def test_centered_bins(self):
"""See that we can compute some nice centered bins"""
a = plot.centered_bins(10, bins=9)
self.assertEquals(a[0], -12)
a = plot.centered_bins(55, bins=9)
self.assertEquals(a[0], -56)
a = plot.centered_bins(99, bins=9)
self.assertEquals(a[0], -100)
a = plot.centered_bins(0.9, bins=9)
self.assertEquals(a[0], -0.9)
def test_centered_bins():
"""See that we can compute some nice centered bins"""
a = centered_bins(10, bins=9)
assert a[0] == -12
a = centered_bins(55, bins=9)
assert a[0] == -56
a = centered_bins(99, bins=9)
assert a[0] == -100
a = centered_bins(0.9, bins=9)
assert a[0] == -0.9
def test_centered_bins(self):
"""See that we can compute some nice centered bins"""
a = plot.centered_bins(10, bins=9)
self.assertEquals(a[0], -12)
a = plot.centered_bins(55, bins=9)
self.assertEquals(a[0], -56)
a = plot.centered_bins(99, bins=9)
self.assertEquals(a[0], -100)
a = plot.centered_bins(0.9, bins=9)
self.assertEquals(a[0], -0.9)
def plotter(fdict):
""" Go """
import matplotlib
matplotlib.use('agg')
import matplotlib.pyplot as plt
pgconn = psycopg2.connect(database='coop', host='iemdb', user='******')
ctx = get_autoplot_context(fdict, get_description())
state = ctx['state'][:2]
sector = ctx['sector']
opt = ctx['opt']
p1syear = ctx['p1syear']
p1eyear = ctx['p1eyear']
p2syear = ctx['p2syear']
p2eyear = ctx['p2eyear']
varname = ctx['var']
table = "alldata"
if sector == 'state':
table = "alldata_%s" % (state, )
df = read_sql("""
WITH season1 as (
SELECT station, year,
min(case when month > 7 and low < 32 then
extract(doy from day) else 366 end) as first_freeze,
max(case when month < 7 and low < 32 then
extract(doy from day) else 0 end) as last_freeze
from """ + table + """ WHERE
year >= %s and year <= %s GROUP by station, year),
season2 as (
SELECT station, year,
min(case when month > 7 and low < 32 then
extract(doy from day) else 366 end) as first_freeze,
max(case when month < 7 and low < 32 then
extract(doy from day) else 0 end) as last_freeze
from """ + table + """ WHERE
year >= %s and year <= %s GROUP by station, year),
agg as (
SELECT p1.station, avg(p1.first_freeze) as p1_first_fall,
avg(p1.last_freeze) as p1_last_spring,
avg(p2.first_freeze) as p2_first_fall,
avg(p2.last_freeze) as p2_last_spring
from season1 as p1 JOIN season2 as p2 on (p1.station = p2.station)
GROUP by p1.station)
SELECT station, ST_X(geom) as lon, ST_Y(geom) as lat,
d.* from agg d JOIN stations t ON (d.station = t.id)
WHERE t.network in ('IACLIMATE', 'NDCLIMATE', 'SDCLIMATE', 'NECLIMATE',
'KSCLIMATE', 'MOCLIMATE', 'ILCLIMATE', 'WICLIMATE', 'MNCLIMATE',
'MICLIMATE', 'INCLIMATE', 'OHCLIMATE', 'KYCLIMATE')
and substr(station, 3, 1) != 'C' and substr(station, 3, 4) != '0000'
""",
pgconn,
params=[p1syear, p1eyear, p2syear, p2eyear],
index_col='station')
df['p1_season'] = df['p1_first_fall'] - df['p1_last_spring']
df['p2_season'] = df['p2_first_fall'] - df['p2_last_spring']
df['season_delta'] = df['p2_season'] - df['p1_season']
