Setname':setname,
'opentimes': result.shape[0],
'end_cash': endcash,
'SR': SR,
'Annual': Annual,
'Sharpe': Sharpe,
'DrawBack': DrawBack,
'max_single_loss_rate': max_single_loss_rate
}
closeopr = result.loc[:, 'closetime':'tradetype']
return result, rawdata, closeopr, results
'''
return result
def oldnonzero(a):
return a.nonzero()[0]
else:
from Numeric import *
import MA
def gettypecode(x):
return x.typecode()
oldnonzero = nonzero
import os.path
import time
if not with_numpy:
#-- Set this to a more reasonable value
MA.set_print_limit(10000)
# --- Since gist is only loaded on PE0 in the parallel
# --- environment, the parallel module must also be loaded in to check
# --- whether this is in fact running in parallel.
# --- This creates a logical lparallel which is true when running in parallel
# --- and false when running serial.
# --- This also creates a number of routines which are needed for parallel
# --- data handling. These routines all have sensible default results when
# --- not running in parallel.
from ueparallel import *
try:
if me == 0 and sys.platform != 'mac':
from gist import *
def LvyiWin(symbolinfo, rawdata, paraset):
setname = paraset['Setname']
KDJ_N = paraset['KDJ_N']
KDJ_M = paraset['KDJ_M']
KDJ_HLim = paraset['KDJ_HLim']
KDJ_LLim = paraset['KDJ_LLim']
DMI_N = paraset['DMI_N']
DMI_M = paraset['DMI_M']
MA_Short = paraset['MA_Short']
MA_Long = paraset['MA_Long']
rawdata['Unnamed: 0'] = range(rawdata.shape[0])
beginindex = rawdata.ix[0, 'Unnamed: 0']
# 处理KDJ数据：KDJ_OPEN做为最终KDJ的触发信号
# KDJ_True=1:80>k>D
# KDJ_True=-1:20<K<D
df_KDJ = KDJ.KDJ(rawdata, N=KDJ_N, M=KDJ_M)
df_KDJ['KDJ_True'] = 0
df_KDJ.loc[(KDJ_HLim > df_KDJ['KDJ_K']) & (df_KDJ['KDJ_K'] > df_KDJ['KDJ_D']), 'KDJ_True'] = 1
df_KDJ.loc[(KDJ_LLim < df_KDJ['KDJ_K']) & (df_KDJ['KDJ_K'] < df_KDJ['KDJ_D']), 'KDJ_True'] = -1
# 处理DMI数据：DMI_GOLD_CROSS做为DMI的触发信号
# DMI_True=1:and PDI>MDI
# DMI_True=-1:and PDI<MDI
df_DMI = DMI.DMI(rawdata, N=DMI_N, M=DMI_M)
'''
#2017-10-26 10:44
DMI_True的计算，移到DMI模块中执行
df_DMI['DMI_True']=0
df_DMI.loc[df_DMI['PDI']> df_DMI['MDI'], 'DMI_True'] = 1
df_DMI.loc[df_DMI['PDI']< df_DMI['MDI'], 'DMI_True'] = -1
'''
# 处理MA数据：MA_Cross做为MA的触发信号
# MA20_True=1:close>MA20
# MA20_True=-1:close<MA20
df_MA = MA.MA(rawdata['close'], MA_Short, MA_Long)
'''
#2017-10-26:
MA的计算转移到MA模块中
df_MA=pd.DataFrame({'close':rawdata['close']})
df_MA['MA_Short']=MA.calMA(df_MA['close'],MA_Short)
df_MA['MA_Long']=MA.calMA(df_MA['close'],MA_Long)
df_MA['MA_True'],df_MA['MA_Cross']=MA.dfCross(df_MA,'MA_Short','MA_Long')
'''
'''
#处理BOLL数据：BOLL_Cross做为BOLL平仓的触发信号
df_BOLL=BOLL.BOLL(rawdata,N=BOLL_N,M=BOLL_M,P=BOLL_P)
df_BOLL['BOLL_True']=0
df_BOLL.loc[df_BOLL['close']< df_BOLL['BOTTOM'], 'BOLL_True'] = -1
df_BOLL.loc[df_BOLL['close']> df_BOLL['TOP'], 'BOLL_True'] = 1
