def _fetch_data(self, config, provider): """Fetch the latest feed data. :param dict config: RSS resource configuration :param provider: data provider instance, needed as an argument when raising ingest errors :return: fetched RSS data :rtype: str :raises IngestApiError: if fetching data fails for any reason (e.g. authentication error, resource not found, etc.) """ url = config["url"] if config.get("auth_required", False): auth = (config.get("username"), config.get("password")) else: auth = None response = requests.get(url, auth=auth) if response.ok: return response.content else: if response.status_code in (401, 403): raise IngestApiError.apiAuthError(Exception(response.reason), provider) elif response.status_code == 404: raise IngestApiError.apiNotFoundError(Exception(response.reason), provider) else: raise IngestApiError.apiGeneralError(Exception(response.reason), provider)
def _generate_auth_token(self, provider): """ Generates Authentication Token as per the configuration in Ingest Provider. :param provider: dict - Ingest provider details to which the current directory has been configured :type provider: dict :py:class: `superdesk.io.ingest_provider_model.IngestProviderResource` :return: token details if successfully authenticated :rtype: str :raises: IngestApiError.apiGeneralError() if auth_url is missing in the Ingest Provider configuration """ session = requests.Session() session.mount('https://', SSLAdapter()) auth_url = provider.get('config', {}).get('auth_url', None) if not auth_url: raise IngestApiError.apiGeneralError(provider=provider, exception=KeyError( ''' Ingest Provider {} is missing Authentication URL. Please check the configuration. '''.format(provider['name'])) ) payload = { 'username': provider.get('config', {}).get('username', ''), 'password': provider.get('config', {}).get('password', ''), } response = session.get(auth_url, params=payload, verify=False, timeout=30) if response.status_code < 200 or response.status_code >= 300: raise IngestApiError.apiAuthError(provider=provider) tree = etree.fromstring(response.content) # workaround for http mock lib return tree.text
def _fetch_data(self, config, provider): """Fetch the latest feed data. :param dict config: RSS resource configuration :param provider: data provider instance, needed as an argument when raising ingest errors :return: fetched RSS data :rtype: str :raises IngestApiError: if fetching data fails for any reason (e.g. authentication error, resource not found, etc.) """ url = config['url'] if config.get('auth_required', False): auth = (config.get('username'), config.get('password')) else: auth = None response = requests.get(url, auth=auth) if response.ok: return response.content else: if response.status_code in (401, 403): raise IngestApiError.apiAuthError( Exception(response.reason), provider) elif response.status_code == 404: raise IngestApiError.apiNotFoundError( Exception(response.reason), provider) else: raise IngestApiError.apiGeneralError( Exception(response.reason), provider)
def _fetch_data(self, config, provider): url = config['url'] api_key = config['api_key'] last_update = provider.get( 'last_updated', utcfromtimestamp(0)).strftime('%Y-%m-%dT%H:%M:%S') # Results are pagified so we'll read this many at a time offset_jump = 10 params = {'start': last_update, 'limit': offset_jump} headers = {'apikey': api_key} items = [] offset = 0 while True: params['offset'] = offset try: response = requests.get(url, params=params, headers=headers, timeout=30) except requests.exceptions.ConnectionError as err: raise IngestApiError.apiConnectionError(exception=err) if response.ok: # The total number of results are given to us in json, get them # via a regex to read the field so we don't have to convert the # whole thing to json pointlessly item_ident = re.search('\"total\": *[0-9]*', response.text).group() results_str = re.search('[0-9]+', item_ident).group() if results_str is None: raise IngestApiError.apiGeneralError( Exception(response.text), provider) num_results = int(results_str) if num_results > 0: items.append(response.text) if offset >= num_results: return items offset += offset_jump else: if re.match('Error: No API Key provided', response.text): raise IngestApiError.apiAuthError(Exception(response.text), provider) elif response.status_code == 404: raise IngestApiError.apiNotFoundError( Exception(response.reason), provider) else: raise IngestApiError.apiGeneralError( Exception(response.reason), provider) return items
