def best_matches(bigkeys, menu, vector_dict, threshold=90):
"""Matches 2-gram and 3-gram keyword vectors to menu items
Args:
bigkeys (list): a list of 2-grams and 3-grams
menu (dict): a dictionary of a menu for a given restaurant
vector_dict (dict): a dictionary whose keys are keywords and values are vectors
threshold (int): fuzzy-match score required for keyword matching
Returns:
best_matches (dict): a dictionary whose keys menu items and values are tuples of keyword
matches and their corresponding fuzzy match score
"""
best_matches = {}
for name, meals in menu.iteritems():
for meal, course in meals.iteritems():
for food in course:
item = food.replace('-', ' ')
if len(item.split()) > 1 and len(item.split()) < 4:
matches = process.extractBests(item, bigkeys, scorer=fuzz.QRatio, score_cutoff=threshold, limit=3)
matches += process.extractBests(item, bigkeys, scorer=fuzz.UWRatio, score_cutoff=threshold, limit=3)
vectors = []
for match in matches:
vectors.append(vector_dict[match[0]])
best_matches[item] = dict(zip(matches, vectors))
if len(item.split()) > 3:
descriptor = item.split()[-2:]
matches = process.extractBests(item, bigkeys, scorer=fuzz.QRatio, score_cutoff=threshold, limit=3)
matches += process.extractBests(item, bigkeys, scorer=fuzz.UWRatio, score_cutoff=threshold, limit=3)
vectors = []
for match in matches:
vectors.append(vector_dict[match[0]])
best_matches[item] = dict(zip(matches, vectors))
return best_matches
def search(self, data):
books = Book.objects.all()
resID = set([b.id for b in books])
for key in ['title', 'isbn']:
if key in data:
values = [b.__dict__[key] for b in books]
fuzzy = process.extractBests(data[key], values, score_cutoff=60)
fvalues = [val for val, _ in fuzzy]
ids = [b.id for b in books if b.__dict__[key] in fvalues]
resID &= set(ids)
if 'author' in data:
values = [b.author.lname for b in books]
fuzzy = process.extractBests(data['author'], values, score_cutoff=60)
fvalues = [val for val, _ in fuzzy]
ids = [b.id for b in books if b.author.lname in fvalues]
resID &= set(ids)
resBooks = [{
'id': b.id,
'title': b.title,
'author': f'{b.author.fname} {b.author.lname}',
'isbn': b.isbn
} for b in books if b.id in resID]
return resBooks
def common_process(self, proddesc_list, Item_Description,
shouldConsiderOnlyTopScore, matchingThreshold, cursor,
listType):
if len(proddesc_list) == 0:
return '|||||||||||||||||||||'
if shouldConsiderOnlyTopScore:
bestMatch = process.extractBests(Item_Description,
proddesc_list,
scorer=fuzz.token_sort_ratio,
limit=1,
score_cutoff=matchingThreshold)
if len(bestMatch) == 0:
return '|||||||||||||||||||||'
else:
return self.getmatchedrecord(bestMatch[0][0], bestMatch[0][1],
cursor, listType)
else:
bestMatches = process.extractBests(Item_Description,
proddesc_list,
scorer=fuzz.token_sort_ratio,
limit=5,
score_cutoff=matchingThreshold)
if len(bestMatches) == 0:
return '|||||||||||||||||||||'
else:
bestMatches = set(bestMatches)
rtrn_str = ''
for match in bestMatches:
matched_record = self.getmatchedrecord(
match[0], match[1], cursor, listType)
rtrn_str = rtrn_str + '#' + matched_record
rtrn_str.strip('#')
return rtrn_str
def match_ingredients(self, recipe):
"""
Makes a fuzzy match between ingredients in a recipe and in the ingredient:compound dict
:param recipe: ingredients to match
:param comp_ing_dict: compound ingredient dict
:return:
"""
matches = []
for ing in recipe:
component_matches = self.decomposer.get_component_ingredients(ing)
if component_matches:
return self.match_ingredients(component_matches)
match_list = process.extractBests(ing, self.sanitized_comp_ing_dict.keys(),
scorer=fuzz.ratio, score_cutoff=98)
matches += [match[0] for match in match_list]
match_list = process.extractBests(ing, self.sanitized_comp_ing_dict.keys(),
scorer=fuzz.token_sort_ratio,
score_cutoff=85)
matches += [match[0] for match in match_list]
if not matches:
sub_ing = ing.split()
for i in sub_ing:
match_list = process.extractBests(i, self.sanitized_comp_ing_dict.keys(),
scorer=fuzz.ratio, score_cutoff=88)
matches += [match[0] for match in match_list]
return list(set(matches))
def _find_person(self, player_name, choices):
print('Searching %s' % player_name)
matching_results = process.extractBests(unidecode(player_name), choices,
scorer=fuzz.partial_token_set_ratio,
score_cutoff=75)
if len(matching_results) > 0:
# si les meilleurs matchs sont à egalité de score, chercher à nouveau avec méthode différente
best_score = 0
creme = dict()
for name, score, plid in matching_results:
if score >= best_score:
best_score = score
creme.update({plid: name})
else:
