def makeSpatialVx(X, Xhat, thresholds, loc, projection, subset, timevals,
reggrid, Map, locbyrow, fieldtype, units, dataname, obsname,
modelname, q, qs):
#参数：X, Xhat, threshold, loc, projection, subset, timevals,reggrid, map, locbyrow, fieldtype, units, dataname, obsname, modelname, q, qs
'''
#初始化参数
thresholds = np.array([0.01, 20.01]) #清空threshold
loc = None
projection = False
subset = None
timevals = None
reggrid = True #logical, is the grid a regular grid?
map = False #logical, should a map be added to image plots of the data?
locbyrow = False
fieldtype = ""
units = ""
dataname = ""
obsname = "X"
modelname = "Xhat"
q = (0, 0.1, 0.25, 0.33, 0.5, 0.66, 0.75, 0.9, 0.95)
qs = None
'''
#if (X.shape == None):
if (type(X) == list): #判断X是否为列表，X实际为二维矩阵
nobs = X.shape[0] * X.shape[1]
xdim = None
else:
nobs = 1
xdim = X.shape
#if (Xhat.shape == None):
if (type(Xhat) == list): #判断X是否为列表
nforecast = Xhat.shape[0] * Xhat.shape[1]
ydim = None
else:
nforecast = 1
ydim = Xhat.shape
if (xdim != ydim):
try:
sys.exit(0)
except:
print(
"make.SpatialVx: dim of X{0} must be the same as dim of Xhat{1} "
.format(X.shape, Xhat.shape))
out = {"X": X, "Xhat": Xhat} #将矩阵存为列表
if (all(thresholds) == None):
if (nobs > 1): #猜测是输入的观测数据大于一个时次，对每个时次的数据分别求分位数
othresh = np.zeros((9, 1)) #定义othresh为空数组，长度与q一致，9
for i in range(1, nobs):
#X_1 = np.array(X)
#X_flat = X_1.flatten() #dataframe没有flatten属性，先转array
quantile_X = np.quantile(X, q) #计算X的分位数
othresh = np.hstack([quantile_X, othresh])
else:
othresh = np.quantile(X, q) #数据大小为（9，1）是写死的，因为所要求的的分位数的个数是确定的
if (nforecast > 1): #猜测是输入的观测数据大于一个时次，对每个时次的数据分别求分位数
fthresh = np.zeros() #定义fthresh为空数组，长度未知
for i in range(1, nforecast):
quantile_Xhat = np.quantile(Xhat, q) #计算Xhat的q分位数
fthresh = np.hstack([fthresh,
quantile_Xhat]) #原函数里面直接生成的矩阵，没有赋值到任何变量
else:
fthresh = np.quantile(Xhat, q) #数据大小为（9，1）是写死的，因为所要求的的分位数的个数是确定的
qs = str(q)
X = othresh
Xhat = fthresh
thresholds = [X, Xhat]
elif (type(thresholds) == list): #判断thresholds是否为一个list
threshnames = ["X", "Xhat"] #获取列表的名称
if (("X" not in threshnames) or ("Xhat" not in threshnames)):
try:
sys.exit(0)
except:
print(
"make.SpatialVx: invalid thresholds argument. List must have components named X and Xhat."
)
odim = np.array([len(thresholds),
np.ndim(thresholds)]) #thresholds 第一项(X)的数组长度赋值给odim
fdim = np.array([len(thresholds), np.ndim(thresholds)
]) #thresholds 第二项(Xhat)的数组长度赋值给fdim
if (odim[0] != fdim[0]):
try:
sys.exit(0)
except:
print(
"make.SpatialVx: invalid thresholds argument. X must have same number of thresholds (rows) as Xhat."
)
nth = odim[0]
if (any(odim) == None):
thresholds[0] = np.array(thresholds[0]).reshape(
len(thresholds[0]), nobs)
elif (odim[1] == 1 and nobs > 1):
thresholds[0] = np.array(thresholds[0]).reshape(
len(thresholds[0]), nobs)
elif (odim[1] > 1 and odim[1] != nobs):
try:
sys.exit(0)
except:
print("make.SpatialVx: invalid thresholds argument.")
threshold = 'threshold'
qs = []
for i in range(nth):
qs.append(threshold + ' ' + str(i))
elif (
type(thresholds) == list
): #判断thresholds是否为vector,R里面检查是否最多只有一个属性：name。即查看其属性，如果没有属性或者只有一个name属性，才返回TRUE
qs = str(thresholds)
nth = len(thresholds)
thresholds = [
np.array(thresholds).reshape(nth, nobs),
np.array(thresholds).reshape(nth, nforecast)
]
#else stop("make.SpatialVx: invalid thresholds argument. Must be a vector or list.")
else:
try:
sys.exit(0)
except:
print(
"make.SpatialVx: invalid thresholds argument. Must be a vector or list."