def _fetch_data(self, config, provider): url = config['url'] api_key = config['api_key'] last_update = provider.get('last_updated', utcfromtimestamp(0)).strftime('%Y-%m-%dT%H:%M:%S') # Results are pagified so we'll read this many at a time offset_jump = 10 params = {'start': last_update, 'limit': offset_jump} headers = {'apikey': api_key} items = [] offset = 0 while True: params['offset'] = offset try: response = requests.get(url, params=params, headers=headers, timeout=30) except requests.exceptions.ConnectionError as err: raise IngestApiError.apiConnectionError(exception=err) if response.ok: # The total number of results are given to us in json, get them # via a regex to read the field so we don't have to convert the # whole thing to json pointlessly item_ident = re.search('\"total\": *[0-9]*', response.text).group() results_str = re.search('[0-9]+', item_ident).group() if results_str is None: raise IngestApiError.apiGeneralError( Exception(response.text), provider) num_results = int(results_str) if num_results > 0: items.append(response.text) if offset >= num_results: return items offset += offset_jump else: if re.match('Error: No API Key provided', response.text): raise IngestApiError.apiAuthError( Exception(response.text), provider) elif response.status_code == 404: raise IngestApiError.apiNotFoundError( Exception(response.reason), provider) else: raise IngestApiError.apiGeneralError( Exception(response.reason), provider) return items
def _generate_auth_token(self, provider): """ Generates Authentication Token as per the configuration in Ingest Provider. :param provider: dict - Ingest provider details to which the current directory has been configured :type provider: dict :py:class: `superdesk.io.ingest_provider_model.IngestProviderResource` :return: token details if successfully authenticated :rtype: str :raises: IngestApiError.apiGeneralError() if auth_url is missing in the Ingest Provider configuration """ session = requests.Session() session.mount("https://", SSLAdapter()) auth_url = provider.get("config", {}).get("auth_url", None) if not auth_url: raise IngestApiError.apiGeneralError( provider=provider, exception=KeyError( """ Ingest Provider {} is missing Authentication URL. Please check the configuration. """.format( provider["name"] ) ), ) payload = { "username": provider.get("config", {}).get("username", ""), "password": provider.get("config", {}).get("password", ""), } response = session.get(auth_url, params=payload, verify=False, timeout=30) if response.status_code < 200 or response.status_code >= 300: try: response.raise_for_status() except Exception: err = IngestApiError.apiAuthError(provider=provider) self.close_provider(provider, err, force=True) raise err tree = etree.fromstring(response.content) # workaround for http mock lib return tree.text
def test_raise_apiAuthError(self): with assert_raises(IngestApiError) as error_context: ex = Exception("Testing API authorization error") raise IngestApiError.apiAuthError(ex, self.provider) exception = error_context.exception self.assertEqual(exception.code, 4007) self.assertEqual(exception.message, "API authorization error") self.assertEqual(exception.provider_name, "TestProvider") self.assertIsNotNone(exception.system_exception) self.assertEqual(exception.system_exception.args[0], "Testing API authorization error") self.assertEqual(len(self.mock_logger_handler.messages['error']), 1) self.assertEqual( self.mock_logger_handler.messages['error'][0], "IngestApiError Error 4007 - API authorization error: " "Testing API authorization error on channel TestProvider")
def test_raise_apiAuthError(self): with assert_raises(IngestApiError) as error_context: ex = Exception("Testing API authorization error") raise IngestApiError.apiAuthError(ex, self.provider) exception = error_context.exception self.assertEqual(exception.code, 4007) self.assertEqual(exception.message, "API authorization error") self.assertEqual(exception.provider_name, "TestProvider") self.assertIsNotNone(exception.system_exception) self.assertEqual(exception.system_exception.args[0], "Testing API authorization error") self.assertEqual(len(self.mock_logger_handler.messages["error"]), 1) self.assertEqual( self.mock_logger_handler.messages["error"][0], "IngestApiError Error 4007 - API authorization error: " "Testing API authorization error on channel TestProvider", )
