def check_wiktionary_language(target_lang):
seen_codes = {}
for lang_name in LANGUAGE_NAMES[target_lang]:
if lang_name.startswith('Proto-'):
lang_name = lang_name[6:]
code = str(find_name('language', lang_name, target_lang))
assert code not in seen_codes, \
"%r and %r have the same code" % (seen_codes[code], lang_name)
seen_codes[code] = lang_name
def CQS_match_query_phrase(self, phrase):
"""Analyze phrase to see if it is a play-able phrase with this skill.
Needs to be implemented by the skill.
Arguments:
phrase (str): User phrase, "What is an aardwark"
Returns:
(match, CQSMatchLevel[, callback_data]) or None: Tuple containing
a string with the appropriate matching phrase, the PlayMatch
type, and optionally data to return in the callback if the
match is selected.
"""
response = None
match = self.intents.calc_intent(phrase)
level = CQSMatchLevel.CATEGORY
data = match.matches
intent = match.name
score = match.conf
data["intent"] = intent
data["score"] = score
if score > 0.8:
level = CQSMatchLevel.EXACT
elif score > 0.5:
level = CQSMatchLevel.CATEGORY
elif score > 0.3:
level = CQSMatchLevel.GENERAL
else:
intent = None
if intent:
# Validate extracted entities
country = data.get("country")
region = data.get("region")
language = data.get("language")
if country:
data["query"] = country
# ensure we really have a country name
response = self.dialog_renderer.render("bad_country", {})
match, score = match_one(country.lower(),
list(self.countries_data.keys()))
self.log.debug("Country fuzzy match: {n}, Score: {s}".format(
n=match, s=score))
if score > 0.5:
country = match
data.update(self.countries_data[country])
else:
countries = self.search_country(country)
if not len(countries) > 0:
level = CQSMatchLevel.GENERAL
else:
country = countries[0]["name"]
data.update(countries[0])
# TODO disambiguation
if len(countries) > 1:
data["disambiguation"] = countries[1:]
self.log.debug("multiple matches found: " +
str([c["name"] for c in countries]))
data["country"] = country # normalized from match
if language:
data["query"] = language
# ensure we really have a language name
words = language.split(" ")
clean_up = ["is"]
# remove words commonly caught by mistake in padatious
language = " ".join(
[word for word in words if word not in clean_up])
lang_code = langcodes.find_name(
'language', language, langcodes.standardize_tag(self.lang))
lang_code = str(lang_code)
self.log.debug("Detected lang code: " + lang_code)
if not lang_code:
return None
data["lang_code"] = lang_code
# TODO
countries = self.search_country_by_language(lang_code)
data["country_list"] = countries
if region:
data["query"] = region
# ensure we really have a region name
response = self.dialog_renderer.render("bad_region")
countries = None
match, score = match_one(region, self.regions)
data["region_score"] = score
if score > 0.5:
region = match
countries = self.search_country_by_region(region)
match, score2 = match_one(region, self.subregions)
data["subregion_score"] = score2
if score2 > score:
region = match
countries = self.search_country_by_subregion(region)
if score > 0.8 and not country:
level = CQSMatchLevel.EXACT
elif score > 0.5 and not country:
level = CQSMatchLevel.CATEGORY
elif score > 0.3 and not country:
level = CQSMatchLevel.GENERAL
data["region"] = region
self.log.debug("Detected region: " + region)
data["country_list"] = countries
# Get response from intents
response = self.intent2answer(intent, data) or response
if response:
return (phrase, level, response, data)
return None
def _language_name_to_code(name, name_language_code):
try:
found = langcodes.find_name('language', name, name_language_code)
return str(found)
except LookupError:
return None
def _language_name_to_code(name, name_language_code):
try:
found = langcodes.find_name('language', name, name_language_code)
return str(found)
except LookupError:
return None
