def do_report(index, k):
b1_data = read_res_file(B1_RES_FILE)
a2_data = read_res_file(A2_RES_FILE)
process_k(k, b1_data, a2_data, index)
def main(index = DEFAULT_INDEX):
b1_res = read_res_file(B1_RES_FILE)
a2_res = read_res_file(A2_RES_FILE)
if len(b1_res) and len(a2_res):
cldat = ClDat()
if cldat.loaded:
b1_hist = gen_hist(b1_res, cldat.b1, HIST_FILE_B1, TYPE_1_COL)
a2_hist = gen_hist(a2_res, cldat.a2, HIST_FILE_A2, TYPE_2_COL)
if index != DEFAULT_INDEX:
success, k, pro, pre, p = load_local_data(index, cl.b1, cl.a2)
if success:
generate_final_hist(index, b1_hist, a2_hist, k, pro, pre, p, cldat)
else:
print "ERROR: Generation of historical not possible, local data not available."
else:
print "ERROR: Generation of historical not possible, cl not available."
else:
print "ERROR: Generation of historical not possible, res not available."
def do_report(index, k, cl, pred_rf, pre, ex_mean):
print "Generating report ..."
b1_data = read_res_file(B1_RES_FILE)
a2_data = read_res_file(A2_RES_FILE)
process_k(k, b1_data, a2_data, cl, index, pred_rf, pre, ex_mean)
def generate_hist(index, cl, k, pro, pre, p):
b1_res = read_res_file(B1_RES_FILE)
a2_res = read_res_file(A2_RES_FILE)
if len(b1_res) and len(a2_res):
b1_hist = gen_hist(b1_res, cl.b1, HIST_FILE_B1, TYPE_1_COL)
a2_hist = gen_hist(a2_res, cl.a2, HIST_FILE_A2, TYPE_2_COL)
generate_final_hist(index, b1_hist, a2_hist, k, pro, pre, p, cl)
else:
print "ERROR: Generation of historical not possible, res not available."
def generate_hist(index, cl, k, pro, pre, p):
b1_res = read_res_file(B1_RES_FILE)
a2_res = read_res_file(A2_RES_FILE)
if len(b1_res) and len(a2_res):
b1_hist = gen_hist(b1_res, cl.b1, HIST_FILE_B1, TYPE_1_COL)
a2_hist = gen_hist(a2_res, cl.a2, HIST_FILE_A2, TYPE_2_COL)
generate_final_hist(index, b1_hist, a2_hist, k, pro, pre, p, cl)
else:
print "ERROR: Generation of historical not possible, res not available."
def _calc_from_res(self, res_file_name, cl_data):
final_dif = {}
cl_dict = self._dict_from_cl(cl_data)
res = read_res_file(res_file_name)
for r in res:
name1 = r[R_NAME_1_COL]
name2 = r[R_NAME_2_COL]
m = r[R_M_COL]
dif = cl_dict[name1] - cl_dict[name2]
sum = SUM_DIF_POS[m]
try:
val = final_dif[dif]
new_val = [ val[i] + sum[i] for i in range(len(val))]
final_dif.update( {dif: new_val} )
except KeyError:
final_dif.update( {dif: sum} )
return final_dif
def main():
clda = ClDat()
if clda.loaded:
b1_res = read_res_file(B1_RES_FILE)
a2_res = read_res_file(A2_RES_FILE)
if len(b1_res) and len(a2_res):
mdls_b1 = evaluate_all_models(clda.b1, b1_res)
mdls_a2 = evaluate_all_models(clda.a2, a2_res)
save_data_to_csv(MODELS_FILENAME, mdls_b1 + mdls_a2)
else:
print "Res data couldn't be read."
else:
print "Cl data couldn't be loaded."
def _generate(self, force_calc):
pre_file_name = PRE_FILE_NAME_PREFIX + self._index + INPUT_FILE_NAME_EXT
if not force_calc:
self._pre = read_input_file(pre_file_name)
if not len(self._pre):
b1_res = read_res_file(B1_RES_FILE)
a2_res = read_res_file(A2_RES_FILE)
for k in self._k:
if k[TYPE_COL] == TYPE_1_COL:
cl_data = self._b1
res_data = b1_res
else:
cl_data = self._a2
res_data = a2_res
lo_data, lo_target_data = Pre.get_data_for_pre(k[NAME_LO_COL],
cl_data,
res_data, True)
lo_mdl = Pre.get_mdl(k[NAME_LO_COL])
vi_data, vi_target_data = Pre.get_data_for_pre(k[NAME_VI_COL],
cl_data,
res_data, False)
vi_mdl = Pre.get_mdl(k[NAME_VI_COL])
print "Predicting: %s - %s" % (k[NAME_LO_COL], k[NAME_VI_COL])
lo_pre = self.get_pre_values(lo_data, lo_target_data,
lo_mdl.lo_mdls)
vi_pre = self.get_pre_values(vi_data, vi_target_data,
vi_mdl.vi_mdls)
self._pre.append(combine_lo_vi(lo_pre, vi_pre))
if self.generated:
save_data_to_csv(pre_file_name, self._pre)
def _generate(self, force_calc):
pre_file_name = PRE_FILE_NAME_PREFIX + self._index + INPUT_FILE_NAME_EXT
if not force_calc:
self._pre = read_input_file(pre_file_name)
if not len(self._pre):
b1_res = read_res_file(B1_RES_FILE)
a2_res = read_res_file(A2_RES_FILE)
for k in self._k:
if k[TYPE_COL] == TYPE_1_COL:
cl_data = self._b1
res_data = b1_res
else:
cl_data = self._a2
res_data = a2_res
lo_data, lo_target_data = Pre.get_data_for_pre(
k[NAME_LO_COL], cl_data, res_data, True)
lo_mdl = Pre.get_mdl(k[NAME_LO_COL])
vi_data, vi_target_data = Pre.get_data_for_pre(
k[NAME_VI_COL], cl_data, res_data, False)
vi_mdl = Pre.get_mdl(k[NAME_VI_COL])
print "Predicting: %s - %s" % (k[NAME_LO_COL], k[NAME_VI_COL])
lo_pre = self.get_pre_values(lo_data, lo_target_data,
lo_mdl.lo_mdls)
vi_pre = self.get_pre_values(vi_data, vi_target_data,
vi_mdl.vi_mdls)
self._pre.append(combine_lo_vi(lo_pre, vi_pre))
if self.generated:
save_data_to_csv(pre_file_name, self._pre)