def _fetch_data(self, config, provider): """Fetch the latest feed data. :param dict config: RSS resource configuration :param provider: data provider instance, needed as an argument when raising ingest errors :return: fetched RSS data :rtype: str :raises IngestApiError: if fetching data fails for any reason (e.g. authentication error, resource not found, etc.) """ url = config['url'] if config.get('auth_required', False): auth = (config.get('username'), config.get('password')) self.auth_info = { 'username': config.get('username', ''), 'password': config.get('password', '') } else: auth = None try: response = requests.get(url, auth=auth, timeout=30) except requests.exceptions.ConnectionError as err: raise IngestApiError.apiConnectionError(exception=err, provider=provider) except requests.exceptions.RequestException as err: raise IngestApiError.apiURLError(exception=err, provider=provider) if response.ok: return response.content else: if response.status_code in (401, 403): raise IngestApiError.apiAuthError( Exception(response.reason), provider) elif response.status_code == 404: raise IngestApiError.apiNotFoundError( Exception(response.reason), provider) else: raise IngestApiError.apiGeneralError( Exception(response.reason), provider)
def _generate_auth_token(self, provider): """ Generates Authentication Token as per the configuration in Ingest Provider. :param provider: dict - Ingest provider details to which the current directory has been configured :type provider: dict :py:class: `superdesk.io.ingest_provider_model.IngestProviderResource` :return: token details if successfully authenticated :rtype: str :raises: IngestApiError.apiGeneralError() if auth_url is missing in the Ingest Provider configuration """ session = requests.Session() session.mount('https://', SSLAdapter()) auth_url = provider.get('config', {}).get('auth_url', None) if not auth_url: raise IngestApiError.apiGeneralError(provider=provider, exception=KeyError(''' Ingest Provider {} is missing Authentication URL. Please check the configuration. '''.format( provider['name']))) payload = { 'username': provider.get('config', {}).get('username', ''), 'password': provider.get('config', {}).get('password', ''), } response = session.get(auth_url, params=payload, verify=False, timeout=30) if response.status_code < 200 or response.status_code >= 300: raise IngestApiError.apiAuthError(provider=provider) tree = etree.fromstring( response.content) # workaround for http mock lib return tree.text
from datetime import datetime from superdesk.errors import IngestApiError, ParserError from superdesk.io import register_provider from superdesk.io.ingest_service import IngestService from superdesk.utils import merge_dicts from urllib.parse import quote as urlquote, urlsplit, urlunsplit PROVIDER = "rss" utcfromtimestamp = datetime.utcfromtimestamp errors = [ IngestApiError.apiAuthError().get_error_description(), IngestApiError.apiNotFoundError().get_error_description(), IngestApiError.apiGeneralError().get_error_description(), ParserError.parseMessageError().get_error_description(), ] class RssIngestService(IngestService): """Ingest service for providing feeds received in RSS 2.0 format. (NOTE: it should also work with other syndicated feeds formats, too, since the underlying parser supports them, but for our needs RSS 2.0 is assumed) """ ItemField = namedtuple("ItemField", ["name", "name_in_data", "type"])
class RSSFeedingService(FeedingService): """ Feeding service for providing feeds received in RSS 2.0 format. (NOTE: it should also work with other syndicated feeds formats, too, since the underlying parser supports them, but for our needs RSS 2.0 is assumed) """ NAME = 'rss' ERRORS = [IngestApiError.apiAuthError().get_error_description(), IngestApiError.apiNotFoundError().get_error_description(), IngestApiError.apiGeneralError().get_error_description(), ParserError.parseMessageError().get_error_description()] ItemField = namedtuple('ItemField', ['name', 'name_in_data', 'type']) item_fields = [ ItemField('guid', 'guid', str), ItemField('uri', 'guid', str), ItemField('firstcreated', 'published_parsed', datetime), ItemField('versioncreated', 'updated_parsed', datetime), ItemField('headline', 'title', str), ItemField('abstract', 'summary', str), ItemField('body_html', 'body_text', str), ItemField('timescalled', 'timescalled', int), ItemField('test', 'test', str), ItemField('testing', 'testing', str), ItemField('mobilecircle', 'mobilecircle', str), ItemField('audiofile', 