def miss(test_filename, train_filename, param_filename):
fn = test_filename.replace("-data", "-index")
meta = None
if os.path.exists(fn):
with open(fn, "r") as idxfile:
meta = idxfile.readlines()
X, y, df = design_matrix(test_filename, train_filename, get_df=True)
predictor = gen_predictor(param_filename)
probs = predictor(X)
indices = get_misses(probs, y)
for i in indices:
print "----------------"
if meta: print "META:",",".join(meta[i].split(",")).strip()
print df.ix[i]
return indices
def miss(test_filename, train_filename, param_filename):
fn = test_filename.replace("-data", "-index")
meta = None
if os.path.exists(fn):
with open(fn, "r") as idxfile:
meta = idxfile.readlines()
X, y, df = design_matrix(test_filename, train_filename, get_df=True)
predictor = gen_predictor(param_filename)
probs = predictor(X)
indices = get_misses(probs, y)
for i in indices:
print "----------------"
if meta: print "META:", ",".join(meta[i].split(",")).strip()
print df.ix[i]
return indices
def eval(test_filename, train_filename, param_filename, method, **kwparams):
X, y = design_matrix(test_filename, train_filename)
predictor = gen_predictor(param_filename)
probs = predictor(X)
return run_eval(probs, y, method, **kwparams)
def pred(test_filename, train_filename, param_filename):
X, y = design_matrix(test_filename, train_filename)
predictor = gen_predictor(param_filename)
probs = predictor(X)
return probs, y
def calc_risk(self):
if "AGE" in values and values["AGE"]:
age = 30
try:
age = float(values["AGE"])
except ValueError:
pass
except TypeError:
pass
if age < 10 or 50 < age:
res_scr.curr_risk_color = [
153.0 / 255, 93.0 / 255, 77.0 / 255, 1
]
res_scr.curr_risk_label = 'HIGH RISK'
res_scr.curr_risk_level = 1
sm.current = 'result'
return
# Find highest ranking model that contained in the provided variables
model_dir = None
model_vars = None
vv = set([])
for k in values:
if values[k]: vv.add(k)
print vv
for info in models_info:
v = set(info[1])
res = v.issubset(vv)
#print res, info[1]
if res:
model_dir = info[0]
model_vars = info[1]
break
if not model_dir or not models_info:
res_scr.curr_risk_color = [0.5, 0.5, 0.5, 1]
res_scr.curr_risk_label = 'INSUFFICIENT DATA'
res_scr.curr_risk_level = 0
sm.transition = RiseInTransition()
sm.current = 'result'
return
print "FOUND MODEL !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!"
print model_dir
print model_vars
predictor = gen_predictor(os.path.join(model_dir, 'nnet-params'))
N = len(model_vars)
model_min = []
model_max = []
with open(os.path.join(model_dir, 'bounds.txt')) as bfile:
lines = bfile.readlines()
for line in lines:
line = line.strip()
parts = line.split()
model_min.append(float(parts[1]))
model_max.append(float(parts[2]))
print model_min
print model_max
v = [None] * (N + 1)
v[0] = 1
for i in range(N):
var = model_vars[i]
if var in values:
try:
v[i + 1] = float(values[var])
if var in units and units[var] != var_def_unit[var]:
# Need to convert units
c, f = var_unit_conv[var]
print "convert", var, v[i +
1], "->", f * (v[i + 1] + c)
v[i + 1] = f * (v[i + 1] + c)
except ValueError:
pass
except TypeError:
pass
if None in v:
res_scr.curr_risk_color = [0.5, 0.5, 0.5, 1]
res_scr.curr_risk_label = 'INSUFFICIENT DATA'
res_scr.curr_risk_level = 0
sm.current = 'result'
return
for i in range(N):
f = (v[i + 1] - model_min[i]) / (model_max[i] - model_min[i])
if f < 0: v[i + 1] = 0
elif 1 < f: v[i + 1] = 1
else: v[i + 1] = f
print values
print v
X = np.array([v])
probs = predictor(X)
pred = probs[0]
print "------------->", pred, type(pred)
res_scr.curr_risk_level = float(pred)
if pred < 0.5:
res_scr.curr_risk_color = [
121.0 / 255, 192.0 / 255, 119.0 / 255, 1
]
res_scr.curr_risk_label = 'LOW RISK'
else:
level = float((pred - 0.5) / 0.5)
