def write_hesuvi_zip():
manufacturers = ManufacturerIndex()
zip_object = ZipFile(os.path.join(DIR_PATH, 'hesuvi.zip'), 'w')
dir_paths = [
os.path.join(DIR_PATH, 'oratory1990'),
os.path.join(DIR_PATH, 'crinacle', 'gras_43ag-7_harman_over-ear_2018'),
os.path.join(DIR_PATH, 'crinacle', 'ears-711_harman_over-ear_2018'),
os.path.join(DIR_PATH, 'innerfidelity'),
os.path.join(DIR_PATH, 'rtings'),
os.path.join(DIR_PATH, 'headphonecom'),
os.path.join(DIR_PATH, 'referenceaudioanalyzer'),
]
zip_files = set()
for dir_path in dir_paths:
for fp in glob(os.path.join(dir_path, '**', '* GraphicEQ.txt'),
recursive=True):
_, name = os.path.split(fp)
name = name.replace(' GraphicEQ.txt', '')
if re.search(MOD_REGEX, name, flags=re.IGNORECASE):
# Skip samples, there are averaged results available
continue
manufacturer, _ = manufacturers.find(name)
if manufacturer is None:
print(f'Manufacturer could not be found for {name}')
continue
name = manufacturers.model(name)
arcname = f'eq/{manufacturer}/{name}.txt'
if arcname in zip_files:
# Skip duplicates
continue
with open(fp, 'r', encoding='utf-8') as fh:
s = fh.read()
data = np.array(
[x.split() for x in s.split(': ')[1].split('; ')],
dtype='float')
sl = np.logical_and(data[:, 0] > 100, data[:, 0] < 10000)
data[:, 1] -= np.mean(data[sl, 1])
s = 'GraphicEQ: '
s += '; '.join([f'{x[0]:.0f} {x[1]:.1f}' for x in data])
zip_object.writestr(arcname, s)
zip_files.add(arcname)
zip_object.close()
def write_hesuvi_index():
os.makedirs(os.path.join(DIR_PATH, 'hesuvi'), exist_ok=True)
manufacturers = ManufacturerIndex()
zip_object = ZipFile(os.path.join(DIR_PATH, 'hesuvi.zip'), 'w')
dir_paths = [
os.path.join(DIR_PATH, 'oratory1990'),
os.path.join(DIR_PATH, 'crinacle', 'harman_in-ear_2019v2'),
os.path.join(DIR_PATH, 'crinacle', 'crinacl_over-ear'),
os.path.join(DIR_PATH, 'innerfidelity'),
os.path.join(DIR_PATH, 'rtings'),
os.path.join(DIR_PATH, 'headphonecom'),
]
zip_files = set()
for dir_path in dir_paths:
for fp in glob(os.path.join(dir_path, '**', '* GraphicEQ.txt'),
recursive=True):
_, name = os.path.split(fp)
name = name.replace(' GraphicEQ.txt', '')
if re.search(r' \(?(sample |sn)[a-zA-Z0-9]+\)?$',
name,
flags=re.IGNORECASE):
# Skip samples, there are averaged results available
continue
manufacturer, _ = manufacturers.find(name)
name = manufacturers.model(name)
arcname = f'eq/{manufacturer}/{name}.txt'
if arcname in zip_files:
# Skip duplicates
continue
with open(fp, 'r', encoding='utf-8') as fh:
s = fh.read()
data = np.array(
[x.split() for x in s.split(': ')[1].split('; ')],
dtype='float')
sl = np.logical_and(data[:, 0] > 100, data[:, 0] < 10000)
data[:, 1] -= np.mean(data[sl, 1])
s = 'GraphicEQ: '
s += '; '.join([f'{x[0]:.0f} {x[1]:.1f}' for x in data])
zip_object.writestr(arcname, s)
zip_files.add(arcname)
zip_object.close()
class Crawler(ABC):
def __init__(self, driver=None):
self.driver = driver
self.name_index = self.read_name_index()
self.manufacturers = ManufacturerIndex()
self.name_proposals = None
self.init_name_proposals()