df['spring_delta'] = df['p2_last_spring'] - df['p1_last_spring']
df['fall_delta'] = df['p2_first_fall'] - df['p1_first_fall']
# Reindex so that most extreme values are first
df = df.reindex(df[varname +
'_delta'].abs().sort_values(ascending=False).index)
title = PDICT3[varname]
m = MapPlot(sector=sector,
state=state,
axisbg='white',
title=('%.0f-%.0f minus %.0f-%.0f %s Difference') %
(p2syear, p2eyear, p1syear, p1eyear, title),
subtitle=('based on IEM Archives'),
titlefontsize=14)
# Create 9 levels centered on zero
abval = df[varname + '_delta'].abs().max()
levels = centered_bins(abval)
if opt in ['both', 'contour']:
m.contourf(df['lon'].values,
df['lat'].values,
df[varname + '_delta'].values,
levels,
cmap=plt.get_cmap('seismic'),
units='days')
if sector == 'state':
m.drawcounties()
if opt in ['both', 'values']:
m.plot_values(df['lon'].values,
df['lat'].values,
df[varname + '_delta'].values,
fmt='%.1f',
labelbuffer=5)
return m.fig, df
def plotter(fdict):
""" Go """
pgconn = get_dbconn('coop')
ctx = get_autoplot_context(fdict, get_description())
state = ctx['state']
varname = ctx['var']
sector = ctx['sector']
threshold = ctx['threshold']
opt = ctx['opt']
month = ctx['month']
p1syear = ctx['p1syear']
p1eyear = ctx['p1eyear']
p1yearreq = (p1eyear - p1syear)
p2syear = ctx['p2syear']
p2eyear = ctx['p2eyear']
p2yearreq = (p2eyear - p2syear)
opt1 = ctx['opt1']
if month == 'all':
months = range(1, 13)
elif month == 'fall':
months = [9, 10, 11]
elif month == 'winter':
months = [12, 1, 2]
elif month == 'spring':
months = [3, 4, 5]
elif month == 'summer':
months = [6, 7, 8]
elif month == 'gs':
months = [5, 6, 7, 8, 9]
else:
ts = datetime.datetime.strptime("2000-"+month+"-01", '%Y-%b-%d')
# make sure it is length two for the trick below in SQL
months = [ts.month]
table = "alldata"
if sector == 'state':
# optimization
table = "alldata_%s" % (state,)
df = read_sql("""
WITH period1 as (
SELECT station, year, sum(precip) as total_precip,
avg((high+low) / 2.) as avg_temp, avg(high) as avg_high,
avg(low) as avg_low,
sum(gddxx(50, 86, high, low)) as sum_gdd,
sum(case when high > 86 then high - 86 else 0 end) as sum_sdd,
sum(case when high >= %s then 1 else 0 end) as days_high_above
from """ + table + """ WHERE year >= %s and year < %s
and month in %s GROUP by station, year),
period2 as (
SELECT station, year, sum(precip) as total_precip,
avg((high+low) / 2.) as avg_temp, avg(high) as avg_high,
avg(low) as avg_low,
sum(gddxx(50, 86, high, low)) as sum_gdd,
sum(case when high > 86 then high - 86 else 0 end) as sum_sdd,
sum(case when high >= %s then 1 else 0 end) as days_high_above
from """ + table + """ WHERE year >= %s and year < %s
and month in %s GROUP by station, year),
p1agg as (
SELECT station, avg(total_precip) as precip,
avg(avg_temp) as avg_temp, avg(avg_high) as avg_high,
avg(avg_low) as avg_low, avg(sum_sdd) as sdd,
avg(sum_gdd) as gdd,
avg(days_high_above) as avg_days_high_above,