#BOLL的金叉，BOLL_True从0变成1是金叉，从0变成-1是死叉
df_BOLL['BOLL_Cross']=0
df_BOLL['bigger1'] = df_BOLL['BOLL_True'].shift(1).fillna(0)
df_BOLL.loc[(df_BOLL['BOLL_True']==1)&(df_BOLL['bigger1']==0), 'BOLL_Cross'] = 1
df_BOLL.loc[(df_BOLL['BOLL_True']==-1)&(df_BOLL['bigger1']==0), 'BOLL_Cross'] = -1
df_BOLL.drop('bigger1', axis=1, inplace=True)
df_BOLL.to_csv('df_BOLL.csv')
df_BOLL.drop('close',axis=1,inplace=True)
'''
df_MA.drop('close', axis=1, inplace=True)
df = pd.concat([rawdata, df_DMI, df_KDJ, df_MA], axis=1)
# 找出买卖点：
# 1.先找出MA金叉的买卖点
# 2.找到结合判决条件的买点
# 3.从MA买点中滤出真实买卖点
# 取出金叉点
goldcrosslist = pd.DataFrame({'goldcrosstime': df.loc[df['MA_Cross'] == 1, 'strtime']})
goldcrosslist['goldcrossutc'] = df.loc[df['MA_Cross'] == 1, 'utc_time']
goldcrosslist['goldcrossindex'] = df.loc[df['MA_Cross'] == 1, 'Unnamed: 0'] - beginindex
goldcrosslist['goldcrossprice'] = df.loc[df['MA_Cross'] == 1, 'close']
# 取出死叉点
deathcrosslist = pd.DataFrame({'deathcrosstime': df.loc[df['MA_Cross'] == -1, 'strtime']})
deathcrosslist['deathcrossutc'] = df.loc[df['MA_Cross'] == -1, 'utc_time']
deathcrosslist['deathcrossindex'] = df.loc[df['MA_Cross'] == -1, 'Unnamed: 0'] - beginindex
deathcrosslist['deathcrossprice'] = df.loc[df['MA_Cross'] == -1, 'close']
goldcrosslist = goldcrosslist.reset_index(drop=True)
deathcrosslist = deathcrosslist.reset_index(drop=True)
# 生成多仓序列（金叉在前，死叉在后）
if goldcrosslist.ix[0, 'goldcrossindex'] < deathcrosslist.ix[0, 'deathcrossindex']:
longcrosslist = pd.concat([goldcrosslist, deathcrosslist], axis=1)
else: # 如果第一个死叉的序号在金叉前，则要将死叉往上移1格
longcrosslist = pd.concat([goldcrosslist, deathcrosslist.shift(-1).fillna(0)], axis=1)
longcrosslist = longcrosslist.set_index(pd.Index(longcrosslist['goldcrossindex']), drop=True)
# 生成空仓序列（死叉在前，金叉在后）
if deathcrosslist.ix[0, 'deathcrossindex'] < goldcrosslist.ix[0, 'goldcrossindex']:
shortcrosslist = pd.concat([deathcrosslist, goldcrosslist], axis=1)
else: # 如果第一个金叉的序号在死叉前，则要将金叉往上移1格
shortcrosslist = pd.concat([deathcrosslist, goldcrosslist.shift(-1).fillna(0)], axis=1)
shortcrosslist = shortcrosslist.set_index(pd.Index(shortcrosslist['deathcrossindex']), drop=True)
# 取出开多序号和开空序号
openlongindex = df.loc[(df['MA_Cross'] == 1) & (df['KDJ_True'] == 1) & (df['DMI_True'] == 1)].index
openshortindex = df.loc[(df['MA_Cross'] == -1) & (df['KDJ_True'] == -1) & (df['DMI_True'] == -1)].index
# 从多仓序列中取出开多序号的内容，即为开多操作
longopr = longcrosslist.loc[openlongindex]
longopr['tradetype'] = 1
longopr.rename(columns={'goldcrosstime': 'opentime',
'goldcrossutc': 'openutc',
'goldcrossindex': 'openindex',
'goldcrossprice': 'openprice',
'deathcrosstime': 'closetime',
'deathcrossutc': 'closeutc',
'deathcrossindex': 'closeindex',
'deathcrossprice': 'closeprice'}, inplace=True)
# 从空仓序列中取出开空序号的内容，即为开空操作
shortopr = shortcrosslist.loc[openshortindex]