)
if (any(loc) == None): #判断loc是否为空
rep00_1 = np.arange(1, xdim[0], 1)
rep00 = np.tile(rep00_1, xdim[1])
rep0_1 = np.arnge(1, xdim[1], 1)
rep0 = rep0_1.repeat(xdim[0], axis=0) #按行进行元素重复
loc = np.vstack(rep00, rep0)
if (timevals == None):
if (len(xdim) == 3):
timevals = np.arnge(1, xdim[2])
else:
timevals = 1
if (dataname != None):
msg = dataname
else:
msg = ""
if (fieldtype != "" and units != ""):
msg = msg + '\n' + fieldtype + ' ' + '(' + units + ')'
elif (fieldtype != ""):
msg = msg + '\n' + fieldtype
elif (units != ""):
msg = msg + ' ' + '(' + units + ')'
out = {"X": X, "Xhat": Xhat, "class":"SpatialVx", "xdim":xdim, "time":timevals,\
"thresholds":thresholds, "loc":loc, "locbyrow":locbyrow, \
"subset":subset, "dataname":dataname, "obsname":obsname, \
"modelname":modelname, "nobs":nobs, "nforecast":nforecast, \
"fieldtype":fieldtype , "units":units , "projection":projection , \
"reggrid":reggrid , "Map":Map , "qs":qs, "msg":msg}
return out
print("No date found for ind: ", ind)
try:
if len(record["MedlineCitation"]["Article"]["ArticleTitle"].split(COI)) > 1:
yearcolondict[year].appepnd(1)
else:
yearcolondict[year].append(0)
except:
print("The next try statement didn"t work.")
# Plot.
fig = plt.figure()
ax = fig.add_subplot(111)
x = []; y = []
for year in yearcolondict.keys():
totescolons = float(sum(yearcolondict[year]))
pubsthisyear = float(len(yearcolondict[year]))
ax.scatter(year, totescolons/pubsthisyear, c="k", s=80, edgecolor="w")
x.append(year)
y.append(totescolons/pubsthisyear)
# Linear regression
slope, intercept, rvalue, pvalue, stderr = stats.linregress(x, y)
xacls = np.arnge(min(x), max(x), 1)
yvals = slope*xvals + intercept
ax.plot(xvals, yvals, c="k")
ax.set_title("Proportions per year for search term " + searchterm + ", slope =" + str(slope)[:6] + "/yr", style="italic")
ax.set_xlabel("Year", style="italic")
ax.set_ylabel("Proportion of the titles containing search term", style="italic")
ax.set_ylim([0, .55])
plt.savefig("searchterms.png")
def make_spatialVx(grd_ob,
grd_fo,
loc,
fieldtype="",
units="",
dataname="",
obsname="ob",
modelname="fo"):
#参数：X, Xhat, threshold, loc, projection, subset, timevals,reggrid, map, locbyrow, fieldtype, units, dataname, obsname, modelname, q, qs
thresholds = [None, None] #画图阈值
ob = grd_ob
fo = grd_fo
projection = False
subset = None
timevals = None
reggrid = True
Map = False
locbyrow = False
qs = None
q = (0, 0.1, 0.25, 0.33, 0.5, 0.66, 0.75, 0.9, 0.95)
#if (X.shape == None):
if (type(ob) == list): #判断X是否为列表，X实际为二维矩阵
nobs = ob.shape[0] * ob.shape[1]
xdim = None
else:
nobs = 1
xdim = ob.shape
#if (Xhat.shape == None):
if (type(fo) == list): #判断X是否为列表
nforecast = fo.shape[0] * fo.shape[1]
ydim = None
else:
nforecast = 1
ydim = fo.shape
if (xdim != ydim):
try:
sys.exit(0)
except:
print(
"make.SpatialVx: dim of ob{0} must be the same as dim of fo{1} "
.format(ob.shape, fo.shape))
out = {"ob": ob, "fo": fo} #将矩阵存为列表
if (all(thresholds) == None):
if (nobs > 1): #猜测是输入的观测数据大于一个时次，对每个时次的数据分别求分位数
othresh = np.zeros((9, 1)) #定义othresh为空数组，长度与q一致，9
for i in range(1, nobs):
#X_1 = np.array(X)
#X_flat = X_1.flatten() #dataframe没有flatten属性，先转array
quantile_X = np.quantile(ob, q) #计算X的分位数
othresh = np.hstack([quantile_X, othresh])
else:
othresh = np.quantile(ob, q) #数据大小为（9，1）是写死的，因为所要求的的分位数的个数是确定的
if (nforecast > 1): #猜测是输入的观测数据大于一个时次，对每个时次的数据分别求分位数
fthresh = np.zeros() #定义fthresh为空数组，长度未知
for i in range(1, nforecast):
quantile_Xhat = np.quantile(fo, q) #计算Xhat的q分位数
fthresh = np.hstack([fthresh,
quantile_Xhat]) #原函数里面直接生成的矩阵，没有赋值到任何变量
else:
fthresh = np.quantile(fo, q) #数据大小为（9，1）是写死的，因为所要求的的分位数的个数是确定的
qs = str(q)
#X = othresh
#Xhat = fthresh
thresholds = [othresh, fthresh]
elif (type(thresholds) == list): #判断thresholds是否为一个list
threshnames = ["ob", "fo"] #获取列表的名称
if (("ob" not in threshnames) or ("fo" not in threshnames)):
try:
sys.exit(0)
except:
print(
"make.SpatialVx: invalid thresholds argument. List must have components named ob and fo."
