def HSS(clim_data, hindcast, n_dry = 10, n_wet = 10):
from utils import categorize
from numpy import argsort
#_Categorize the data
clim_cat = categorize(clim_data)
hind_cat = categorize(hindcast)
#_Sort the observations
idx = argsort(clim_data)
#_Create indices for the all, dry, and wet categories
index = { 'all' : idx,
'dry' : idx[:n_dry],
'wet' : idx[-n_wet:]
}
#_Initialize the skill score dictionary
hss = {}
#_Calculate HSS for each category
for cat in index:
total = len(index[cat])
exp_hits = total / 3.
hits = sum(clim_cat[index[cat]] == hind_cat[index[cat]])
hss[cat] = float(hits - exp_hits) / (total - exp_hits)
### Optional printing ###
# print '%s years: total is %i, exp_hits is %.2f, hits is %i, hss is %.2f' \
# % (yrs, total, exp_hits, hits, hss[yrs])
return hss
def _get_yacc_variables(self, module):
frame, variables = get_global_vars(module)
filter_rules = [lambda x: x == "tokens",
lambda x: x.startswith("p_")]
variables = sorted(filter_variables(filter_rules, variables),
key=partial(by_appearance, frame))
tokens, production_rules = categorize(filter_rules, variables)
if len(tokens) == 0:
raise SyntaxError("'tokens' (a tuple of strings) must be defined")
return tokens[-1][-1], production_rules # tokens = [('tokens', (...))]
def get_lexer_variables(self, module=None):
frame, variables = get_global_vars(module)
filter_rules = [lambda x: x == "tokens",
lambda x: x == "states",
lambda x: x.startswith("t_")]
variables = sorted(filter_variables(filter_rules, variables),
key=partial(by_appearance, frame))
token_names, states, t_rules = categorize(filter_rules, variables)
if len(token_names) == 0:
raise SyntaxError("'tokens' (a tuple of strings) must be defined")
token_names = token_names[0] if token_names else []
states = states[0] if states else []
return token_names, states, t_rules
insert_item({'id': article.id}, 'errorArticles')
continue
try:
article.download()
article.parse()
title = article.title
title_split = article.title.split('|')
if len(title_split) != 1:
title = title_split[0].strip()
pattern = re.compile(source.brand, re.IGNORECASE)
body = pattern.sub('', article.text)
categories = categorize(body)
try:
# if langdetect.detect(body) != 'en' or langdetect.detect(title) != 'en':
if langdetect.detect(body) != 'en':
if PY_ENV == 'development':
print('\n(NOT ENGLISH) Skipped: ' +
str(article.url) + '\n')
slp_time = insert_log(
source_id, 'articleCrawl', 'error',
float(time.clock() - start_time), {
'articleUrl': article.url,
'articleTitle': title,
'errorMessage': 'NOT ENGLISH',
'crawlerName': 'credible crawler'
})