shortopr['tradetype'] = -1
shortopr.rename(columns={'deathcrosstime': 'opentime',
'deathcrossutc': 'openutc',
'deathcrossindex': 'openindex',
'deathcrossprice': 'openprice',
'goldcrosstime': 'closetime',
'goldcrossutc': 'closeutc',
'goldcrossindex': 'closeindex',
'goldcrossprice': 'closeprice'}, inplace=True)
# 结果分析
result = pd.concat([longopr, shortopr])
result = result.sort_index()
result = result.reset_index(drop=True)
result.drop(result.shape[0] - 1, inplace=True)
# 去掉跨合约的操作
# 使用单合约，不用再去掉跨合约
# result = removeContractSwap(result, contractswaplist)
slip = symbolinfo.getSlip()
result['ret'] = ((result['closeprice'] - result['openprice']) * result['tradetype']) - slip
result['ret_r'] = result['ret'] / result['openprice']
results = {}
'''
# 使用单合约，策略核心内不再计算结果
if calcResult:
result['commission_fee'], result['per earn'], result['own cash'], result['hands'] = RS.calcResult(result,
symbolinfo,
initialCash,
positionRatio)
endcash = result['own cash'].iloc[-1]
Annual = RS.annual_return(result)
Sharpe = RS.sharpe_ratio(result)
DrawBack = RS.max_drawback(result)[0]
SR = RS.success_rate(result)
max_single_loss_rate = abs(result['ret_r'].min())
results = {
'Setname':setname,
'opentimes': result.shape[0],
'end_cash': endcash,
'SR': SR,
'Annual': Annual,
'Sharpe': Sharpe,
'DrawBack': DrawBack,
'max_single_loss_rate': max_single_loss_rate
}
closeopr = result.loc[:, 'closetime':'tradetype']
return result, df, closeopr, results
'''
return result
def LvyiNoKDJWin(symbolinfo,
setname,
rawdata,
paraset,
contractswaplist,
calcResult=True):
print setname
DMI_N = paraset['DMI_N']
DMI_M = paraset['DMI_M']
MA_Short = paraset['MA_Short']
MA_Long = paraset['MA_Long']
beginindex = rawdata.ix[0, 'Unnamed: 0']
df_DMI = DMI.DMI(rawdata, N=DMI_N, M=DMI_M)
df_MA = MA.MA(rawdata['close'], MA_Short, MA_Long)
df_MA.drop('close', axis=1, inplace=True)
df = pd.concat([rawdata, df_DMI, df_MA], axis=1)
# 找出买卖点：
# 1.先找出MA金叉的买卖点
# 2.找到结合判决条件的买点
# 3.从MA买点中滤出真实买卖点
# 取出金叉点
goldcrosslist = pd.DataFrame(
{'goldcrosstime': df.loc[df['MA_Cross'] == 1, 'strtime']})
goldcrosslist['goldcrossutc'] = df.loc[df['MA_Cross'] == 1, 'utc_time']
goldcrosslist['goldcrossindex'] = df.loc[df['MA_Cross'] == 1,
'Unnamed: 0'] - beginindex
goldcrosslist['goldcrossprice'] = df.loc[df['MA_Cross'] == 1, 'close']
# 取出死叉点
deathcrosslist = pd.DataFrame(
{'deathcrosstime': df.loc[df['MA_Cross'] == -1, 'strtime']})
deathcrosslist['deathcrossutc'] = df.loc[df['MA_Cross'] == -1, 'utc_time']
deathcrosslist['deathcrossindex'] = df.loc[df['MA_Cross'] == -1,
'Unnamed: 0'] - beginindex
deathcrosslist['deathcrossprice'] = df.loc[df['MA_Cross'] == -1, 'close']
goldcrosslist = goldcrosslist.reset_index(drop=True)
deathcrosslist = deathcrosslist.reset_index(drop=True)