'audiofile', str), ItemField('timesrecorded', 'timesrecorded', int), ItemField('timespublished', 'timespublished', int), ] """A list of fields that items created from the ingest data should contain. Each list item is a named tuple with the following three attribues: * name - the name of the field (attribute) in the resulting ingest item * name_in_data - the expected name of the data field in the retrieved ingest data (this can be overriden by providing a field name alias) * type - field's data type """ IMG_MIME_TYPES = ( 'image/gif', 'image/jpeg', 'image/png', 'image/tiff', ) """ Supported MIME types for ingesting external images referenced by the RSS entries. """ IMG_FILE_SUFFIXES = ('.gif', '.jpeg', '.jpg', '.png', '.tif', '.tiff') """ Supported image filename extensions for ingesting (used for the <media:thumbnail> tags - they lack the "type" attribute). """ def __init__(self): super().__init__() self.auth_info = None def prepare_href(self, url, mimetype=None): """ Prepare a link to an external resource (e.g. an image file) so that it can be directly used by the ingest machinery for fetching it. If provider requires authentication, basic HTTP authentication info is added to the given url, otherwise it is returned unmodified. :param str url: the original URL as extracted from an RSS entry :return: prepared URL :rtype: str """ if self.auth_info: userinfo_part = '{}:{}@'.format( urlquote(self.auth_info['username']), urlquote(self.auth_info['password']) ) scheme, netloc, path, query, fragment = urlsplit(url) netloc = userinfo_part + netloc url = urlunsplit((scheme, netloc, path, query, fragment)) return url def _update(self, provider): """ Check data provider for data updates and returns new items (if any). :param provider: data provider instance :return: a list containing a list of new content items :rtype: list :raises IngestApiError: if data retrieval error occurs :raises ParserError: if retrieved RSS data cannot be parsed """ config = provider.get('config', {}) if config.get('auth_required'): self.auth_info = { 'username': config.get('username', ''), 'password': config.get('password', '') } try: xml_data = self._fetch_data(config, provider) data = feedparser.parse(xml_data) except IngestApiError: raise except Exception as ex: raise ParserError.parseMessageError(ex, provider) # If provider last updated time is not available, set it to 1.1.1970 # so that it will be recognized as "not up to date". # Also convert it to a naive datetime object (removing tzinfo is fine, # because it is in UTC anyway) t_provider_updated = provider.get('last_updated', utcfromtimestamp(0)) t_provider_updated = t_provider_updated.replace(tzinfo=None) new_items = [] field_aliases = config.get('field_aliases') for entry in data.entries: t_entry_updated = utcfromtimestamp(timegm(entry.updated_parsed)) if t_entry_updated <= t_provider_updated: continue item = self._create_item(entry, field_aliases, provider.get('source', None)) self.add_timestamps(item) # If the RSS entry references any images, create picture items from # them and create a package referencing them and the entry itself. # If there are no image references, treat entry as a simple text # item, even if it might reference other media types, e.g. videos. image_urls = self._extract_image_links(entry) if image_urls: image_items = self._create_image_items(image_urls, item) new_items.extend(image_items) new_items.append(item) item = self._create_package(item, image_items) new_items.append(item) return [new_items] def _fetch_data(self, config, provider): """Fetch the latest feed data. :param dict config: RSS resource configuration :param provider: data provider instance, needed as an argument when raising ingest errors :return: fetched RSS data :rtype: str :raises IngestApiError: if fetching data fails for any reason (e.g. authentication error, resource not found, etc.) """ url = config['url'] if config.get('auth_required', False): auth = (config.get('username'), config.get('password')) else: auth = None response = requests.get(url, auth=auth) if response.ok: return response.content else: if response.status_code in (401, 403): raise IngestApiError.apiAuthError( Exception(response.reason), provider) elif response.status_code == 404: raise IngestApiError.apiNotFoundError( Exception(response.reason), provider) else: raise IngestApiError.apiGeneralError( Exception(response.reason), provider) def _extract_image_links(self, rss_entry): """Extract