# la liste renvoyée par extractBests est triée donc on peut s'arreter dès que le niveau baisse.
break
# combien de meilleurs scores ?
if len(creme) == 1:
plid, plname = creme.popitem()
print('Found %s at first round with ratio %s' % (plname, best_score))
matching_player = FootballPerson.objects.get(pk=plid)
return matching_player, best_score
else:
print('Multiple matches found with ratio %s, refining...' % best_score)
refine_results = process.extractBests(player_name, creme,
scorer=myfuzz.partial_token_set_ratio_with_avg)
plname, ratio, plid = refine_results[0]
print('Found %s at second round with ratio %s then %s' % (plname, best_score, ratio))
matching_player = FootballPerson.objects.get(pk=plid)
return matching_player, best_score
else:
print("Alert : no match for %s" % player_name)
return None, 0.0
def fuzzy_match(
mistyped_command: str, score_cutoff: int = 75
) -> Optional[Tuple[str, int]]:
match = process.extractBests(
mistyped_command, FUZZY_COMMANDS.keys(), score_cutoff=score_cutoff, limit=1
)
if len(match) == 0:
match = process.extractBests(
mistyped_command, FUZZY_ALIASES.keys(), score_cutoff=score_cutoff, limit=1
)
return match[0] if len(match) != 0 else None
def buildEachLocatorTagPattern(self, eachLocatorArray, i, sourceDataProcessed, patternBuild):
try:
import re
eachLocator = eachLocatorArray[i].upper()# mjj
"""considering if a string contains number, prevent it for fuzzy search"""
# searchOnlyNumAndCharObj = re.search(r'^[0-9-`!@#$%^&*()_+=\\|}\]\[{\';:\/\?>\.,<~ ]+$', eachLocator)
searchOnlyNumAndCharObj = re.search(r'[0-9]', eachLocator)
if searchOnlyNumAndCharObj:
""" Converting number to coresponding regular expression """
bestMatch = re.sub('\d', '\d', eachLocator)
# print('pattern found', bestMatch)
if i == 0:
patternBuild = patternBuild + bestMatch
else:
patternBuild = patternBuild + '(.*)' + bestMatch
return patternBuild
else:
""" Get fuzzy matching string array """
matchingFuzzyWord = self.getFuzzySearchData(eachLocator, sourceDataProcessed)
if len(process.extractBests(eachLocator, matchingFuzzyWord)) > 0:
"""
Find the best amoung them
It is very importent for selecting appropriate confidence limit, one thing is based on string length
:todo: find other parameters for improving
"""
if len(eachLocator) < 5:
confidenceLimit = 90
else:
confidenceLimit = 80
bestMatch, confidence = process.extractBests(eachLocator, matchingFuzzyWord, limit=1)[0]
if len(matchingFuzzyWord) > 0 and confidenceLimit < confidence:
bestMatch = self.regularExpressionHandling(bestMatch, 0)
if i == 0:
patternBuild = patternBuild + bestMatch
else:
patternBuild = patternBuild + '(.*)' + bestMatch
return patternBuild
else:
if i == 0 or len(eachLocatorArray) == i + 1:
""" if first or last locator doesnot match then no need for further process (Improvement in searching) """
return False
elif len(matchingFuzzyWord) == 0 and (i == 0 or len(eachLocatorArray) == i + 1):
""" if first or last locator doesnot match then no need for further process (Improvement in searching) """
return False
except Exception as e:
print('error in buildEachLocatorTagPattern in StringHandling', e)
return False
def search(bank_query: str, branch_query: str):
"""
:param bank_query: Bank string. Example: "icici", "hdfc bank"
:param branch_query: Branch address string. Example:
:return: DF row/object of the matching branch with IFSC & MICR code
Step1: Load CSV to a DF
Step2: Slice the results to the substring i.e. bank_string
Step3:
3.1) Fuzzy search the BRANCH col. with branch_query. BRANCH_NAME
3.2) Fuzzy search the ADDRESS col. with branch_query. ADDRESS
3.3) Compare the fuzzy search score of 3.1 and 3.2. If both scores are equal or BRANCH_NAME > ADDRESS,
we assume the BRANCH_NAME row to be returned else ADDRESS row.
Order of CSV Headers: ['IFSC', 'BRANCH', 'CENTRE', 'DISTRICT', 'STATE', 'ADDRESS', 'CONTACT',
'IMPS', 'RTGS', 'CITY', 'NEFT', 'MICR']
"""
# Step 1: Find the branch from `BANK` column
# bank_dict = process.extractBests('ICIC', df1.BANK.to_dict(), score_cutoff=90, limit=None)
# bank_branches = df1.loc[[l[2] for l in bank_dict]]
# OR =>
# PS: The following will fail if the bank name provided exceeds the bank name we have
bank_branches = df1[df1['BANK'].str.lower().str.contains(bank_query.lower())]
print('===== ALL BRANCHES OF BANK ======'.center(10))
print(bank_branches)
# FYI to convert whole column to lower case: bank_branches.ADDRESS.str.lower()
branch_matches = process.extractBests(branch_query, bank_branches.BRANCH.to_dict()) # , limit=2)
address_matches = process.extractBests(branch_query, bank_branches.ADDRESS.to_dict())
best_branch = branch_matches[0]
best_address = address_matches[0]
print(best_address, best_branch)
if best_branch[1] == best_address[1]:
optimal_result = best_branch
elif best_branch[1] > best_address[1]:
optimal_result = best_branch
else:
optimal_result = best_address
result_index = optimal_result[2]
result_row = bank_branches.loc[result_index]
print('==== Queried BRANCH ===='.center(10))
print(result_row)
logger.error("{}".format(result_row))
return result_row
def get_categories_related(categories_raw, min_score=90, min_bad_score=80, is_name=False, names=None):
if categories_raw:
categories_raw = clean(categories_raw, is_name)
if names:
names = clean(names, True)
# print(categories_raw)
if categories_raw:
match_categories = []
for name, attrs in categories_json.items():
not_choices = attrs.get('banned')
bad_results = process.extractBests(categories_raw, not_choices, scorer=fuzz.partial_token_set_ratio,
score_cutoff=min_bad_score)
if names:
bad_results += process.extractBests(names, not_choices, scorer=fuzz.partial_token_set_ratio,
score_cutoff=min_bad_score)
if not bad_results:
choices = attrs.get('tokens')
results = process.extractBests(categories_raw, choices, scorer=fuzz.partial_token_set_ratio, score_cutoff=min_score)
if results:
# print("++++++ \t", name, ': ', results)
match_categories.append(name)
result_keys = {result[0] for result in results}
for cat in attrs.get('subcats'):
choices = set(list(cat.values())[0].get('tokens'))
if (choices & result_keys):
not_choices = list(cat.values())[0].get('banned')
bad_results_sub = process.extractBests(categories_raw, not_choices,
scorer=fuzz.partial_token_set_ratio,
score_cutoff=min_bad_score)
if names:
bad_results_sub += process.extractBests(names, not_choices,
scorer=fuzz.partial_token_set_ratio,
score_cutoff=min_bad_score)
if not bad_results_sub:
cat_name = list(cat.keys())[0]
match_categories.append(cat_name)
# print("++++++ \t", cat_name, ': ', results)
# else:
# cat_name = list(cat.keys())[0]
# print("------ \t", cat_name, ': ', bad_results_sub)
# else:
# print("------ \t", name, ': ', bad_results)
if {"Mascotas", "Autos, Motos y llantas"} & set(match_categories):
aux = set(match_categories) - {"Super"}
if len(aux) > 1:
return []
return match_categories
else:
return []
def get_best_match_location(manga_name, choices):
"""
Extracts the best possible matching manga.
If it cannot find the exact match asks the user for the best matches.