# 生成多仓序列（金叉在前，死叉在后）
if goldcrosslist.ix[0, 'goldcrossindex'] < deathcrosslist.ix[
0, 'deathcrossindex']:
longcrosslist = pd.concat([goldcrosslist, deathcrosslist], axis=1)
else: # 如果第一个死叉的序号在金叉前，则要将死叉往上移1格
longcrosslist = pd.concat(
[goldcrosslist, deathcrosslist.shift(-1).fillna(0)], axis=1)
longcrosslist = longcrosslist.set_index(pd.Index(
longcrosslist['goldcrossindex']),
drop=True)
# 生成空仓序列（死叉在前，金叉在后）
if deathcrosslist.ix[0, 'deathcrossindex'] < goldcrosslist.ix[
0, 'goldcrossindex']:
shortcrosslist = pd.concat([deathcrosslist, goldcrosslist], axis=1)
else: # 如果第一个金叉的序号在死叉前，则要将金叉往上移1格
shortcrosslist = pd.concat(
[deathcrosslist, goldcrosslist.shift(-1).fillna(0)], axis=1)
shortcrosslist = shortcrosslist.set_index(pd.Index(
shortcrosslist['deathcrossindex']),
drop=True)
#取出开多序号和开空序号
openlongindex = df.loc[(df['MA_Cross'] == 1) & (df['DMI_True'] == 1)].index
openshortindex = df.loc[(df['MA_Cross'] == -1)
& (df['DMI_True'] == -1)].index
# 从多仓序列中取出开多序号的内容，即为开多操作
longopr = longcrosslist.loc[openlongindex]
longopr['tradetype'] = 1
longopr.rename(columns={
'goldcrosstime': 'opentime',
'goldcrossutc': 'openutc',
'goldcrossindex': 'openindex',
'goldcrossprice': 'openprice',
'deathcrosstime': 'closetime',
'deathcrossutc': 'closeutc',
'deathcrossindex': 'closeindex',
'deathcrossprice': 'closeprice'
},
inplace=True)
# 从空仓序列中取出开空序号的内容，即为开空操作
shortopr = shortcrosslist.loc[openshortindex]
shortopr['tradetype'] = -1
shortopr.rename(columns={
'deathcrosstime': 'opentime',
'deathcrossutc': 'openutc',
'deathcrossindex': 'openindex',
'deathcrossprice': 'openprice',
'goldcrosstime': 'closetime',
'goldcrossutc': 'closeutc',
'goldcrossindex': 'closeindex',
'goldcrossprice': 'closeprice'
},
inplace=True)
# 结果分析
result = pd.concat([longopr, shortopr])
result = result.sort_index()
result = result.reset_index(drop=True)
result.drop(result.shape[0] - 1, inplace=True)
# 去掉跨合约的操作
result = removeContractSwap(result, contractswaplist)
initial_cash = 20000
margin_rate = 0.2
slip = symbolinfo.getSlip()
multiplier = symbolinfo.getMultiplier()
poundgeType, poundgeFee, poundgeRate = symbolinfo.getPoundage()
result['ret'] = ((result['closeprice'] - result['openprice']) *
result['tradetype']) - slip
result['ret_r'] = result['ret'] / result['openprice']
results = {}
if calcResult:
firsttradecash = initial_cash / margin_rate
result['commission_fee'] = 0
if poundgeType == symbolinfo.POUNDGE_TYPE_RATE:
result.ix[0, 'commission_fee'] = firsttradecash * poundgeRate * 2
else:
result.ix[0, 'commission_fee'] = firsttradecash / (
multiplier * result.ix[0, 'openprice']) * poundgeFee * 2
result['per earn'] = 0 # 单笔盈亏
result['own cash'] = 0 # 自有资金线
result['trade money'] = 0 # 杠杆后的可交易资金线
result.ix[0, 'per earn'] = firsttradecash * result.ix[0, 'ret_r']