self.existing = self.get_existing()
self.urls = self.get_urls()
# UI
self.prompts = widgets.VBox([])
self.iframe = widgets.VBox([])
self.widget = widgets.HBox([self.prompts, self.iframe])
@staticmethod
@abstractmethod
def read_name_index():
"""Reads name index as Index
Returns:
NameIndex
"""
pass
def init_name_proposals(self):
"""Gets name proposals for new measurements
Returns:
NameIndex
"""
name_proposals = NameIndex()
for db in [
'crinacle', 'oratory1990', 'rtings', 'referenceaudioanalyzer'
]:
name_index = NameIndex.read_tsv(
os.path.join(DIR_PATH, db, 'name_index.tsv'))
name_proposals.concat(name_index)
for db in ['innerfidelity', 'headphonecom']:
name_index = NameIndex.read_files(
os.path.join(DIR_PATH, db, 'data', '**', '*.csv'))
name_proposals.concat(name_index)
name_proposals.remove_duplicates()
manufacturer_pattern = rf'^({"|".join([m[0] for m in self.manufacturers.manufacturers])})'
proposal_data = {'form': [], 'manufacturer': [], 'model': []}
for item in name_proposals.items:
if not item.true_name or item.form == 'ignore':
continue
manufacturer = re.search(manufacturer_pattern,
item.true_name,
flags=re.IGNORECASE)
if not manufacturer:
continue
manufacturer = manufacturer[0]
proposal_data['form'].append(item.form)
proposal_data['manufacturer'].append(manufacturer)
proposal_data['model'].append(
item.true_name.replace(manufacturer, '').strip())
self.name_proposals = pd.DataFrame(proposal_data)
@abstractmethod
def write_name_index(self):
"""Writes name index to a file
Returns:
Index
"""
pass
@staticmethod
@abstractmethod
def get_existing():
"""Reads existing files as Index
Returns:
Index
"""
pass
@abstractmethod
def get_urls(self):
"""Crawls measurement URLs
Returns:
Dict where headphone names are keys and URLs are values
"""
pass
@abstractmethod
def process(self, item, url):
"""Downloads a single URL and processes it
Args:
item: Item
url: URL to measurement
Returns:
None
"""
pass
def update_name_index(self, item):
"""Updates name index"""
exact_match = self.name_index.find_one(false_name=item.false_name,
true_name=item.true_name,
form=item.form)
if not exact_match:
self.name_index.update(item, false_name=item.false_name)
self.write_name_index()
def prompt_callback(self, false_name, url):
def callback(true_name, form):
if form == 'ignore':
self.update_name_index(NameItem(false_name, None, form))
return
item = NameItem(false_name, true_name, form)
try:
self.process(NameItem(false_name, true_name, form), url)
except FileNotFoundError as err:
print(err)
return
self.update_name_index(item)
return callback
def process_new(self, prompt=True):
"""Processes all new measurements
Updates name index with the new entries now found in the name index previously.
Returns:
None
"""
prompts = []
unknown_manufacturers = []
for false_name, url in self.urls.items():
item = self.name_index.find_one(false_name=false_name)
if item and item.form == 'ignore':
continue
if not item:
if not prompt:
print(
f'{false_name} is unknown and prompting is prohibited, skipping the item.'
