def __init__(self, name, spec, item_schemas, all_extractors, settings=None,
**kw):
super(IblSpider, self).__init__(name, **kw)
spec = deepcopy(spec)
for key, val in kw.items():
if isinstance(val, basestring) and key in STRING_KEYS:
val = val.splitlines()
spec[key] = val
self._item_template_pages = sorted(
((t['scrapes'], t) for t in spec['templates']
if t.get('page_type', 'item') == 'item'), key=itemgetter(0))
self._templates = [templ for _, templ in self._item_template_pages]
self.plugins = IndexedDict()
for plugin_class, plugin_name in zip(load_plugins(settings),
load_plugin_names(settings)):
instance = plugin_class()
instance.setup_bot(settings, spec, item_schemas, all_extractors)
self.plugins[plugin_name] = instance
self.login_requests = []
self.form_requests = []
self._start_requests = []
self.generic_form = GenericForm(**kw)
self._create_init_requests(spec.get("init_requests", []))
self._process_start_urls(spec)
self.allowed_domains = spec.get(
'allowed_domains',
self._get_allowed_domains(self._templates)
)
if not self.allowed_domains:
self.allowed_domains = None
def _configure_plugins(self, settings, spec, schemas, extractors):
plugins = IndexedDict()
for plugin_class, plugin_name in zip(load_plugins(settings),
load_plugin_names(settings)):
instance = plugin_class()
instance.setup_bot(settings, spec, schemas, extractors)
plugins[plugin_name] = instance
return plugins
def _configure_plugins(self, settings, spec, schemas, extractors):
plugins = IndexedDict()
for plugin_class, plugin_name in zip(load_plugins(settings),
load_plugin_names(settings)):
instance = plugin_class()
instance.setup_bot(settings, spec, schemas, extractors, self.logger)
plugins[plugin_name] = instance
return plugins
def __init__(self,
name,
spec,
item_schemas,
all_extractors,
settings=None,
**kw):
super(IblSpider, self).__init__(name, **kw)
self._job_id = settings.get('JOB', '')
spec = deepcopy(spec)
for key, val in kw.items():
if isinstance(val, six.string_types) and key in STRING_KEYS:
val = val.splitlines()
spec[key] = val
self._item_template_pages = sorted(
((t['scrapes'], t) for t in spec['templates']
if t.get('page_type', 'item') == 'item'),
key=itemgetter(0))
self._templates = [templ for _, templ in self._item_template_pages]
self.plugins = IndexedDict()
for plugin_class, plugin_name in zip(load_plugins(settings),
load_plugin_names(settings)):
instance = plugin_class()
instance.setup_bot(settings, spec, item_schemas, all_extractors)
self.plugins[plugin_name] = instance
self.js_enabled = False
self.SPLASH_HOST = None
if settings.get('SPLASH_URL'):
self.SPLASH_HOST = urlparse(settings.get('SPLASH_URL')).hostname
self.js_enabled = spec.get('js_enabled', False)
if self.js_enabled and (settings.get('SPLASH_PASS') is not None
or settings.get('SPLASH_USER') is not None):
self.splash_auth = basic_auth_header(
settings.get('SPLASH_USER', ''),
settings.get('SPLASH_PASS', ''))
self._filter_js_urls = self._build_js_url_filter(spec)
self.login_requests = []
self.form_requests = []
self._start_requests = []
self.generic_form = GenericForm(**kw)
self._create_init_requests(spec.get("init_requests", []))
self._process_start_urls(spec)
self.allowed_domains = spec.get(
'allowed_domains', self._get_allowed_domains(self._templates))
self.page_actions = spec.get('page_actions', [])
if not self.allowed_domains:
self.allowed_domains = None
def __init__(self, name, spec, item_schemas, all_extractors, settings=None,
**kw):
super(IblSpider, self).__init__(name, **kw)
self._job_id = settings.get('JOB', '')
spec = deepcopy(spec)
for key, val in kw.items():
if isinstance(val, six.string_types) and key in STRING_KEYS:
val = val.splitlines()
spec[key] = val
self._item_template_pages = sorted(
((t['scrapes'], t) for t in spec['templates']
if t.get('page_type', 'item') == 'item'), key=itemgetter(0))
self._templates = [templ for _, templ in self._item_template_pages]
self.plugins = IndexedDict()
for plugin_class, plugin_name in zip(load_plugins(settings),
load_plugin_names(settings)):
instance = plugin_class()
instance.setup_bot(settings, spec, item_schemas, all_extractors)
self.plugins[plugin_name] = instance
self.js_enabled = False
self.SPLASH_HOST = None
if settings.get('SPLASH_URL'):
self.SPLASH_HOST = urlparse(settings.get('SPLASH_URL')).hostname
self.js_enabled = spec.get('js_enabled', False)
if self.js_enabled and (settings.get('SPLASH_PASS') is not None or
settings.get('SPLASH_USER') is not None):
self.splash_auth = basic_auth_header(
settings.get('SPLASH_USER', ''),
settings.get('SPLASH_PASS', ''))
self._filter_js_urls = self._build_js_url_filter(spec)
self.login_requests = []
self.form_requests = []
self._start_requests = []
self.generic_form = GenericForm(**kw)
self._create_init_requests(spec.get("init_requests", []))
self._process_start_urls(spec)
self.allowed_domains = spec.get(
'allowed_domains',
self._get_allowed_domains(self._templates)
)
self.page_actions = spec.get('page_actions', [])
if not self.allowed_domains:
self.allowed_domains = None
def __init__(self,
name,
spec,
item_schemas,
all_extractors,
settings=None,
**kw):
super(IblSpider, self).__init__(name, **kw)
spec = deepcopy(spec)
for key, val in kw.items():
if isinstance(val, basestring) and key in STRING_KEYS:
val = val.splitlines()
spec[key] = val
self._item_template_pages = sorted(
((t['scrapes'], t) for t in spec['templates']
if t.get('page_type', 'item') == 'item'),
key=itemgetter(0))
self._templates = [templ for _, templ in self._item_template_pages]
self.plugins = IndexedDict()
for plugin_class, plugin_name in zip(load_plugins(settings),
load_plugin_names(settings)):
instance = plugin_class()
instance.setup_bot(settings, spec, item_schemas, all_extractors)
self.plugins[plugin_name] = instance
self.login_requests = []
self.form_requests = []
self._start_requests = []
self.generic_form = GenericForm(**kw)
self._create_init_requests(spec.get("init_requests", []))
self._process_start_urls(spec)
self.allowed_domains = spec.get(
'allowed_domains', self._get_allowed_domains(self._templates))
if not self.allowed_domains:
self.allowed_domains = None