"""
matches=process.extractBests(manga_name, choices, limit=10, score_cutoff=100)
if len(matches) == 0:
matches=process.extractBests(manga_name, choices, limit=10, score_cutoff=80)
match=ask_best_match(matches)
return website_specific.absoulute_location( match )
def get_normalized(starting_name: str, names_coll: list):
reading = process.extractBests(starting_name, names_coll, score_cutoff=75)
candidates = [starting_name]
best_n = starting_name
max_n = len(reading)
while reading != []:
(new, points) = reading.pop()
if points < 100 and new not in candidates:
candidates.append(new)
tmp = process.extractBests(new, li, score_cutoff=75)
reading = reading + tmp
if len(tmp) > max_n:
best_n = new
return (best_n, candidates)
def get_match(word, word_list, precision=90):
if len(word) == 0:
return ''
best_match = process.extractBests(word, word_list, limit=1)
if best_match[0][1] <= precision:
best_match = ''
return best_match
def extract(self, expectation, limit=4):
"""extract(extraction, limit=4)
Returns the address or addresses within the set of the reverse
geocoded addresses that best match an expected result. Uses fuzzywuzzy
under the hood for matching.
Args:
expectation (str): The string indicating your expected result for a
reverse geocoding operation. It should probably look like an
address. Results are returned in the order that best meets this
expected address.
Kwargs:
limit (int): The maximum number of match candidates to retrieve
from fuzzywuzzy. The length of the returned array may be longer, if
the set of addresses has identical addresses that are good matches
for the expected address (i.e. if two geocoders resolve to the same
string address).
Returns:
list. Return value is a list of tuples, where each tuple contains a
geopy Location, and a matching score based on an extension of the
Levenshtien distance between the expectation and the Location's
address (a higher score is a better match). The algorithm is
implemented by SeatGeek's fuzzywuzzy, and you can read more here:
http://chairnerd.seatgeek.com/fuzzywuzzy-fuzzy-string-matching-in-python/
"""
extractions = fuzzyprocess.extractBests(
expectation, [str(a) for a in self.addresses],
limit=limit)
result = []
for extraction in extractions:
result.extend([(x, extraction[1]) for x in self.addresses
if str(x) == extraction[0]])
return result
def test_identical_strings_extracted(scorer, processor, data):
"""
Test that identical strings will always return a perfect match.
:param scorer:
:param processor:
:param data:
:return:
"""
# Draw a list of random strings
strings = data.draw(
st.lists(st.text(min_size=10, max_size=100),
min_size=1, max_size=50))
# Draw a random integer for the index in that list
choiceidx = data.draw(st.integers(min_value=0, max_value=(len(strings) - 1)))
# Extract our choice from the list
choice = strings[choiceidx]
# Check process doesn't make our choice the empty string
assume(processor(choice) != '')
# Extract all perfect matches
result = process.extractBests(choice,
strings,
scorer=scorer,
processor=processor,
score_cutoff=100,
limit=None)
# Check we get a result
assert result != []
# Check the original is in the list
assert (choice, 100) in result
def item(self, name, fuzzy_threshold=100):
"""
Extract a password from an unlocked Keychain using fuzzy
matching. ``fuzzy_threshold`` can be an integer between 0 and
100, where 100 is an exact match.
"""
matches = process.extractBests(
name,
self._items.keys(),
score_cutoff=(fuzzy_threshold-1),
)
if matches:
if len(matches) > 1:
for i, m in enumerate(matches):
sys.stderr.write('[%s] %s\n' % (i, m[0]))
sys.stdin = open('/dev/tty')
try:
exact_name = matches[int(raw_input())][0]
except:
return None
else:
exact_name = matches[0][0]
item = self._items[exact_name]
item.decrypt_with(self)
return item
else:
return None
def bot_message(self, cpu_lookup):
self.cpu_lookup = cpu_lookup
logging.info('Looking for CPU...')
try:
choices = []
for cpu in self.cpu_list:
match_criteria = fuzz.token_set_ratio(
self.clean_input(cpu), self.clean_input(cpu_lookup))
if match_criteria >= 45:
choices.append(cpu)
closest_match = process.extractOne(cpu_lookup,
choices,
scorer=fuzz.token_set_ratio,
score_cutoff=85)
logging.info(f"Searching: {cpu_lookup}, Closest: {closest_match}")
closest_match_name = closest_match[0]
bot_reply = self.display_cpu_info(closest_match_name)
except TypeError:
limit_choices = process.extractBests(cpu_lookup, choices)
if limit_choices:
bot_reply = f"No direct CPU match found for **{cpu_lookup}**, displaying {len(limit_choices)} potential matches:\n\n"
search_results = ''
for result in limit_choices[:6]:
cpu_name = result[0]
search_results += f"[{cpu_name}]({self.cpu_list[cpu_name]})\n\n"
bot_reply += search_results
bot_reply += "\n\nFeel free to ask me again (`CPUBot! cpu model`) with these models or visit PassMark directly!\n"
# Handles no results being found in search
if not limit_choices:
bot_reply = f"\n\nI'm sorry, I couldn't find any information on **{cpu_lookup}**.\n\nPlease feel free to try again; perhaps you had a spelling mistake, or your CPU is not listed in the [Passmark CPU list]({self.passmark_page})."