)
continue
# Name doesn't exist in the name index
intermediate_name = self.intermediate_name(false_name)
manufacturer, manufacturer_match = self.manufacturers.find(
intermediate_name)
if manufacturer:
model = re.sub(re.escape(manufacturer_match),
'',
intermediate_name,
flags=re.IGNORECASE).strip()
name_proposals = self.get_name_proposals(false_name)
similar_names = self.get_name_proposals(
false_name, n=6, normalize_digits=True, threshold=0)
similar_names = [
item.true_name for item in similar_names.items
]
else:
unknown_manufacturers.append(intermediate_name)
model = intermediate_name
name_proposals = None
similar_names = None
# Not sure about the name, ask user
prompts.append(
NamePrompt(model,
self.prompt_callback(false_name, url),
manufacturer=manufacturer,
name_proposals=name_proposals,
search_callback=self.search,
false_name=false_name,
similar_names=similar_names).widget)
else:
existing = self.existing.find_one(true_name=item.true_name)
if not existing:
# Name found in name index but the measurement doesn't exist
self.process(item, url)
if len(unknown_manufacturers) > 0:
print('Headphones with unknown manufacturers\n ' +
'\n '.join(unknown_manufacturers))
print('Add them to manufacturers.tsv and run this cell again')
self.prompts.children = prompts
def search(self, name):
quoted = urllib.parse.quote_plus(name)
url = f'https://google.com/search?q={quoted}&tbm=isch'
webbrowser.open(url)
def get_name_proposals(self,
false_name,
n=4,
normalize_digits=False,
normalize_extras=False,
threshold=60):
"""Prompts manufacturer, model and form from the user
Args:
false_name: Name as it exists in the measurement source
n: Number of proposals to return
normalize_digits: Normalize all digits to zeros before calculating fuzzy string matching score
normalize_extras: Remove extra details in the parentheses
threshold: Score threshold
Returns:
NameItem
"""
def fuzzy(fn, a, b):
a = a.lower()
b = b.lower()
if normalize_digits:
a = re.sub(r'\d', '0', a).strip()
b = re.sub(r'\d', '0', b).strip()
if normalize_extras:
a = re.sub(r'\(.+\)$', '', a).strip()
b = re.sub(r'\(.+\)$', '', b).strip()
return fn(a, b)
manufacturer, manufacturer_match = self.manufacturers.find(false_name)
if not manufacturer:
return NameIndex([])
false_model = re.sub(re.escape(manufacturer_match),
'',
false_name,
flags=re.IGNORECASE).strip()
# Select only the items with the same manufacturer
models = self.name_proposals[self.name_proposals.manufacturer ==
manufacturer]
# Calculate ratios
partial_ratios = [
fuzzy(fuzz.partial_ratio, model, false_model)
for model in models.model.tolist()
]
ratios = [
fuzzy(fuzz.ratio, model, false_model)
for model in models.model.tolist()
]
models = models.assign(partial_ratio=partial_ratios)
models = models.assign(ratio=ratios)
models = models[models.partial_ratio >= threshold]
models.sort_values('ratio', ascending=False, inplace=True)
proposals = []
for i, row in models.iterrows():
proposals.append(
NameItem(None, f'{manufacturer} {row.model}', row.form))
ni = NameIndex(items=proposals)
ni.df = ni.df.head(n)
return ni
def intermediate_name(self, false_name):
"""Gets intermediate name with false name."""
return false_name
@staticmethod
def download(url, true_name, output_dir, file_type=None):
"""Downloads a file from a URL
Args:
url: URL to download
true_name: True name of the item to download
output_dir: Where to write the downloaded file
file_type: File extension. Detected automatically if None.