count(*) as count
from period1 GROUP by station),
p2agg as (
SELECT station, avg(total_precip) as precip,
avg(avg_temp) as avg_temp, avg(avg_high) as avg_high,
avg(avg_low) as avg_low, avg(sum_sdd) as sdd,
avg(sum_gdd) as gdd,
avg(days_high_above) as avg_days_high_above,
count(*) as count
from period2 GROUP by station),
agg as (
SELECT p2.station,
p2.precip as p2_total_precip, p1.precip as p1_total_precip,
p2.gdd as p2_gdd, p1.gdd as p1_gdd,
p2.sdd as p2_sdd, p1.sdd as p1_sdd,
p2.avg_temp as p2_avg_temp, p1.avg_temp as p1_avg_temp,
p1.avg_high as p1_avg_high, p2.avg_high as p2_avg_high,
p1.avg_low as p1_avg_low, p2.avg_low as p2_avg_low,
p1.avg_days_high_above as p1_days_high_above,
p2.avg_days_high_above as p2_days_high_above
from p1agg p1 JOIN p2agg p2 on
(p1.station = p2.station)
WHERE p1.count >= %s and p2.count >= %s)
SELECT station, ST_X(geom) as lon, ST_Y(geom) as lat,
d.* from agg d JOIN stations t ON (d.station = t.id)
WHERE t.network ~* 'CLIMATE'
and substr(station, 3, 1) != 'C' and substr(station, 3, 4) != '0000'
""", pgconn, params=[threshold, p1syear, p1eyear, tuple(months),
threshold, p2syear, p2eyear, tuple(months),
p1yearreq, p2yearreq],
index_col=None)
df['total_precip'] = df['p2_total_precip'] - df['p1_total_precip']
df['avg_temp'] = df['p2_avg_temp'] - df['p1_avg_temp']
df['avg_high'] = df['p2_avg_high'] - df['p1_avg_high']
df['avg_low'] = df['p2_avg_low'] - df['p1_avg_low']
df['gdd'] = df['p2_gdd'] - df['p1_gdd']
df['sdd'] = df['p2_sdd'] - df['p1_sdd']
df['days_high_above'] = (df['p2_days_high_above'] -
df['p1_days_high_above'])
column = varname
title = "%s %s" % (MDICT[month], PDICT3[varname])
title = title.replace("[Threshold]", '%.1f' % (threshold,))
if opt1 == 'p1':
column = 'p1_%s' % (varname,)
title = '%.0f-%.0f %s' % (p1syear, p1eyear, title)
else:
title = ('%.0f-%.0f minus %.0f-%.0f %s Difference'
) % (p2syear, p2eyear, p1syear, p1eyear, title)
# Reindex so that most extreme values are first
df = df.reindex(df[column].abs().sort_values(ascending=False).index)
# drop 5% most extreme events, too much?
df2 = df.iloc[int(len(df.index) * 0.05):]
mp = MapPlot(sector=sector, state=state, axisbg='white', title=title,
subtitle=('based on IEM Archives'),
titlefontsize=12)
if opt1 == 'diff':
# Create 9 levels centered on zero
abval = df2[column].abs().max()
levels = centered_bins(abval)
else:
levels = [round(v, PRECISION[varname])
for v in np.percentile(df2[column].values, range(0, 101,
10))]
if opt in ['both', 'contour']:
mp.contourf(df2['lon'].values, df2['lat'].values,
df2[column].values, levels,
cmap=plt.get_cmap(('seismic_r' if varname == 'total_precip'
else 'seismic')),
units=UNITS[varname])
if sector == 'state':
mp.drawcounties()
if opt in ['both', 'values']:
mp.plot_values(df2['lon'].values, df2['lat'].values,
df2[column].values,
fmt='%%.%if' % (PRECISION[varname],), labelbuffer=5)
return mp.fig, df
def plotter(fdict):
""" Go """
pgconn = get_dbconn("coop")