return bot_reply
def search_without_properties_list(self):
"""
Первоначальный поиск не использующий список связанных слов
"""
search_list = Suggest.root.suggest_db[self.first_let][str(
self.len_word)]
tokens = self.search_token(self.search_word, list(search_list.keys()),
50, 3)
for token, percent in tokens:
tokens_list = search_list[token[:self.len_word]]
token = process.extractBests(self.search_word,
tokens_list,
limit=3,
score_cutoff=50)
self.logger.debug(msg=[token, tokens_list])
for word, percent in token:
if word not in self.root.stop_words:
word_prop = process.extractOne(
word, Suggest.root.search_words_db[word[0]].keys())[0]
gm_names = Suggest.root.search_words_db[
word[0]][word_prop]['gm_name']
self.res_list.append(
(self.start_phrase + word, percent, gm_names))
def upload_file ():
data = {"success": False}
print(request.files)
if request.files.get('file'):
# read the file
file = request.files['file']
# read the filename
filename = file.filename
# create a path to the uploads folder
filepath = os.path.join(app.config['UPLOAD_FOLDER'], filename)
file.save(filepath)
prediction = predict(filepath)
breed = prediction[0][1]
animal_outcome = pd.read_csv('../Animals.csv')
r = process.extractBests(breed.replace('_', ' '), animal_outcome.Breed.to_dict(), scorer=partial_ratio, score_cutoff=70, limit=1000000000)
animal_analysis = animal_outcome.loc[map(lambda x: x[-1], r)]
return animal_analysis.to_json(orient='records')
# return render_template('table.html', dogs=animal_analysis.to_dict(orient='records'))
return jsonify({"error": "there is an error!"})
def fuzzyExtract(self, qs, ls, threshold):
'''
todo fuzzy search seperation in words
:param qs: query string
:param ls: large string
:param threshold: threshold
:return:
'''
'''fuzzy matches 'qs' in 'ls' and returns list of
tuples of (word,index)
'''
if len(qs) < self.fuzzySearchOptimumLength:
processThreshold = 60
max_l_dist = 0
else:
processThreshold = threshold
max_l_dist = 1
for word, confidence in process.extractBests(qs, (ls,), score_cutoff=processThreshold):
print('word {}'.format(word), confidence)
for match in find_near_matches(qs, word, max_l_dist=max_l_dist):
match = word[match.start:match.end]
print('match {}'.format(match))
index = ls.find(match)
def get_best_events_matching(event_list: list, search: str):
if len(event_list) == 0:
return []
if search is None or len(search) == 0:
return event_list
try:
i = int(search)
for event in event_list:
if event["id"] == event_id:
return [event]
except ValueError:
pass
except TypeError:
pass
# try doing a search through the names of events:
event_names = [(e["event_name"], e) for e in event_list]
def process_by_first(v):
return fuzzy_fullproc(v[0])
results = process.extractBests((search, None),
event_names,
processor=process_by_first,
scorer=fuzz.partial_ratio,
score_cutoff=90)
rl = []
for e, score in results:
rl.append(e[1])
return rl
def fuzzy_combine_def(self, definitions, scorer=None):
d = dict(definitions)
order_key = lambda name: d[name]
result = []
names = set(d.keys())
if scorer is None:
scorer = fuzz.token_set_ratio
while names:
first = names.pop()
similar = process.extractBests(first, names, scorer=scorer, limit=1000, score_cutoff=65)
similar_names = [name for name, _ in similar]
for name in similar_names:
names.remove(name)
res_names = [first] + similar_names
ordered_by_score = sorted(res_names, key=order_key, reverse=True)
total_score = sum([d[name] for name in ordered_by_score])
result.append((ordered_by_score, total_score))
resorted = sorted(result, key=operator.itemgetter(1), reverse=True)
return resorted
def DupliFuzzyMatch(query):
#choices = list(result[['FullTitle'].itertuples(index=False, name=None))
#choices=list(dict.fromkeys(choices))
choices = result['FullTitle'].tolist()
prc=process.extractBests(query,choices,score_cutoff=60, limit=10)
prc=list(dict.fromkeys(prc))
'''
labels = ['FullTitles', 'Score']
df = pd.DataFrame.from_records(prc, columns=labels)
df1=pd.DataFrame(df['FullTitles'].values.tolist(), index=df.index)
df = pd.concat([df1, df['Score']], axis=1, sort=False)
df.rename(columns={0: 'FullTitle',1:'System'}, inplace=True)
'''
lim=len(prc)
if lim>0:
df = pd.DataFrame([])
for i in range(0,lim):
prcc=Dup_Count[Dup_Count['FullTitle'].str.match((prc[i][0]))]
df = df.append(prcc)
df=df.drop_duplicates(['Titles'],keep= 'last')
df=df[['FullTitle','System', 'Count']]
df=df.sort_values(['Count'],ascending=False)
df.reset_index(drop = True, inplace = True)
else:
df = pd.DataFrame(columns=['FullTitle','System','ISBN13'])
return df
def fuzzy_search(query_string,
options_dict,
scorer=fuzz.QRatio,
score_cutoff=81):
"""
Uses fuzzy search to find best matches for the `query_string` in
`options_dict`
Args:
query_string (str):
String used to search for matches
options_dict (list):
List of options to find matches in.
scorer (fuzz.Scorer, optional):
Strategy to use when searching for matches.
Defaults to fuzz.QRatio.
score_cutoff (int, optional):
Similarity score cutoff threshold.
Defaults to 81.
Returns:
array:
Array of matching words
"""
fuzzy_results = process.extractBests(query_string,
options_dict,
scorer=fuzz.QRatio,
score_cutoff=score_cutoff)
return fuzzy_results
def fuzzy_match(input_string, choices, n=3, score_cutoff=50):
try:
return list(
zip(*(fuzzy_process.extractBests(
input_string, choices, score_cutoff=score_cutoff))))[0][:n]
except IndexError:
return []
def fuzzy_extract_matches(word, possibilities, junk_seq=None, n=3, cutoff=0.5):
score_cutoff = int(cutoff*100)
matches = process.extractBests(word,possibilities,score_cutoff=score_cutoff,limit=n)
final_matches = []
for m,s in matches:
final_matches.append((m,s/100))
return final_matches
def get_match(word_list: list,
word: str,
score_cutoff: int = 60,
isPartial: bool = False,
limit: int = 1):
"""Uses fuzzywuzzy to see if word is close to entries in word_list
Returns a tuple of (MATCH, SCORE)
"""
if not word:
return (None, None)
result = None
scorer = fuzz.ratio
if isPartial:
scorer = fuzz.partial_ratio
if limit == 1:
result = process.extractOne(word,
word_list,
scorer=scorer,
score_cutoff=score_cutoff)
else:
result = process.extractBests(word,
word_list,
scorer=scorer,
score_cutoff=score_cutoff,
limit=limit)
if not result:
return (None, None)
return result
def get_fuzzy_search_movies(fuzzy, repo):
"""
"""
movies = list(repo.dataset_of_movies)
movie_title_dict = dict(enumerate([movie.title for movie in movies]))
best_matches = process.extractBests(fuzzy, movie_title_dict, score_cutoff=50)
return [utils.movie_to_dict(movie) for movie in [movies[z] for (x, y, z) in best_matches]]