res_scr.curr_risk_color = [153.0 / 255, 93.0 / 255, 77.0 / 255, 1]
res_scr.curr_risk_label = 'HIGH RISK'
sm.transition = RiseInTransition()
#sm.transition = SlideTransition(direction='left')
sm.current = 'result'
def eval(test_filename, train_filename, param_filename, method, **kwparams):
X, y = design_matrix(test_filename, train_filename)
predictor = gen_predictor(param_filename)
probs = predictor(X)
return run_eval(probs, y, method, **kwparams)
def pred(test_filename, train_filename, param_filename):
X, y = design_matrix(test_filename, train_filename)
predictor = gen_predictor(param_filename)
probs = predictor(X)
return probs, y
model_vars = None
for info in models_info:
v = set(info[1])
res = v.issubset(vv)
#print res, info[1]
if res:
model_dir = info[0]
model_vars = info[1]
break
if not model_dir or not models_info:
continue
count += 1
predictor = gen_predictor(os.path.join(model_dir, 'nnet-params'))
N = len(model_vars)
model_min = []
model_max = []
with open(os.path.join(model_dir, 'bounds.txt')) as bfile:
lines = bfile.readlines()
for line in lines:
line = line.strip()
parts = line.split()
model_min.append(float(parts[1]))
model_max.append(float(parts[2]))
# print model_min
# print model_max
def calc_risk(self):
if "AGE" in values and values["AGE"]:
age = 30
try:
age = float(values["AGE"])
except ValueError:
pass
except TypeError:
pass
if age < 10 or 50 < age:
res_scr.curr_risk_color = [153.0/255, 93.0/255, 77.0/255, 1]
res_scr.curr_risk_label = 'HIGH RISK'
res_scr.curr_risk_level = 1
sm.current = 'result'
return
# Find highest ranking model that contained in the provided variables
model_dir = None
model_vars = None
vv = set([])
for k in values:
if values[k]: vv.add(k)
print vv
for info in models_info:
v = set(info[1])
res = v.issubset(vv)
#print res, info[1]
if res:
model_dir = info[0]
model_vars = info[1]
break
if not model_dir or not models_info:
res_scr.curr_risk_color = [0.5, 0.5, 0.5, 1]
res_scr.curr_risk_label = 'INSUFFICIENT DATA'
res_scr.curr_risk_level = 0
sm.transition = RiseInTransition()
sm.current = 'result'
return
print "FOUND MODEL !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!"
print model_dir
print model_vars
predictor = gen_predictor(os.path.join(model_dir, 'nnet-params'))
N = len(model_vars)
model_min = []
model_max = []
with open(os.path.join(model_dir, 'bounds.txt')) as bfile:
lines = bfile.readlines()
for line in lines:
line = line.strip()
parts = line.split()
model_min.append(float(parts[1]))
model_max.append(float(parts[2]))
print model_min
print model_max
v = [None] * (N + 1)
v[0] = 1
for i in range(N):
var = model_vars[i]
if var in values:
try:
v[i + 1] = float(values[var])
if var in units and units[var] != var_def_unit[var]:
# Need to convert units
c, f = var_unit_conv[var]
print "convert",var,v[i + 1],"->",f*(v[i + 1] + c)
v[i + 1] = f * (v[i + 1] + c)
except ValueError:
pass
except TypeError:
pass
if None in v:
res_scr.curr_risk_color = [0.5, 0.5, 0.5, 1]
res_scr.curr_risk_label = 'INSUFFICIENT DATA'
res_scr.curr_risk_level = 0
sm.current = 'result'
return
for i in range(N):
f = (v[i + 1] - model_min[i]) / (model_max[i] - model_min[i])
if f < 0: v[i + 1] = 0
elif 1 < f: v[i + 1] = 1
else: v[i + 1] = f
print values
print v
X = np.array([v])
probs = predictor(X)
pred = probs[0]
print "------------->",pred,type(pred)
res_scr.curr_risk_level = float(pred)
if pred < 0.5:
res_scr.curr_risk_color = [121.0/255, 192.0/255, 119.0/255, 1]
res_scr.curr_risk_label = 'LOW RISK'
else:
level = float((pred - 0.5) / 0.5)
res_scr.curr_risk_color = [153.0/255, 93.0/255, 77.0/255, 1]
res_scr.curr_risk_label = 'HIGH RISK'
sm.transition = RiseInTransition()
#sm.transition = SlideTransition(direction='left')
sm.current = 'result'