ctx = get_autoplot_context(fdict, get_description())
state = ctx["state"][:2]
sector = ctx["sector"]
opt = ctx["opt"]
p1syear = ctx["p1syear"]
p1eyear = ctx["p1eyear"]
p2syear = ctx["p2syear"]
p2eyear = ctx["p2eyear"]
varname = ctx["var"]
table = "alldata"
if sector == "state":
table = "alldata_%s" % (state,)
df = read_sql(
"""
WITH season1 as (
SELECT station, year,
min(case when month > 7 and low < 32 then
extract(doy from day) else 366 end) as first_freeze,
max(case when month < 7 and low < 32 then
extract(doy from day) else 0 end) as last_freeze
from """
+ table
+ """ WHERE
year >= %s and year <= %s GROUP by station, year),
season2 as (
SELECT station, year,
min(case when month > 7 and low < 32 then
extract(doy from day) else 366 end) as first_freeze,
max(case when month < 7 and low < 32 then
extract(doy from day) else 0 end) as last_freeze
from """
+ table
+ """ WHERE
year >= %s and year <= %s GROUP by station, year),
agg as (
SELECT p1.station, avg(p1.first_freeze) as p1_first_fall,
avg(p1.last_freeze) as p1_last_spring,
avg(p2.first_freeze) as p2_first_fall,
avg(p2.last_freeze) as p2_last_spring
from season1 as p1 JOIN season2 as p2 on (p1.station = p2.station)
GROUP by p1.station)
SELECT station, ST_X(geom) as lon, ST_Y(geom) as lat,
d.* from agg d JOIN stations t ON (d.station = t.id)
WHERE t.network ~* 'CLIMATE'
and substr(station, 3, 1) != 'C' and substr(station, 3, 4) != '0000'
""",
pgconn,
params=[p1syear, p1eyear, p2syear, p2eyear],
index_col="station",
)
if df.empty:
raise NoDataFound("No Data Found")
df["p1_season"] = df["p1_first_fall"] - df["p1_last_spring"]
df["p2_season"] = df["p2_first_fall"] - df["p2_last_spring"]
df["season_delta"] = df["p2_season"] - df["p1_season"]
df["spring_delta"] = df["p2_last_spring"] - df["p1_last_spring"]
df["fall_delta"] = df["p2_first_fall"] - df["p1_first_fall"]
# Reindex so that most extreme values are first
df = df.reindex(
df[varname + "_delta"].abs().sort_values(ascending=False).index
)
title = PDICT3[varname]
mp = MapPlot(
sector=sector,
state=state,
axisbg="white",
title=("%.0f-%.0f minus %.0f-%.0f %s Difference")
% (p2syear, p2eyear, p1syear, p1eyear, title),
subtitle=("based on IEM Archives"),
titlefontsize=14,
)
# Create 9 levels centered on zero
abval = df[varname + "_delta"].abs().max()
levels = centered_bins(abval)
if opt in ["both", "contour"]:
mp.contourf(
df["lon"].values,
df["lat"].values,
df[varname + "_delta"].values,
levels,
cmap=get_cmap(ctx["cmap"]),
units="days",
)
if sector == "state":
mp.drawcounties()
if opt in ["both", "values"]:
mp.plot_values(
df["lon"].values,
df["lat"].values,
df[varname + "_delta"].values,
fmt="%.1f",
labelbuffer=5,
)
return mp.fig, df
def plotter(fdict):
""" Go """
pgconn = get_dbconn("coop")
ctx = get_autoplot_context(fdict, get_description())
state = ctx["state"]
varname = ctx["var"]
sector = ctx["sector"]
threshold = ctx["threshold"]
opt = ctx["opt"]
month = ctx["month"]
p1syear = ctx["p1syear"]