def address_candidates(input_data, cutoff_info):
"""
Returns address candidates
"""
output = {}
st_num = input_data['st_num']
st_name = input_data['st_name']
if input_data['appeal_type'] == "detroit_single_family":
cutoff = cutoff_info['detroit']
region = 'detroit'
elif input_data['appeal_type'] == "cook_county_single_family":
cutoff = cutoff_info['cook']
region = 'cook'
mini = address_candidates_query(region, st_num)
candidate_matches = process.extractBests(st_name, mini.st_name, score_cutoff=50)
selected = mini[mini['st_name'].isin([i for i, _, _ in candidate_matches])].copy()
selected['Distance'] = 0
if input_data['appeal_type'] == "detroit_single_family":
selected = prettify_detroit(selected, False)
elif input_data['appeal_type'] == "cook_county_single_family":
selected = prettify_cook(selected, False)
selected['eligible'] = selected.assessed_value <= cutoff
selected.dropna(axis=0, inplace=True)
output['candidates'] = selected.to_dict(orient='records')
if len(output['candidates']) == 0: #if none found raise
raise Exception('No Matches Found')
return output
def view_script(title_to_href):
#This is what will get the actual script of a single movie
movie_names_row = input(
'\033[31m Entrez le nom du film et l\'année entre parenthèses, ex : 99 Homes (2014) : \033[0m'
)
#The first attribute of extract will be user's input, second is the list of all movie scripts, third is number of results determined by user
movie_names = process.extractBests(movie_names_row,
title_to_href.keys(),
limit=1,
score_cutoff=70)
titles = [movie_name[0] for movie_name in movie_names]
title = titles[0]
print(title)
if input('\033[31m Entrez "yes" pour continuer : \033[0m') == 'yes':
if title in title_to_href:
print(title_to_href[title])
else:
print('Aucun résultat')
page_url = "https://www.springfieldspringfield.co.uk/movie_script.php?movie=" + title_to_href[
title]
page = urllib.request.urlopen(page_url)
soup = BeautifulSoup(page, 'html.parser')
script = soup.find("div", {"class": "movie_script"})
print(script.text)
else:
pass
def normalize_country(country):
if process.extractBests(country, choices=['N/A', 'n/a'], score_cutoff=HIGH_CUTOFF):
return None
if country == 'null':
return None
if country in ['USA', 'UK']:
return country
if process.extractOne(country, choices=['United Kingdom', 'UK'], score_cutoff=HIGH_CUTOFF):
return 'UK'
if process.extractOne(country, choices=['England', 'Wales', 'Scotland'], score_cutoff=HIGH_CUTOFF):
return 'UK'
if process.extractOne(country, choices=['Russia', 'Russian Federation'], score_cutoff=LOW_CUTOFF):
return 'Russia'
try:
country = pycountry.historic_countries.get(name=country.capitalize())
return country.name
except KeyError: pass
try:
country = pycountry.historic_countries.get(alpha2=country.upper())
return country.name
except KeyError: pass
countries = [country.name for country in pycountry.countries]
best_match = process.extractOne(country, choices=countries, score_cutoff=90)
return best_match if best_match else country
def test_identical_strings_extracted(scorer, processor, data):
"""
Test that identical strings will always return a perfect match.
:param scorer:
:param processor:
:param data:
:return:
"""
# Draw a list of random strings
strings = data.draw(
st.lists(st.text(min_size=10, max_size=100), min_size=1, max_size=50))
# Draw a random integer for the index in that list
choiceidx = data.draw(
st.integers(min_value=0, max_value=(len(strings) - 1)))
# Extract our choice from the list
choice = strings[choiceidx]
# Check process doesn't make our choice the empty string
assume(processor(choice) != '')
# Extract all perfect matches
result = process.extractBests(choice,
strings,
scorer=scorer,
processor=processor,
score_cutoff=100,
limit=None)
# Check we get a result
assert result != []
# Check the original is in the list
assert (choice, 100) in result
def get_grid(name):
name = name.lower()
if name in grid_lookup.keys():
return grid_lookup[name]
if name in grid_names:
index = grid_names.index(name)
return grid_uk[index].get('id')
elif name.replace('the ', '', 1) in grid_names:
index = grid_names.index(name.replace('the ', '', 1))
return grid_uk[index].get('id')
fuzznames = grid_names + [name for name in grid_lookup.keys()]
fuzzmatch = process.extractBests(
name.replace('university', '').replace('the', ''), fuzznames)
print('\nName to match:', name)
for i, match in enumerate(fuzzmatch):
print(' ' + str(i) + ': ' + match[0])
i = input('Select a match (0-indexed) or give the correct GRID:')
try:
i = int(i)
return grid_uk[grid_names.index(fuzzmatch[i][0])].get('id')
except ValueError:
return i
def fuzMatch(query):
#choices = result['FullTitle'].tolist()
choices = list(result[['FullTitle', 'System','ISBN13']].itertuples(index=False, name=None))
choices=list(dict.fromkeys(choices))
prc=process.extractBests(query,choices,score_cutoff=60, limit=10)
prc=list(dict.fromkeys(prc))
lim=len(prc)
if lim>0:
labels = ['FullTitles', 'Score']
df = pd.DataFrame.from_records(prc, columns=labels)
df1=pd.DataFrame(df['FullTitles'].values.tolist(), index=df.index)
df = pd.concat([df1, df['Score']], axis=1, sort=False)
df.rename(columns={0: 'FullTitle',1:'System',2:'ISBN13'}, inplace=True)
else:
df = pd.DataFrame(columns=['FullTitle','System','Score','ISBN13'])
#for i in range(0,lim):
# prcc=result[result['FullTitle'].str.match((prc[i][0]))]
# df = df.append(prcc,sort=True)
'''
lent=len(df)
for j in range(0,lent):
kk= df['FullTitle'].values[j]
score= fuzz.WRatio(query,kk)
df['Score'].values[j]=score
df=df.drop_duplicates(['FullTitle','System'],keep= 'last')
'''
#df.reset_index(drop = True, inplace = True)
return df
def main() -> None:
args = sys.argv
first = args[1]
second: Optional[str]
try:
second = args[2]
except IndexError:
second = None
first_res: Tuple[str, int] = process.extractOne(first, first_args)
if first_res[1] >= MIN_SCORE:
if first_res[0] == "spells":
if second is None:
print(spells)
else:
spell_keywords = get_spell_keywords()
second_res: List[Tuple[str, int]] = process.extractBests(
second, spell_keywords, score_cutoff=MIN_SCORE)
potential_matches = [second_res[0][0]]
potential_matches.extend([
x[0] for x in second_res[1:]
if second_res[0][1] - x[1] <= 5 and second_res[0][1] != 100
])
if len(potential_matches) > 1:
string = "Multiple potential matches found:"
for match in potential_matches:
string += f"\n{match}"
print(string)
elif len(potential_matches) == 1:
print_keyword_info(potential_matches[0])
else:
print(second_res)
def thumbnail(self):
"""
To find a list of thumbnail candidates, collections can be filtered inexactly against the title of the page (to catch spelling errors).
