def __next__(self):
while True:
if self.test_limit is not None:
self.test_limit -= 1
if self.test_limit <= 0:
raise StopIteration
line = self.fh.readline().strip()
if line == "":
raise StopIteration
if line == "page_title":
continue
basename = line.strip()
three_dirs, filename = to3dirs.get_path_file(basename)
path = os.path.join(self.dest_dir, three_dirs)
disk_name = os.path.join(path, filename)
if not os.path.exists(disk_name.encode('utf-8')):
if not os.path.exists(path.encode('utf-8')):
os.makedirs(path.encode('utf-8'))
quoted_url = urllib.parse.quote(basename)
# Skip wikipedia automatic redirect
wiki = WIKI % dict(lang=self.language)
url = wiki + "w/index.php?title=%s&redirect=no" % (
quoted_url, )
data = DataURLs(url=url,
temp_dir=self.temp_dir,
disk_name=disk_name,
basename=basename)
return data
def scrap_portal(language, lang_config):
"""Get the portal index and scrap it."""
# get the portal url, get out if don't have it
portal_index_title = lang_config.get('portal_index')
if portal_index_title is None:
logger.info("Not scraping portals, url not configured.")
return
logger.info("Scraping portal main page %s", portal_index_title)
with NamedTemporaryFile('wt',
encoding='utf8',
dir='/tmp/',
prefix='cdpedia-') as tf:
tf.write(portal_index_title + '\n')
tf.flush()
_call_scraper(language, tf.name)
dir3, quoted_page = to3dirs.get_path_file(portal_index_title)
portal_filepath = os.path.join(location.articles, dir3, quoted_page)
logger.info("Parsing portal page")
with open(portal_filepath, 'rt', encoding='utf8') as fh:
soup = bs4.BeautifulSoup(fh, features="html.parser")
cnt = 0
_path = os.path.join(location.langdir, PORTAL_PAGES)
with open(_path, 'wt', encoding='utf8') as fh:
for page in preprocessors.extract_pages(soup):
cnt += 1
fh.write(page + '\n')
logger.info("Scraping portal sub pages (total=%d)", cnt)
_call_scraper(language, _path)
logger.info("Portal scraping done")
def to_filename(title):
"""Compute the filename from the title."""
tt = title.replace(" ", "_")
if len(tt) >= 2:
tt = tt[0].upper() + tt[1:]
elif len(tt) == 1:
tt = tt[0].upper()
else:
raise ValueError("Title must have at least one character")
dir3, arch = to3dirs.get_path_file(tt)
expected = os.path.join(dir3, arch)
return expected
def top_pages(self):
"""Calculate the HTMLs with more score and store both lists."""
all_pages = []
colsep = config.SEPARADOR_COLUMNAS
with PATH_TEMP.joinpath("page_scores_final.txt").open(
"rt", encoding='utf8') as fh:
for line in fh:
page, score = line.strip().split(colsep)
dir3, fname = to3dirs.get_path_file(page)
all_pages.append((dir3, fname, int(score)))
# order by score, and get top N
all_pages.sort(key=operator.itemgetter(2), reverse=True)
return all_pages
def calculate(self):
"""Calculate the HTMLs with more score and store both lists."""
self._calculated = True
# read the preprocessed file
all_pages = []
colsep = config.SEPARADOR_COLUMNAS
with codecs.open(LOG_SCORES_FINAL, 'rt', encoding='utf8') as fh:
for line in fh:
page, score = line.strip().split(colsep)
dir3, fname = to3dirs.get_path_file(page)
all_pages.append((dir3, fname, page, int(score)))
# order by score, and get top N
all_pages.sort(key=operator.itemgetter(3), reverse=True)
page_limit = config.imageconf['page_limit']
self._top_pages = all_pages[:page_limit]
# get all items after N that still has the same score that last one
last_score = self._top_pages[-1][3]
for more_info in all_pages[page_limit:]:
if more_info[3] == last_score:
self._top_pages.append(more_info)
separator = config.SEPARADOR_COLUMNAS
if os.path.exists(config.PAG_ELEGIDAS):
# previous run for this info! same content?
with codecs.open(config.PAG_ELEGIDAS, "rt", "utf8") as fh:
old_stuff = []
for linea in fh:
dir3, arch, score = linea.strip().split(separator)
old_stuff.append((dir3, arch, int(score)))
if sorted(old_stuff) == sorted(self._top_pages):
self._same_info_through_runs = True
if not self._same_info_through_runs:
# previous info not there, or different: write to disk
with codecs.open(config.PAG_ELEGIDAS, "wt", "utf8") as fh:
for dir3, fname, page, score in self._top_pages:
info = (dir3, page, str(score))
fh.write(separator.join(info) + "\n")
def get_data_urls(listado_nombres, dest_dir, language, test_limit=None):
"""Get a list of DataURLs to download (verifying which are already downloaded)."""
logger.info('Generando DataURLs')
wiki_base = WIKI.format(language=language)
temp_dir = dest_dir + ".tmp"
if not os.path.exists(temp_dir):
os.makedirs(temp_dir)
data_urls_list = []
previous_count = 0
with open(listado_nombres, 'rt', encoding='utf-8') as fh:
for line in fh:
if test_limit is not None:
test_limit -= 1
if test_limit <= 0:
break
line = line.strip()
if line == "page_title":
continue
basename = line.strip()
three_dirs, filename = to3dirs.get_path_file(basename)
path = os.path.join(dest_dir, three_dirs)
disk_name = os.path.join(path, filename)
if os.path.exists(disk_name):
previous_count += 1
continue
if not os.path.exists(path):
os.makedirs(path)
quoted_url = urllib.parse.quote(basename)
# Skip wikipedia automatic redirect
url = "{}/w/index.php?title={}&redirect=no".format(wiki_base, quoted_url)
data = DataURLs(url=url, temp_dir=temp_dir, disk_name=disk_name, basename=basename)
data_urls_list.append(data)
return (previous_count, data_urls_list)
def test_encoding(self):
r = to3dirs.get_path_file("2.3")
self.assertEqual(r, ("2/%/2", "2%2E3"))
def test_very_short(self):
r = to3dirs.get_path_file("m")
self.assertEqual(r, ("m/_/_", "m"))
def test_short(self):
r = to3dirs.get_path_file("mo")
self.assertEqual(r, ("m/o/_", "mo"))
def test_simple(self):
r = to3dirs.get_path_file("moño")
self.assertEqual(r, ("m/o/ñ", "moño"))