p1eyear = ctx["p1eyear"]
p1yearreq = p1eyear - p1syear
p2syear = ctx["p2syear"]
p2eyear = ctx["p2eyear"]
p2yearreq = p2eyear - p2syear
opt1 = ctx["opt1"]
if month == "all":
months = range(1, 13)
elif month == "fall":
months = [9, 10, 11]
elif month == "winter":
months = [12, 1, 2]
elif month == "spring":
months = [3, 4, 5]
elif month == "summer":
months = [6, 7, 8]
elif month == "gs":
months = [5, 6, 7, 8, 9]
else:
ts = datetime.datetime.strptime("2000-" + month + "-01", "%Y-%b-%d")
# make sure it is length two for the trick below in SQL
months = [ts.month]
table = "alldata"
if sector == "state":
# optimization
table = "alldata_%s" % (state, )
df = read_sql(
"""
WITH period1 as (
SELECT station, year, sum(precip) as total_precip,
avg((high+low) / 2.) as avg_temp, avg(high) as avg_high,
avg(low) as avg_low,
sum(gddxx(50, 86, high, low)) as sum_gdd,
sum(case when high > 86 then high - 86 else 0 end) as sum_sdd,
sum(case when high >= %s then 1 else 0 end) as days_high_above
from """ + table + """ WHERE year >= %s and year < %s
and month in %s GROUP by station, year),
period2 as (
SELECT station, year, sum(precip) as total_precip,
avg((high+low) / 2.) as avg_temp, avg(high) as avg_high,
avg(low) as avg_low,
sum(gddxx(50, 86, high, low)) as sum_gdd,
sum(case when high > 86 then high - 86 else 0 end) as sum_sdd,
sum(case when high >= %s then 1 else 0 end) as days_high_above
from """ + table + """ WHERE year >= %s and year < %s
and month in %s GROUP by station, year),
p1agg as (
SELECT station, avg(total_precip) as precip,
avg(avg_temp) as avg_temp, avg(avg_high) as avg_high,
avg(avg_low) as avg_low, avg(sum_sdd) as sdd,
avg(sum_gdd) as gdd,
avg(days_high_above) as avg_days_high_above,
count(*) as count
from period1 GROUP by station),
p2agg as (
SELECT station, avg(total_precip) as precip,
avg(avg_temp) as avg_temp, avg(avg_high) as avg_high,
avg(avg_low) as avg_low, avg(sum_sdd) as sdd,
avg(sum_gdd) as gdd,
avg(days_high_above) as avg_days_high_above,
count(*) as count
from period2 GROUP by station),
agg as (
SELECT p2.station,
p2.precip as p2_total_precip, p1.precip as p1_total_precip,
p2.gdd as p2_gdd, p1.gdd as p1_gdd,
p2.sdd as p2_sdd, p1.sdd as p1_sdd,
p2.avg_temp as p2_avg_temp, p1.avg_temp as p1_avg_temp,
p1.avg_high as p1_avg_high, p2.avg_high as p2_avg_high,
p1.avg_low as p1_avg_low, p2.avg_low as p2_avg_low,
p1.avg_days_high_above as p1_days_high_above,
p2.avg_days_high_above as p2_days_high_above
from p1agg p1 JOIN p2agg p2 on
(p1.station = p2.station)
WHERE p1.count >= %s and p2.count >= %s)
SELECT station, ST_X(geom) as lon, ST_Y(geom) as lat,
d.* from agg d JOIN stations t ON (d.station = t.id)
WHERE t.network ~* 'CLIMATE'
and substr(station, 3, 1) != 'C' and substr(station, 3, 4) != '0000'
""",
pgconn,
params=[
threshold,
p1syear,
p1eyear,
tuple(months),
threshold,
p2syear,
p2eyear,
tuple(months),
p1yearreq,
p2yearreq,
],
index_col=None,
)
if df.empty:
raise NoDataFound("No Data Found.")