"""
colls = []
q_title = Collection.objects.filter(name__iexact=self.title)
q_slug = Collection.objects.filter(name__iexact=self.slug)
# We cast a very broad net to find all collections with a name matching the slug or title of the page. Collections created by Mediabox will have a name matching the name of the subdirectory they came from, so names are not unique.
colls.append([e for e in q_title])
colls.append([e for e in q_slug])
candidates = []
for coll in colls:
imgs_for_coll = Image.objects.filter(collection=coll)
candidates.append([entry.title for entry in imgs_for_coll])
for query in ['thumbnail', 'directory', self.title]:
hit = process.extractBests(query, candidates, score_cutoff=70)
if hit:
return Image.objects.get(title__iexact=hit[0])
consolation = random.choice(candidates) # Just pick one.
return Image.objects.get(title__iexact=consolation)
def _handle_not_found(self, query):
"""
Handles when a query does not match a valid command or cog.
Will pass on possible close matches along with the
``HelpQueryNotFound`` exception.
Parameters
----------
query: str
The full query that was requested.
Raises
------
HelpQueryNotFound
"""
# combine command and cog names
choices = list(self._bot.all_commands) + list(self._bot.cogs)
result = process.extractBests(query,
choices,
scorer=fuzz.ratio,
score_cutoff=90)
raise HelpQueryNotFound(f'Query "{query}" not found.', dict(result))
def enrich_supports(rows):
relevant_rows = 0
matched_rows = 0
for row in rows:
bests = []
if row['request_type'] == 'א3':
relevant_rows += 1
payments = row['payments']
if payments and len(payments) > 0:
payment = payments[0]
key = (payment['support_title'], payment['supporting_ministry'])
if key in cache:
bests = cache[key]
else:
bests = process.extractBests(
{
'purpose': key[0],
'office': key[1]
},
criteria,
processor=id,
scorer=criteria_scorer
)
cache[key] = bests
if len(bests)>0:
matched_rows += 1
row['criteria_docs'] = [x[0] for x in bests]
yield row
logging.info('MATCH STATS: rel: %d, matched: %d', relevant_rows, matched_rows)
def word_match(words, menu, vector_dict, threshold=93, best_matches=None):
"""Matches single-word keyword vectors to menu items
Args:
littlekeys (list): a list of single keywords
menu (dict): a dictionary of a menu for a given restaurant
vector_dict (dict): a dictionary whose keys are keywords and values are vectors
threshold (int): fuzzy-match score required for keyword matching
best_matches (dict): dict created in best_matches to add to
Returns:
best_matches (dict): a nested dictionary whose outer keys are menu items, inner keys
are tuples of keyword matches and their corresponding fuzzy match score, and values
are the review vectors of each keyword
"""
if not best_matches:
best_matches = {}
for name, meals in menu.iteritems():
for meal, value in meals.iteritems():
for food in value:
item = remove_accents(food)
if len(item.split()) == 1:
matches = process.extractBests(item, words, scorer=fuzz.QRatio, score_cutoff=threshold, limit=3)
vectors = []
for match in matches:
vectors.append(vector_dict[match[0]])
best_matches[item] = dict(zip(matches, vectors))
return best_matches
def get_breeze_user(self, name):
self.get_all_users()
# Put name in format so get_name applies
temp_name = {'first_name': name, 'last_name': ''}
get_name = lambda u: ' '.join([u['first_name'], u['last_name']])
bests = process.extractBests(temp_name,
self._people,
get_name,
limit=3)
# Penalize 10 points if first names start with different letters
best = (bests[0], 0) # Hold best candidate and the score
for item in bests:
penalty = 0
if name[0].casefold() != item[0]['first_name'][0].casefold():
penalty = 10
if item[1] - penalty > best[1]:
best = (item, item[1] - penalty)
if best[1] < 90:
raise MatchNotFoundError
return best[0]
def fuzzy_combine_def(self, definitions, scorer=None):
d = dict(definitions)
order_key = lambda name: d[name]
result = []
names = set(d.keys())
if scorer is None:
scorer = fuzz.token_set_ratio
while names:
first = names.pop()
similar = process.extractBests(first,
names,
scorer=scorer,
limit=1000,
score_cutoff=65)
similar_names = [name for name, _ in similar]
for name in similar_names:
names.remove(name)
res_names = [first] + similar_names
ordered_by_score = sorted(res_names, key=order_key, reverse=True)
total_score = sum([d[name] for name in ordered_by_score])
result.append((ordered_by_score, total_score))
resorted = sorted(result, key=operator.itemgetter(1), reverse=True)
return resorted
def fuzzy_matches(x=None, query='', where='both', score_cutoff=50, limit=10):
"""
Get the internal matches used in fuzzy search. Exposses the scores of each match.
If item is a 2 tuple match is on a key, if the item is a 3 tuple match is on a value.
"""
if where is 'keys':
matches = process.extractBests(query, x.keys(), score_cutoff=score_cutoff, limit=limit)
elif where is 'values':
matches = process.extractBests(query, x, score_cutoff=score_cutoff, limit=limit)
elif where is 'both':
key_matches = process.extractBests(query, x.keys(), score_cutoff=score_cutoff, limit=limit)
val_matches = process.extractBests(query, x, score_cutoff=score_cutoff, limit=limit)
# Combine results
results = key_matches + val_matches
return results
def getclosestelement(name, array):
if name in array:
return name
try:
close = process.extractBests(name, array)[0][0]
except:
close = difflib.get_close_matches(name, array)[0]
return close
def search_collection(items, query, key='title'):
"""Levenstein fuzzy search"""
d = {i[key]: i for i in items}
bests = process.extractBests(query, list(d.keys()))
if bests:
return [d[match[0]] for match in bests]
else:
return []
def busca(self, query):
query = normaliza_string(query)
amostra = self.cria_amostra()
scorer = seleciona_scorer(query)
resultado = process.extractBests(query, amostra, limit=10, scorer=scorer, score_cutoff=65)
if scorer == fuzz.token_set_ratio:
resultado = process.extract(query, lapida_extracao(resultado), limit=20, scorer=fuzz.partial_ratio)
return lapida_extracao(resultado)
def lookup(self, keyword):
data = _concat(self.db.all())
title = data.pop('Title')
auth = data.pop('Authors')
choices = title + auth
searchKey = process.extractBests(keyword, choices)
for i in searchKey:
if i[1] >= 90:
self.search(i[0])
def search(self, name): '''searches list of movies for a movie''' movies_path = self.movies_path; onlyfolders = [f for f in os.listdir(movies_path) if os.path.isdir(os.path.join(movies_path, f))] results = process.extractBests(name,onlyfolders) for result in results: print(result[0])
def list(self, query=None, by_nick=False):
if query is None:
return self.licenses
field = 'nick' if by_nick else 'title'
choices = {getattr(lic, field): lic for lic in self.licenses}
print choices
field_values = extractBests(query, choices.iterkeys())
lics = [choices[v] for v, _ in field_values]
return lics
def fuzzy_search(x=None, query='', where='both', score_cutoff=50, limit=10):
"""
Search thru keys, values, or both via fuzzy string comparison.
