Cutlery/calibre
1
0
Fork 0
mirror of https://github.com/kovidgoyal/calibre.git synced 2025-07-09 03:04:10 -04:00
Code Issues Packages Projects Releases Wiki Activity

Make strptime more robust and add --verbose support in feed parsing

Browse Source
This commit is contained in:
Kovid Goyal 2007-10-28 19:50:03 +00:00
parent 50114d70ce
commit c269d0a670
2 changed files with 24 additions and 6 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
src/libprs500/ebooks/lrf/web/convert_from.py
View File

@ -112,7 +112,7 @@ def process_profile(args, options, logger=None):
except ValueError: except ValueError:
raise CommandLineError('Unknown profile: %s\nValid profiles: %s'%(args[1], available_profiles)) raise CommandLineError('Unknown profile: %s\nValid profiles: %s'%(args[1], available_profiles))
profile = DefaultProfile if index == -1 else builtin_profiles[index] profile = DefaultProfile if index == -1 else builtin_profiles[index]
profile = profile(options.username, options.password) profile = profile(logger, options.verbose, options.username, options.password)
if profile.browser is not None: if profile.browser is not None:
options.browser = profile.browser options.browser = profile.browser

28
src/libprs500/ebooks/lrf/web/profiles/__init__.py
View File

@ -78,9 +78,11 @@ class DefaultProfile(object):
######################################################################## ########################################################################
def __init__(self, username=None, password=None): def __init__(self, logger, verbose=False, username=None, password=None):
self.logger = logger
self.username = username self.username = username
self.password = password self.password = password
self.verbose = verbose
self.temp_dir = tempfile.mkdtemp(prefix=__appname__+'_') self.temp_dir = tempfile.mkdtemp(prefix=__appname__+'_')
self.browser = self.get_browser() self.browser = self.get_browser()
self.url = 'file:'+ ('' if iswindows else '//') + self.build_index() self.url = 'file:'+ ('' if iswindows else '//') + self.build_index()
@ -149,7 +151,9 @@ class DefaultProfile(object):
try: try:
src = self.browser.open(url).read() src = self.browser.open(url).read()
except Exception, err: except Exception, err:
print 'Could not fetch feed: %s\nError: %s'%(url, err) self.logger.error('Could not fetch feed: %s\nError: %s'%(url, err))
if self.verbose:
self.logger.exception(' ')
continue continue
articles[title] = [] articles[title] = []
@ -163,7 +167,7 @@ class DefaultProfile(object):
d = { d = {
'title' : item.find('title').string, 'title' : item.find('title').string,
'url' : self.print_version(item.find('guid').string), 'url' : self.print_version(item.find('guid').string),
'timestamp': calendar.timegm(self.strptime(pubdate)), 'timestamp': self.strptime(pubdate),
'date' : pubdate 'date' : pubdate
} }
delta = time.time() - d['timestamp'] delta = time.time() - d['timestamp']
@ -171,6 +175,8 @@ class DefaultProfile(object):
continue continue
except Exception, err: except Exception, err:
if self.verbose:
self.logger.exception('Error parsing article:\n%s'%(item,))
continue continue
try: try:
desc = item.find('description') desc = item.find('description')
@ -205,13 +211,25 @@ class DefaultProfile(object):
@classmethod @classmethod
def strptime(cls, src): def strptime(cls, src):
src = src.strip().split() src = src.replace('GMT', '').replace('UTC', '').strip()
src = src.split()
src[0] = str(cls.DAY_MAP[src[0][:-1]])+',' src[0] = str(cls.DAY_MAP[src[0][:-1]])+','
try: try:
src[2] = str(cls.MONTH_MAP[src[2]]) src[2] = str(cls.MONTH_MAP[src[2]])
except KeyError: except KeyError:
src[2] = str(cls.FULL_MONTH_MAP[src[2]]) src[2] = str(cls.FULL_MONTH_MAP[src[2]])
return time.strptime(' '.join(src), '%w, %d %m %Y %H:%M:%S %Z') fmt = '%w, %d %m %Y %H:%M:%S'
delta = 0
if src[-1].startswith('+') or src[-1].startswith('-'):
delta = src[-1]
src = src[:-1]
hrs, mins = int(delta[1:3]), int(delta[3:5])
delta = 60*(hrs*60 + mins) * (-1 if delta.startswith('-') else 1)
try:
time_t = time.strptime(' '.join(src), fmt)
except ValueError:
time_t = time.strptime(' '.join(src), fmt.replace('%Y', '%y'))
return calendar.timegm(time_t)-delta
def command_line_options(self): def command_line_options(self):
args = [] args = []