df["total_precip"] = df["p2_total_precip"] - df["p1_total_precip"]
df["avg_temp"] = df["p2_avg_temp"] - df["p1_avg_temp"]
df["avg_high"] = df["p2_avg_high"] - df["p1_avg_high"]
df["avg_low"] = df["p2_avg_low"] - df["p1_avg_low"]
df["gdd"] = df["p2_gdd"] - df["p1_gdd"]
df["sdd"] = df["p2_sdd"] - df["p1_sdd"]
df["days_high_above"] = df["p2_days_high_above"] - df["p1_days_high_above"]
column = varname
title = "%s %s" % (MDICT[month], PDICT3[varname])
title = title.replace("[Threshold]", "%.1f" % (threshold, ))
if opt1 == "p1":
column = "p1_%s" % (varname, )
title = "%.0f-%.0f %s" % (p1syear, p1eyear, title)
else:
title = ("%.0f-%.0f minus %.0f-%.0f %s Difference") % (
p2syear,
p2eyear,
p1syear,
p1eyear,
title,
)
# Reindex so that most extreme values are first
df = df.reindex(df[column].abs().sort_values(ascending=False).index)
# drop 5% most extreme events, too much?
df2 = df.iloc[int(len(df.index) * 0.05):]
mp = MapPlot(
sector=sector,
state=state,
axisbg="white",
title=title,
subtitle=("based on IEM Archives"),
titlefontsize=12,
)
if opt1 == "diff":
# Create 9 levels centered on zero
abval = df2[column].abs().max()
levels = centered_bins(abval)
else:
levels = [
round(v, PRECISION[varname])
for v in np.percentile(df2[column].values, range(0, 101, 10))
]
if opt in ["both", "contour"]:
mp.contourf(
df2["lon"].values,
df2["lat"].values,
df2[column].values,
levels,
cmap=plt.get_cmap(ctx["cmap"]),
units=UNITS[varname],
)
if sector == "state":
mp.drawcounties()
if opt in ["both", "values"]:
mp.plot_values(
df2["lon"].values,
df2["lat"].values,
df2[column].values,
fmt="%%.%if" % (PRECISION[varname], ),
labelbuffer=5,
)
return mp.fig, df
def plotter(fdict):
""" Go """
import matplotlib
matplotlib.use('agg')
import matplotlib.pyplot as plt
pgconn = psycopg2.connect(database='coop', host='iemdb', user='******')
ctx = get_autoplot_context(fdict, get_description())
state = ctx['state'][:2]
sector = ctx['sector']
opt = ctx['opt']
p1syear = ctx['p1syear']
p1eyear = ctx['p1eyear']
p2syear = ctx['p2syear']
p2eyear = ctx['p2eyear']
varname = ctx['var']
table = "alldata"
if sector == 'state':
table = "alldata_%s" % (state, )
df = read_sql("""
WITH season1 as (
SELECT station, year,
min(case when month > 7 and low < 32 then
extract(doy from day) else 366 end) as first_freeze,
max(case when month < 7 and low < 32 then
extract(doy from day) else 0 end) as last_freeze
from """ + table + """ WHERE
year >= %s and year <= %s GROUP by station, year),
season2 as (
SELECT station, year,
min(case when month > 7 and low < 32 then
extract(doy from day) else 366 end) as first_freeze,
max(case when month < 7 and low < 32 then
extract(doy from day) else 0 end) as last_freeze
from """ + table + """ WHERE
year >= %s and year <= %s GROUP by station, year),
agg as (
SELECT p1.station, avg(p1.first_freeze) as p1_first_fall,
avg(p1.last_freeze) as p1_last_spring,
avg(p2.first_freeze) as p2_first_fall,
avg(p2.last_freeze) as p2_last_spring
from season1 as p1 JOIN season2 as p2 on (p1.station = p2.station)
GROUP by p1.station)
SELECT station, ST_X(geom) as lon, ST_Y(geom) as lat,
d.* from agg d JOIN stations t ON (d.station = t.id)