Returns:
Bool depicting if download succeeded or not
"""
output_dir = os.path.abspath(output_dir)
os.makedirs(output_dir, exist_ok=True)
res = requests.get(url, stream=True)
if res.status_code != 200:
print(f'Failed to download "{true_name}" at "{url}"')
return None
if file_type is None:
file_type = url.split('.')[-1]
file_type = file_type.split('?')[0]
file_path = os.path.join(output_dir, f'{true_name}.{file_type}')
with open(file_path, 'wb') as f:
res.raw.decode_content = True
shutil.copyfileobj(res.raw, f)
print('Downloaded to "{}"'.format(file_path))
return file_path
def get_beautiful_soup(self, url):
self.driver.get(url)
sleep(1) # Giving some time for Selenium to render the page
html = self.driver.find_element_by_tag_name('html').get_attribute(
'outerHTML')
return BeautifulSoup(html, 'html.parser')
class Crawler(ABC):
def __init__(self, driver=None):
self.driver = driver
self.name_index = self.read_name_index()
self.name_proposals = self.get_name_proposals()
self.existing = self.get_existing()
self.urls = self.get_urls()
self.manufacturers = ManufacturerIndex()
@staticmethod
@abstractmethod
def read_name_index():
"""Reads name index as Index
Returns:
NameIndex
"""
pass
def get_name_proposals(self):
"""Gets name proposals for new measurements
Returns:
NameIndex
"""
name_proposals = NameIndex()
for db in [
'crinacle', 'oratory1990', 'rtings', 'referenceaudioanalyzer'
]:
name_index = NameIndex.read_tsv(
os.path.join(DIR_PATH, db, 'name_index.tsv'))
name_proposals.concat(name_index)
for db in ['innerfidelity', 'headphonecom']:
name_index = NameIndex.read_files(
os.path.join(DIR_PATH, db, 'data', '**', '*.csv'))
name_proposals.concat(name_index)
return name_proposals
@abstractmethod
def write_name_index(self):
"""Writes name index to a file
Returns:
Index
"""
pass
@staticmethod
@abstractmethod
def get_existing():
"""Reads existing files as Index
Returns:
Index
"""
pass
@abstractmethod
def get_urls(self):
"""Crawls measurement URLs
Returns:
Dict where headphone names are keys and URLs are values
"""
pass
@abstractmethod
def process(self, item, url):
"""Downloads a single URL and processes it
Args:
item: Item
url: URL to measurement
Returns:
None
"""
pass
@staticmethod
def prompt_true_name(name_options):
"""Prompts true name from the user."""
name_options = name_options if name_options is not None else []
if 'skip' not in name_options:
name_options.insert(0, 'skip')
s = 'What is it\'s true name?'
if len(name_options):
s += ' Select a number or write the name if none of the options.'
print(s)
if len(name_options):
print(f'\n'.join(f'[{i}] {o}' for i, o in enumerate(name_options)))
while True:
name = input('> ')
try:
name = name_options[int(name)]
if name == 'skip':
return None
break
except (KeyError, ValueError):
break
except IndexError:
print('That didn\'t work, try again.')
return name
@staticmethod
def prompt_manufacturer(name_options):
"""Prompts true manufacturer from the user."""
name_options = name_options if name_options is not None else []
s = 'What is it\'s true manufacturer name?'
if len(name_options):
s += ' Select a number or write the name if none of the options.'
print(s)
if len(name_options):
print(f'\n'.join(f'[{i + 1}] {o}'
for i, o in enumerate(name_options)))
while True:
name = input('> ')
try:
name = name_options[int(name) - 1]
break
except (KeyError, ValueError):
break
except IndexError:
print('That didn\'t work, try again.')
print('Which part of the name to replace')
replace = input('> ')
return name, replace
@staticmethod
def prompt_form():
"""Prompts form from the user."""
options = ['onear', 'inear', 'earbud']
print('What is it\'s type?')
print(f'\n'.join(f'[{i + 1}] {o}' for i, o in enumerate(options)))
while True:
form = input('> ')
try:
return options[int(form) - 1]
except (IndexError, ValueError):
print('That didn\'t work, try again.')
def prompt(self, false_name):
"""Prompts user for true name and form based on false name."""