)
odim = np.array([len(thresholds),
np.ndim(thresholds)]) #thresholds 第一项(X)的数组长度赋值给odim
fdim = np.array([len(thresholds), np.ndim(thresholds)
]) #thresholds 第二项(Xhat)的数组长度赋值给fdim
if (odim[0] != fdim[0]):
try:
sys.exit(0)
except:
print(
"make.SpatialVx: invalid thresholds argument. ob must have same number of thresholds (rows) as fo."
)
nth = odim[0]
if (any(odim) == None):
thresholds[0] = np.array(thresholds[0]).reshape(
len(thresholds[0]), nobs)
elif (odim[1] == 1 and nobs > 1):
thresholds[0] = np.array(thresholds[0]).reshape(
len(thresholds[0]), nobs)
elif (odim[1] > 1 and odim[1] != nobs):
try:
sys.exit(0)
except:
print("make.SpatialVx: invalid thresholds argument.")
threshold = 'threshold'
qs = []
for i in range(nth):
qs.append(threshold + ' ' + str(i))
elif (
type(thresholds) == list
): #判断thresholds是否为vector,R里面检查是否最多只有一个属性：name。即查看其属性，如果没有属性或者只有一个name属性，才返回TRUE
qs = str(thresholds)
nth = len(thresholds)
thresholds = [
np.array(thresholds).reshape(nth, nobs),
np.array(thresholds).reshape(nth, nforecast)
]
#else stop("make.SpatialVx: invalid thresholds argument. Must be a vector or list.")
else:
try:
sys.exit(0)
except:
print(
"make.SpatialVx: invalid thresholds argument. Must be a vector or list."
)
if (loc.any() == None): #判断loc是否为空,如果为空的话，生成坐标信息
rep00_1 = np.arange(1, xdim[0], 1)
rep00 = np.tile(rep00_1, xdim[1])
rep0_1 = np.arange(1, xdim[1], 1)
rep0 = rep0_1.repeat(xdim[0], axis=0) #按行进行元素重复
loc = np.vstack((rep00, rep0))
if (timevals == None):
if (len(xdim) == 3):
timevals = np.arnge(1, xdim[2])
else:
timevals = 1
if (dataname != None):
msg = dataname
else:
msg = ""
if (fieldtype != "" and units != ""):
msg = msg + '\n' + fieldtype + ' ' + '(' + units + ')'
elif (fieldtype != ""):
msg = msg + '\n' + fieldtype
elif (units != ""):
msg = msg + ' ' + '(' + units + ')'
'''
out = {"ob": ob, "fo": fo, "class":"SpatialVx", "xdim":xdim, "time":timevals,
"thresholds":thresholds, "loc":loc, "locbyrow":locbyrow,
"subset":subset, "dataname":dataname, "obsname":obsname,
"modelname":modelname, "nobs":nobs, "nforecast":nforecast,
"fieldtype":fieldtype , "units":units , "projection":projection ,
"reggrid":reggrid , "Map":Map , "qs":qs, "msg":msg}
'''
out = {
"grd_ob": ob,
"grd_fo": fo,
"class": "SpatialVx",
"xdim": xdim,
"loc": loc,
"dataname": dataname,
"obsname": obsname,
"modelname": modelname,
"fieldtype": fieldtype,
"units": units,
"msg": msg
}
return out
b = pd.Series([9, 8, 7, 6], index=['a', 'b', 'c', 'd'])
b
c = pd.Series(25, index=['a', 'b', 'c'])
c
a = pd.Series({"a": 1, "b": 2, "c": 3, "d": 4})
print(a)
print(a[1])
print(a["b"])
d = pd.Series(range(20))
print(d)
d.cumsum()
e = pd.Series(np.arange(5))
e
n = pd.Series(np.arange(5), index=np.arnge(9, 4, -1))
n
k = pd.Series([1, 2, 3], ['c', 'd', 'e'])
v = pd.Series([9, 8, 7, 6], ['a', 'b', 'c', 'd'])
k + v
v = pd.Series([9, 8, 7, 6], ['a', 'b', 'c', 'd'])
v['b'] = 15
v.name = "Series"
v['a', 'c'] = 20
import numpy as np
import matplotlib.pyplot as plt
# 다양한 그래프 그리기
# 343p
days = np.arnge(1, 11)
weight = ([10, 14, 16, 17, 13])
plt.plot(days,
weight,
marker="o",
markerfacecolor="k",
linestyle="--",
color="b")
# 막대그래프 만들기
x = np.arange(1, 11)
y1 = np.arange(11, 21)
y2 = np.arange(5, 10)
labels = ([1, 2, 3, 4, 5])
plt.bar(x, y1, tick_label=labels)
plt.bar(x, y2, bottom=y1) # 누적막대그래프
# 히스토그램
data = np.random.randn(1000)