URLs of all images referenced by the given RSS entry. Images can be referenced via `<enclosure>`, `<media:thumbnail>` or `<media:content>` RSS tag and must be listed among the allowed image types. All other links to external media are ignored. Duplicate URLs are omitted from the result. :param rss_entry: parsed RSS item (entry) :type rss_entry: :py:class:`feedparser.FeedParserDict` :return: a list of all unique image URLs found (as strings) """ img_links = set() for link in getattr(rss_entry, 'links', []): if link.get('type') in self.IMG_MIME_TYPES: img_links.add(link['href']) for item in getattr(rss_entry, 'media_thumbnail', []): url = item.get('url', '') if url.endswith(self.IMG_FILE_SUFFIXES): img_links.add(url) for item in getattr(rss_entry, 'media_content', []): if item.get('type') in self.IMG_MIME_TYPES: img_links.add(item['url']) return list(img_links) def _create_item(self, data, field_aliases=None, source=None): """Create a new content item from RSS feed data. :param dict data: parsed data of a single feed entry :param field_aliases: (optional) field name aliases. Used for content fields that are named differently in retrieved data. :type field_aliases: list of {field_name: alias} dictionaries or None :param str source: the source of provider :return: created content item :rtype: dict """ if field_aliases is None: field_aliases = {} else: field_aliases = merge_dicts(field_aliases) aliased_fields = set(field_aliases.values()) item = dict(type=CONTENT_TYPE.TEXT) # Only consider fields that are not used as an alias (i.e. used to # populate another field) - unless those fields have their own # aliases, too. # The idea is that if e.g. the main text field is aliased to use the # parsed data's summary field, that summary should not be used to # populate the field it was originally meant for. fields_to_consider = ( f for f in self.item_fields if (f.name_in_data not in aliased_fields) or (f.name_in_data in aliased_fields and f.name_in_data in field_aliases) ) for field in fields_to_consider: data_field_name = field_aliases.get( field.name_in_data, field.name_in_data ) field_value = data.get(data_field_name) if (field.type is datetime) and field_value: field_value = utcfromtimestamp(timegm(field_value)) item[field.name] = field_value # Some feeds use <content:encoded> tag for storing the main content, # and that tag is parsed differently. If the body_html has not been # found in its default data field and is not aliased, try to # populate it using the aforementioned content field as a fallback. if ( field.name == 'body_html' and not field_value and field.name_in_data not in field_aliases ): try: item['body_html'] = data.content[0].value except: pass # content either non-existent or parsed differently if item.get('uri', None): if not item.get('body_html', None): item['body_html'] = '' source = source or 'source' item['body_html'] = '<p><a href="%s" target="_blank">%s</a></p>' % (item['uri'], source) + item['body_html'] return item def _create_image_items(self, image_links, text_item): """Create a list of picture items that represent the external images located on given URLs. Each created item's `firstcreated` and `versioncreated` fields are set to the same value as the values of these fields in `text_item`. :param iterable image_links: list of image URLs :param dict text_item: the "main" text item the images are related to :return: list of created image items (as dicts) """ image_items = [] for image_url in image_links: img_item = { 'guid': generate_guid(type=GUID_TAG), ITEM_TYPE: CONTENT_TYPE.PICTURE, 'firstcreated': text_item.get('firstcreated'), 'versioncreated': text_item.get('versioncreated'), 'renditions': { 'baseImage': { 'href': image_url } } } image_items.append(img_item) return image_items def _create_package(self, text_item, image_items): """Create a new content package from given content items. The package's `main` group contains only the references to given items, not the items themselves. In the list of references, the reference to the text item preceeds the references to image items. Package's `firstcreated` and `versioncreated` fields are set to values of these fields in `text_item`, and the `headline` is copied as well. :param dict text_item: item representing the text content :param list image_items: list of items (dicts) representing the images related to the text content :return: the created content package :rtype: dict """ package = { ITEM_TYPE: CONTENT_TYPE.COMPOSITE, 'guid': generate_guid(type=GUID_TAG), 'firstcreated': text_item['firstcreated'], 'versioncreated': text_item['versioncreated'], 'headline': text_item.get('headline', ''), 'groups': [ { 'id': 'root', 'role': 'grpRole:NEP', 'refs': [{'idRef': 'main'}], }, { 'id': 'main', 'role': 'main', 'refs': [], } ] } item_references = package['groups'][1]['refs'] item_references.append({'residRef': text_item['guid']}) for image in image_items: item_references.append({'residRef': image['guid']}) return package