"""
if where is 'keys':
matches = process.extractBests(query, x.keys(), score_cutoff=score_cutoff, limit=limit)
results = {k : x[k] for k, score in matches}
elif where is 'values':
matches = process.extractBests(query, x, score_cutoff=score_cutoff, limit=limit)
results = {k : v for v, score, k in matches}
elif where is 'both':
key_matches = process.extractBests(query, x.keys(), score_cutoff=score_cutoff, limit=limit)
key_results = {k : x[k] for k, score in key_matches}
val_matches = process.extractBests(query, x, score_cutoff=score_cutoff, limit=limit)
val_results = {k : v for v, score, k in val_matches}
# Combine results
results = key_results.copy()
results.update(val_results)
return results
def selectBestMatch(tokens, hits, qualityThreshold=50, creatorWeight=2, titleWeight=1):
results = []
for hit in hits:
try:
creators = set()
normalizedScore = 0
creatorScore = 0
score = 0
if hit.has_key('title'):
foo = process.extractBests(hit['title'], tokens)
if len(foo) > 0:
score += foo[0][1] * titleWeight
if hit.has_key('creator'):
creators = creators.union(set(hit['creator']))
if hit.has_key('person'):
creators = creators.union(set(hit['person']))
if hit.has_key('contributor'):
creators = creators.union(set(hit['contributor']))
if hit.has_key('publisher'):
creators = creators.union(set(hit['publisher']))
if len(creators) > 0:
for creator in creators:
foo = process.extractBests(creator, tokens)
if foo != None and foo != []:
creatorScore += foo[0][1]
score += (creatorScore/len(creators)) * creatorWeight
normalizedScore = score/(creatorWeight+titleWeight)
if normalizedScore >= qualityThreshold:
bisect.insort_left(results, (normalizedScore, hit))
except UnicodeDecodeError as e:
continue
if len(results) > 0:
return results[-1]
return None
def match_lists(list1, list2, precision=90):
'''
Takes two lists
Returns dict of matches (if best match scores above precision)
and non-matches (otherwise); both dicts have key = list1
element and value = best match
'''
in_both = {}
not_in_both = {}
for manager1 in list1:
best_match = process.extractBests(manager1, list2, limit=1)
if best_match[0][1] > precision:
in_both[manager1] = best_match[0][0]
else:
not_in_both[manager1] = best_match[0][0]
return [in_both, not_in_both]
def get_closest_text_bump(search_text):
search_text = {'text': search_text}
text_bumps = filter(lambda bump: bump['text'], get_bumps())
if text_bumps:
best_bumps = process.extractBests(
search_text,
text_bumps,
scorer=fuzz.token_set_ratio,
processor=lambda bump: bump['text'].lower()
)
best_bumps = [
bump[0]
for bump in best_bumps
if bump[1] == best_bumps[0][1]
]
return get_bump_link(random.choice(best_bumps))
else:
return 'no dice m8'
def get_best(self, message):
message = message.lower()
tokens = self.create_tokens(message)
matches = process.extractBests(' '.join(tokens), self.trump_quotes,
score_cutoff=50, limit=3)
if not matches:
return None
if re.search('trump|the donald', message):
matches = [(x[0], x[1] + 10, x[2]) for x in matches ]
adjusted = []
for x in matches:
b_tokens = self.create_tokens(x[0])
score = self.score_ngrams(tokens, b_tokens)
adjusted.append((x[0], x[1] + score, x[2]))
top = [x for x in adjusted if x[1] == adjusted[0][1] ]
return top
def test_only_identical_strings_extracted(scorer, processor, data):
"""
Test that only identical (post processing) strings score 100 on the test.
If two strings are not identical then using full comparison methods they should
not be a perfect (100) match.