form = None
if self.name_proposals is not None:
# Name proposals initialized, add matching entries to options in prompt
matches = []
matches += self.name_proposals.search_by_false_name(false_name)
matches += self.name_proposals.search_by_true_name(false_name)
names_and_ratios = []
for match in matches:
if not match[0].true_name:
# Skip items without true name
continue
if match[1] == 100:
# Exact match
match[0].true_name += ' ✓'
if match[0].true_name not in [x[0] for x in names_and_ratios]:
# New match
names_and_ratios.append(
(match[0].true_name, match[1], match[0].form))
else:
# Existing match, update ratio
for i in range(len(names_and_ratios)):
if match[0].true_name == names_and_ratios[i][
0] and match[1] > names_and_ratios[i][1]:
names_and_ratios[i] = (names_and_ratios[i][0],
match[1],
names_and_ratios[i][2])
name_options = [
x[0] for x in sorted(
names_and_ratios, key=lambda x: x[1], reverse=True)[:4]
]
if false_name not in name_options:
name_options.append(false_name) # Add the false name
# Prompt
true_name = self.prompt_true_name(name_options)
if true_name is None:
return None
# Find and replace true manufacturer name or prompt it
if self.manufacturers.find(true_name)[0] is None:
# Unknown manufacturer, find options with the two first words and prompt it
manufacturer_options = []
for i in range(1, min(3, len(true_name.split()))):
candidate = ' '.join(true_name.split()[:i])
print(candidate)
manufacturer_options += self.manufacturers.search(
candidate)
if candidate not in [x[0] for x in manufacturer_options]:
manufacturer_options.append((candidate, 0))
manufacturer_options = sorted(manufacturer_options,
key=lambda x: x[1],
reverse=True)
manufacturer_options = [x[0] for x in manufacturer_options]
manufacturer, replace = self.prompt_manufacturer(
manufacturer_options)
_, match = self.manufacturers.find(manufacturer)
if match:
# Add as a new variant in existing manufacturer
for m in self.manufacturers.manufacturers:
if m[0] == match:
m.append(replace)
else:
# Add new manufacturer
self.manufacturers.manufacturers.append([manufacturer])
self.manufacturers.write()
# Replace
true_name = self.manufacturers.replace(true_name)
# Find the answer and select form
for name, ratio, f in names_and_ratios:
if true_name == name:
form = f
break
true_name = true_name.replace(' ✓', '')
else:
true_name = self.prompt_true_name([false_name])
form = None
if true_name is None:
# User skipped
return None
if form is None:
# Form not found in name proposals, prompt it
form = self.prompt_form()
return NameItem(false_name, true_name, form)
def process_new(self, prompt=True):
"""Processes all new measurements
Updates name index with the new entries now found in the name index previously.
Returns:
None
"""
for false_name, url in self.urls.items():
try:
ni = self.name_index.find(false_name=false_name)
item = ni.items[0] if ni else None
if item and item.form == 'ignore':
continue
if item and item.true_name:
# Name index contains the entry
if not self.existing.find(true_name=item.true_name):
# Doesn't exist already
self.process(item, url)
else:
# Unknown item
if prompt:
# Prompt true name and form
print(f'\n"{false_name}" is not known.')
item = self.prompt(false_name)
if item is None:
self.name_index.update(NameItem(
false_name, None, 'ignore'),
false_name=false_name)
continue
self.name_index.update(item, false_name=false_name)
self.process(item, url)
else:
print(
f'"{false_name}" is not known. Add true name and form to name index and run again.'
)
self.name_index.update(NameItem(
false_name, None, None),
false_name=false_name)
self.write_name_index()
except Exception as err:
print(f'Processing failed for "{false_name}"')
raise err
@staticmethod
def download(url, true_name, output_dir, file_type=None):
"""Downloads a file from a URL
Args:
url: URL to download
true_name: True name of the item to download
output_dir: Where to write the downloaded file
file_type: File extension. Detected automatically if None.
Returns:
Bool depicting if download succeeded or not
"""
output_dir = os.path.abspath(output_dir)
os.makedirs(output_dir, exist_ok=True)
res = requests.get(url, stream=True)
if res.status_code != 200:
print(f'Failed to download "{true_name}" at "{url}"')
return None
if file_type is None:
file_type = url.split('.')[-1]
file_type = file_type.split('?')[0]
file_path = os.path.join(output_dir,
'{}.{}'.format(true_name, file_type))
with open(file_path, 'wb') as f:
res.raw.decode_content = True
shutil.copyfileobj(res.raw, f)
print('Downloaded to "{}"'.format(file_path))
return file_path
def get_beautiful_soup(self, url):
self.driver.get(url)
html = self.driver.find_element_by_tag_name('html').get_attribute(
'outerHTML')
return BeautifulSoup(html, 'html.parser')