WHERE t.network in ('IACLIMATE', 'NDCLIMATE', 'SDCLIMATE', 'NECLIMATE',
'KSCLIMATE', 'MOCLIMATE', 'ILCLIMATE', 'WICLIMATE', 'MNCLIMATE',
'MICLIMATE', 'INCLIMATE', 'OHCLIMATE', 'KYCLIMATE')
and substr(station, 3, 1) != 'C' and substr(station, 3, 4) != '0000'
""", pgconn, params=[p1syear, p1eyear,
p2syear, p2eyear],
index_col='station')
df['p1_season'] = df['p1_first_fall'] - df['p1_last_spring']
df['p2_season'] = df['p2_first_fall'] - df['p2_last_spring']
df['season_delta'] = df['p2_season'] - df['p1_season']
df['spring_delta'] = df['p2_last_spring'] - df['p1_last_spring']
df['fall_delta'] = df['p2_first_fall'] - df['p1_first_fall']
# Reindex so that most extreme values are first
df = df.reindex(df[varname + '_delta'].abs().sort_values(
ascending=False).index)
title = PDICT3[varname]
m = MapPlot(sector=sector, state=state, axisbg='white',
title=('%.0f-%.0f minus %.0f-%.0f %s Difference'
) % (p2syear, p2eyear, p1syear, p1eyear, title),
subtitle=('based on IEM Archives'),
titlefontsize=14)
# Create 9 levels centered on zero
abval = df[varname + '_delta'].abs().max()
levels = centered_bins(abval)
if opt in ['both', 'contour']:
m.contourf(df['lon'].values, df['lat'].values,
df[varname + '_delta'].values, levels,
cmap=plt.get_cmap('seismic'),
units='days')
if sector == 'state':
m.drawcounties()
if opt in ['both', 'values']:
m.plot_values(df['lon'].values, df['lat'].values,
df[varname + '_delta'].values,
fmt='%.1f', labelbuffer=5)
return m.fig, df
def plotter(fdict):
""" Go """
import matplotlib
matplotlib.use('agg')
import matplotlib.pyplot as plt
pgconn = psycopg2.connect(database='coop', host='iemdb', user='******')
state = fdict.get('state', 'IA')[:2]
varname = fdict.get('var', 'total_precip')
sector = fdict.get('sector', 'state')
threshold = float(fdict.get('threshold', -99))
opt = fdict.get('opt', 'both')
month = fdict.get('month', 'all')
p1syear = int(fdict.get('p1syear', 1951))
p1eyear = int(fdict.get('p1eyear', 1980))
p2syear = int(fdict.get('p2syear', 1981))
p2eyear = int(fdict.get('p2eyear', 2010))
if month == 'all':
months = range(1, 13)
elif month == 'fall':
months = [9, 10, 11]
elif month == 'winter':
months = [12, 1, 2]
elif month == 'spring':
months = [3, 4, 5]
elif month == 'summer':
months = [6, 7, 8]
elif month == 'gs':
months = [5, 6, 7, 8, 9]
else:
ts = datetime.datetime.strptime("2000-"+month+"-01", '%Y-%b-%d')
# make sure it is length two for the trick below in SQL
months = [ts.month]
table = "alldata"
if sector == 'state':
# optimization
table = "alldata_%s" % (state,)
df = read_sql("""
WITH period1 as (
SELECT station, year, sum(precip) as total_precip,
avg((high+low) / 2.) as avg_temp, avg(high) as avg_high,
avg(low) as avg_low,
sum(gddxx(50, 86, high, low)) as sum_gdd,
sum(case when high > 86 then high - 86 else 0 end) as sum_sdd,
sum(case when high >= %s then 1 else 0 end) as days_high_above
from """ + table + """ WHERE year >= %s and year < %s
and month in %s GROUP by station, year),
period2 as (
SELECT station, year, sum(precip) as total_precip,
avg((high+low) / 2.) as avg_temp, avg(high) as avg_high,
avg(low) as avg_low,