:param scorer:
:param processor:
:param data:
:return:
"""
# Draw a list of random strings
strings = data.draw(
st.lists(
st.text(min_size=10, max_size=100, alphabet=HYPOTHESIS_ALPHABET),
min_size=1,
max_size=10)
)
# Draw a random integer for the index in that list
choiceidx = data.draw(st.integers(min_value=0, max_value=(len(strings) - 1)))
# Extract our choice from the list
choice = strings[choiceidx]
# Check process doesn't make our choice the empty string
assume(processor(choice) != '')
# Extract all perfect matches
result = process.extractBests(choice,
strings,
scorer=scorer,
processor=processor,
score_cutoff=100,
limit=None)
# Check we get a result
assert result != []
# Check THE ONLY result(s) we get are a perfect match for the (processed) original data
pchoice = processor(choice)
for r in result:
assert pchoice == processor(r[0])
def extrairCitadosTextoBestExtract(link,f):
listaFamosos = model.buscarListaFamoso()
noticiaDB = model.buscarNoticia(link)
if noticiaDB != False and listaFamosos != False:
f.writelines(str(noticiaDB[1]) + "\n")
noticia = {'id': noticiaDB[0], 'titulo': noticiaDB[1], 'subtitulo': noticiaDB[2], 'link': noticiaDB[3], 'tipo': noticiaDB[4], 'texto': noticiaDB[5]}
print(str(noticia['titulo']))
famosos = []
for j in listaFamosos:
famosos.append(j[1])
textoTotal = noticia['titulo'] + " " + noticia['subtitulo'] + " " + noticia['texto']
result = process.extractBests(textoTotal, famosos,None,None,57,limit=10000)
for r in result:
f.writelines(str(r[0])+","+str(round(r[1],1))+"\n")
return result
def pdf(pdf_files, shared_pdf, bibtex_folder, bibtex_files, gscholar):
for pdf in pdf_files:
txt = re.sub("\W", " ", gs.convert_pdf_to_txt(pdf)).lower()
#Research determined that the cutting of 35 words gives the
#highest accuracy
words = txt.strip().split()[:35]
words = " ".join(words)
print words
if gscholar == True:
bib = load(gs.pdflookup(pdf, all, gs.FORMAT_BIBTEX)[0])
keys = bib.entries[0].keys()
matching = [s for s in keys if "pdf" in s]
if len(matching) - 1 <= 0:
key = 'pdf'
else:
key = 'pdf' + str(len(matching))
#link = os.symlink(pdf, str(shared_pdf) + str(pdf.split('/')[-1]))
bib.entries = [add_field(bib.entries[0], key, bib)]
bibtex(bib.entries, bibtex_folder, bib)
sleep(random.uniform(1, 2))
else:
best_match = process.extractBests(words, bibtex_files, limit=1)
print best_match
if best_match:
bib = best_match[0][0]
score = best_match[0][1]
#Research determined that matching score of 45
#gives the highest accuracy
if score > 45:
with open(bib, 'r') as f:
db = load(f)
entries = db.entries[0]
keys = entries.keys()
matching = [s for s in keys if "pdf" in s]
if len(matching) - 1 <= 0:
key = 'pdf'
else:
key = 'pdf' + str(len(matching))
entries = add_field(entries, key, bib)
with open(bib, 'w') as f:
f.write(writer._entry_to_bibtex(entries))
def extrairCitadosTexto(threadName, queue, listaFamosos):
queueLock.acquire()
while not workQueue.empty():
n = queue.get()
print("Queue: " + str(workQueue.qsize()))
queueLock.release()
noticia = {'id': n[0], 'titulo': n[1], 'subtitulo': n[2], 'link': n[3], 'tipo': n[4], 'texto': n[5]}
tempoBest = time.clock()
textoTotal = noticia['titulo'] + " " + noticia['subtitulo'] + " " + noticia['texto']
bestExtract = process.extractBests(textoTotal, listaFamosos,None,None,57,limit=10000)
print("tempoBest:" + str(time.clock() - tempoBest))
tempoLoop = time.clock()
for j in bestExtract:
ratioTotal = 0
mediaRatio = 0
nomeCompleto = j[0]
todosNomes = nomeCompleto.split(" ")
textoTotal = noticia['titulo'] + " " + noticia['subtitulo'] + " " + noticia['texto']
ratioTotal += fuzz.token_set_ratio(nomeCompleto, textoTotal)
if len(todosNomes) > 1:
for nome in todosNomes:
ratioTotal += fuzz.token_set_ratio(nome, textoTotal)
if len(todosNomes) == 1:
mediaRatio = ratioTotal
elif len(todosNomes) > 1:
mediaRatio = ratioTotal / (len(todosNomes) + 1)
if mediaRatio > 85:
model.relacionarFamosoNoticia(nomeCompleto, noticia['id'])
print("tempo loop:" + str(time.clock() - tempoLoop))
print(str(threadName) + " - " + str(noticia['titulo']))
queueLock.acquire()
queueLock.release()
outlist[21] = row[11] # Room (raw location from input)
outlist[27] = row[6] # Manufacturer website
outlist[29] = row[18] # training_required?
outlist[37] = row[7] # Asset ID
outlist[38] = row[28] # Finance ID
outlist[39] = row[5] # Serial No
outlist[42] = row[24] # Date of Purchase
outlist[43] = row[25] # Purchase_cost
outlist[45] = row[23] # end_of_life
outlist[46] = row[22] # maintenance
outlist[49] = row[8] # comments
# Fuzzy match building:
if not row[11]:
row[11] = 'Unknown'
# a -- results from fuzzy matching
a = process.extractBests(row[11], buildings, limit=2)
if a[0][0] == "SMC":
outlist[20] = 'Scottish Microelectronics Centre'
else:
outlist[20] = a[0][0];
# Fuzzy match names and emails:
# Contact 1:
flat_names = flatten_dict(names)
b = process.extractBests(row[12],flat_names, limit=2, scorer=fuzz.token_set_ratio)
custodian_score = b[0][1]
outlist[22] = b[0][0] # contact 1 name
outlist[23] = names[b[0][0]] # contact 1 email
# Contact 2:
if row[13]:
c = process.extractBests(row[13],flat_names, limit=2, scorer=fuzz.token_set_ratio)
technical_score = c[0][1]
def handle_query(self):
self.results.clear()
res = extractBests(self.query.text(), self.all_results, limit=8)
for r, s in res:
self.results.add(r)
self.results.setCurrentRow(0)
def search_episode_by_filename(filename):
# Split out file extension
basename, extension = os.path.splitext(filename)
# Remove common phrases not part of the title
searchname = re.sub(r"Tatort", '', basename)
# Find match
match_results = process.extractBests(searchname, tatort_titles,
score_cutoff = 60, limit = 5)
# no match was found
if not match_results:
print("No match was found for file {}".format(filename))
return
# only one match was found with the minimum required score
matching_episode = None
if len(match_results) == 1:
chosen_result = match_results[0]
# multiple matches were found above the score threshold: ask the user
# which one is right
if len(match_results) > 1:
if match_results[0][1] - match_results[1][1] > 20:
# if choice 0 is 20 points more likely than choice 1, we directly
# use the first choice
chosen_result = match_results[0]
else:
# print choices
print("Multiple matches were found for file {}".format(filename))
print("Please choose the correct one from the list below.")
for index, match_result in enumerate(match_results):
(matching_title, matching_score, matching_id) = match_result
episode = tatort_episodes[matching_id]
print("{index}: {name} (score: {score:02d}/100)".format(
index = index, name = episode['episodename'],
score = matching_score))
# let user choose
chosen_id = int(input('Your choice: '))
# FIXME: repeat on wrong inputs
chosen_result = match_results[chosen_id]
(matching_title, matching_score, matching_id) = chosen_result
matching_episode = tatort_episodes[matching_id]
# build new file name
try:
absolute_number = int(matching_episode['absolute_number'])
except:
absolute_number = 0
new_filename = "Tatort {:04d} - {:02d}x{:02d} - {}{}".format(
absolute_number, int(matching_episode['seasonnumber']),
int(matching_episode['episodenumber']),
matching_episode['episodename'], extension)
new_filename = new_filename.replace('/', ' ')
print("{} -> {}".format(filename, new_filename))
os.rename(filename, new_filename)