result.ix[0, 'own cash'] = initial_cash + result.ix[
0, 'per earn'] - result.ix[0, 'commission_fee']
result.ix[0, 'trade money'] = result.ix[0, 'own cash'] / margin_rate
oprtimes = result.shape[0]
for i in np.arange(1, oprtimes):
# 根据手续费类型计算手续费
if poundgeType == symbolinfo.POUNDGE_TYPE_RATE:
commission = result.ix[i - 1, 'trade money'] * poundgeRate * 2
else:
commission = result.ix[i - 1, 'trade money'] / (
multiplier * result.ix[i, 'openprice']) * poundgeFee * 2
perearn = result.ix[i - 1, 'trade money'] * result.ix[i, 'ret_r']
owncash = result.ix[i - 1, 'own cash'] + perearn - commission
result.ix[i, 'own cash'] = owncash
result.ix[i, 'commission_fee'] = commission
result.ix[i, 'per earn'] = perearn
result.ix[i, 'trade money'] = owncash / margin_rate
endcash = result.ix[oprtimes - 1, 'own cash']
Annual = RS.annual_return(result)
Sharpe = RS.sharpe_ratio(result)
DrawBack = RS.max_drawback(result)[0]
SR = RS.success_rate(result)
max_single_loss_rate = abs(result['ret_r'].min())
results = {
'Setname': setname,
'MA_Short': MA_Short,
'MA_Long': MA_Long,
'DMI_N': DMI_N,
'opentimes': oprtimes,
'end_cash': endcash,
'SR': SR,
'Annual': Annual,
'Sharpe': Sharpe,
'DrawBack': DrawBack,
'max_single_loss_rate': max_single_loss_rate
}
closeopr = result.loc[:, 'closetime':'tradetype']
return result, df, closeopr, results
# How to use numpy with 'None' value in Python? import MA a = MA.array([1, 2, None], mask = [0, 0, 1]) print "average =", MA.average(a)
def rgrd(self, dataIn=None):
""" --------------------------------------------------------------------------------------------------------
ROUTINE: rgrd
PURPOSE: To perform one of the following two possible computations in 3-space:
To interpolate random data in 3-space using a modified Shepard's algorithm
To find the nearest neighbor to a user specified point
USAGE: To interpolate use
dataOut = r.rgrd(dataIn) where
r -- an instance of Shgrid
dataIn -- input data
dataOut -- output data
To locate the nearest point use
np = r.rgrd(numberPoints) where
r -- an instance of Shgrid
np -- the array with numberPoints indices into the input grid idenifying the nearest points
--------------------------------------------------------------------------------------------------------"""
if self.callGridded == 'yes': # Interpolation in 3-space
if usefilled == 'yes':
dataIn = MA.filled(dataIn)
if debug == 1:
print 'calling shgrid'
iwk = numpy.zeros((2 * self.ni, ), 'i')
rwk = numpy.zeros((11 * self.ni + 6, ), 'f')
dataOut, ier = shgridmodule.shgrid(self.xi, self.yi, self.zi,
dataIn, self.xo, self.yo,
self.zo, iwk, rwk)
dataOut = numpy.transpose(dataOut)
if debug == 1:
print '***************** returning from shgrid with ier = ', ier
if ier != 0:
msg = 'Error in return from shgrid call with -- ' + Shgrid.errorTable(
self)[ier]