sum(gddxx(50, 86, high, low)) as sum_gdd,
sum(case when high > 86 then high - 86 else 0 end) as sum_sdd,
sum(case when high >= %s then 1 else 0 end) as days_high_above
from """ + table + """ WHERE year >= %s and year < %s
and month in %s GROUP by station, year),
p1agg as (
SELECT station, avg(total_precip) as precip,
avg(avg_temp) as avg_temp, avg(avg_high) as avg_high,
avg(avg_low) as avg_low, avg(sum_sdd) as sdd,
avg(sum_gdd) as gdd,
avg(days_high_above) as avg_days_high_above
from period1 GROUP by station),
p2agg as (
SELECT station, avg(total_precip) as precip,
avg(avg_temp) as avg_temp, avg(avg_high) as avg_high,
avg(avg_low) as avg_low, avg(sum_sdd) as sdd,
avg(sum_gdd) as gdd,
avg(days_high_above) as avg_days_high_above
from period2 GROUP by station),
agg as (
SELECT p2.station, p2.precip as p2_precip, p1.precip as p1_precip,
p2.gdd as p2_gdd, p1.gdd as p1_gdd,
p2.sdd as p2_sdd, p1.sdd as p1_sdd,
p2.avg_temp as p2_avg_temp, p1.avg_temp as p1_avg_temp,
p1.avg_high as p1_avg_high, p2.avg_high as p2_avg_high,
p1.avg_low as p1_avg_low, p2.avg_low as p2_avg_low,
p1.avg_days_high_above as p1_avg_days_high_above,
p2.avg_days_high_above as p2_avg_days_high_above
from p1agg p1 JOIN p2agg p2 on
(p1.station = p2.station))
SELECT station, ST_X(geom) as lon, ST_Y(geom) as lat,
d.* from agg d JOIN stations t ON (d.station = t.id)
WHERE t.network in ('IACLIMATE', 'NDCLIMATE', 'SDCLIMATE', 'NECLIMATE',
'KSCLIMATE', 'MOCLIMATE', 'ILCLIMATE', 'WICLIMATE', 'MNCLIMATE',
'MICLIMATE', 'INCLIMATE', 'OHCLIMATE', 'KYCLIMATE')
and substr(station, 3, 1) != 'C' and substr(station, 3, 4) != '0000'
""", pgconn, params=[threshold, p1syear, p1eyear, tuple(months),
threshold, p2syear, p2eyear, tuple(months)],
index_col=None)
df['total_precip'] = df['p2_precip'] - df['p1_precip']
df['avg_temp'] = df['p2_avg_temp'] - df['p1_avg_temp']
df['avg_high'] = df['p2_avg_high'] - df['p1_avg_high']
df['avg_low'] = df['p2_avg_low'] - df['p1_avg_low']
df['gdd'] = df['p2_gdd'] - df['p1_gdd']
df['sdd'] = df['p2_sdd'] - df['p1_sdd']
df['days_high_above'] = (df['p2_avg_days_high_above'] -
df['p1_avg_days_high_above'])
# Reindex so that most extreme values are first
df = df.reindex(df[varname].abs().sort_values(ascending=False).index)
# drop 5% most extreme events, too much?
df2 = df.iloc[int(len(df.index) * 0.05):]
title = "%s %s" % (MDICT[month], PDICT3[varname])
title = title.replace("[Threshold]", '%.1f' % (threshold,))
m = MapPlot(sector=sector, state=state, axisbg='white',
title=('%.0f-%.0f minus %.0f-%.0f %s Difference'
) % (p2syear, p2eyear, p1syear, p1eyear, title),
subtitle=('based on IEM Archives'),
titlefontsize=12)
# Create 9 levels centered on zero
abval = df2[varname].abs().max()
levels = centered_bins(abval)
if opt in ['both', 'contour']:
m.contourf(df2['lon'].values, df2['lat'].values,
df2[varname].values, levels,
cmap=plt.get_cmap(('seismic_r' if varname == 'total_precip'
else 'seismic')),
units=UNITS[varname])
if sector == 'state':
m.drawcounties()
if opt in ['both', 'values']:
m.plot_values(df2['lon'].values, df2['lat'].values,
df2[varname].values,
fmt='%%.%if' % (PRECISION[varname],))
return m.fig, df