raise ValueError, msg
# is a reverse the order in the returned arrays necessary?
if (self.xreverse == 'yes') or (self.yreverse
== 'yes') or (self.zreverse
== 'yes'):
needReverse = 'yes'
else:
needReverse = 'no'
if needReverse == 'yes':
dataOut = Shgrid.reverseData(self, dataOut)
return dataOut
#---------------------------------------------------------------------------------
else: # Computation of nearest neighbors
if debug == 1:
print 'calling shgetnp'
numberPoints = dataIn
np = numpy.zeros((numberPoints, ), numpy.int32)
n = self.ni
iwk = numpy.zeros((2 * n, ), numpy.int32)
rwk = numpy.zeros((11 * n + 6, ), numpy.float32)
iflag = 0
np[0], ier = shgridmodule.shgetnp(self.xo, self.yo, self.zo,
self.xi, self.yi, self.zi, iflag,
iwk, rwk)
if debug == 1:
print '***************** returning from shgrid with ier = ', ier
if ier != 0:
msg = 'Error in return from shgetnp call with -- ' + Shgrid.errorTable(
self)[ier]
raise ValueError, msg
iflag = 1
for i in range(1, numberPoints):
np[i], ier = shgridmodule.shgetnp(self.xo, self.yo, self.zo,
self.xi, self.yi, self.zi,
iflag, iwk, rwk)
if debug == 1:
print '***************** returning from shgrid with ier = ', ier
if ier != 0:
msg = 'Error in return from shgetnp call with -- ' + Shgrid.errorTable(
self)[ier]
raise ValueError, msg
return np
def makeFigure(self):
"""
:todo: Uses pylab so may not be threadsafe.
"""
lon = self.grid.lon0 + N.arrayrange(self.grid.nlon) * self.grid.dlon
lat = self.grid.lat0 + N.arrayrange(self.grid.nlat) * self.grid.dlat
ll_lon, ll_lat = lon[0], lat[0]
ur_lon, ur_lat = lon[-1], lat[-1]
# Account for variable lat/lon axis ordering
#!TODO: Could we move this (along with the equivilent in render_imp.py into grid.py?
if self.grid.ilat < self.grid.ilon:
latLonOrdering = True
else:
latLonOrdering = False
if latLonOrdering:
var = self.grid.value
else:
var = MA.transpose(self.grid.value)
fig = p.figure()
map = Basemap(projection='cyl',
llcrnrlon=ll_lon,
llcrnrlat=ll_lat,
urcrnrlon=ur_lon,
urcrnrlat=ur_lat,
resolution='l')
## if self.grid.units:
## p.title("%s\n(%s)" % (self.grid.long_name, self.grid.units))
## else:
## p.title(self.grid.long_name)
p.title(self.grid.long_name)
if self.type == 'colour':
# transform_scalar doesn't support masked arrays so we must fill then replace the mask.
var_dat = map.transform_scalar(var.filled(1.0e20), lon, lat,
len(lon), len(lat))
var = MA.masked_values(var_dat, 1.0e20)
map.imshow(var,
cmap=self.cmap,
vmin=self.vmin,
vmax=self.vmax,
interpolation='nearest')
cbar = p.colorbar(orientation='horizontal', format='%.2g')
if self.grid.units:
cbar.ax.set_xlabel(self.grid.units)
else:
x, y = map(*p.meshgrid(lon, lat))
c = map.contour(x, y, var, 12, colors='black')
c.clabel(fontsize=8)
map.fillcontinents(color='#e0e0e0')
map.drawcoastlines(color='gray')
map.drawmeridians(p.arange(-180, 180, 30),
labels=[1, 0, 0, 1],
color='gray')
map.drawparallels(p.arange(-90, 90, 15),
labels=[1, 0, 0, 1],
color='gray')
# Wrap the caption
caption = word_wrap(self.caption, 80)
fig.text(0.1,
0.08,
caption,
fontsize=10,
horizontalalignment='left',
verticalalignment='top',
transform=fig.transFigure)
return fig