from calendar import timegm from collections import namedtuple from datetime import datetime from superdesk.errors import IngestApiError, ParserError from superdesk.io import register_provider from superdesk.io.ingest_service import IngestService from superdesk.utils import merge_dicts PROVIDER = 'rss' utcfromtimestamp = datetime.utcfromtimestamp errors = [IngestApiError.apiAuthError().get_error_description(), IngestApiError.apiNotFoundError().get_error_description(), IngestApiError.apiGeneralError().get_error_description(), ParserError.parseMessageError().get_error_description()] class RssIngestService(IngestService): """Ingest service for providing feeds received in RSS 2.0 format. (NOTE: it should also work with other syndicated feeds formats, too, since the underlying parser supports them, but for our needs RSS 2.0 is assumed) """ ItemField = namedtuple('ItemField', ['name', 'name_in_data', 'type']) item_fields = [
def get_url(self, url=None, **kwargs): """Do an HTTP Get on URL :param string url: url to use (None to use self.HTTP_URL) :param **kwargs: extra parameter for requests :return requests.Response: response """ if not url: url = self.HTTP_URL config = self.config user = config.get('username') password = config.get('password') if user: user = user.strip() if password: password = password.strip() auth_required = config.get('auth_required', self.HTTP_AUTH) if auth_required is None: # auth_required may not be user in the feeding service # in this case with use authentification only if user # and password are set. auth_required = bool(user and password) if auth_required: if not user: raise SuperdeskIngestError.notConfiguredError( "user is not configured") if not password: raise SuperdeskIngestError.notConfiguredError( "password is not configured") kwargs.setdefault('auth', (user, password)) params = kwargs.pop("params", {}) if params or self.HTTP_DEFAULT_PARAMETERS: # if we have default parameters, we want them to be overriden # by conflicting params given in arguments if self.HTTP_DEFAULT_PARAMETERS: params.update(self.HTTP_DEFAULT_PARAMETERS) kwargs["params"] = params try: response = requests.get(url, timeout=self.HTTP_TIMEOUT, **kwargs) except requests.exceptions.Timeout as exception: raise IngestApiError.apiTimeoutError(exception, self.provider) except requests.exceptions.ConnectionError as exception: raise IngestApiError.apiConnectionError(exception, self.provider) except requests.exceptions.RequestException as exception: raise IngestApiError.apiRequestError(exception, self.provider) except Exception as exception: traceback.print_exc() raise IngestApiError.apiGeneralError(exception, self.provider) if not response.ok: exception = Exception(response.reason) if response.status_code in (401, 403): raise IngestApiError.apiAuthError(exception, self.provider) elif response.status_code == 404: raise IngestApiError.apiNotFoundError(exception, self.provider) else: raise IngestApiError.apiGeneralError(exception, self.provider) return response
class BBCLDRSFeedingService(FeedingService): """ Feeding Service class for reading BBC's Local Democracy Reporting Service """ # Following the api spec at https://docs.ldrs.org.uk/ NAME = 'bbc_ldrs' ERRORS = [ IngestApiError.apiAuthError().get_error_description(), IngestApiError.apiNotFoundError().get_error_description(), IngestApiError.apiGeneralError().get_error_description(), ParserError.parseMessageError().get_error_description() ] label = 'BBC Local Democracy Reporter Service' fields = [{ 'id': 'url', 'type': 'text', 'label': 'LDRS URL', 'placeholder': 'LDRS URL', 'required': True, 'default': 'https://api.ldrs.org.uk/v1/item' }, { 'id': 'api_key', 'type': 'text', 'label': 'API Key', 'placeholder': 'API Key', 'required': True, 'default': '' }] def __init__(self): super().__init__() def _test(self, provider): config = provider.get('config', {}) url = config['url'] api_key = config['api_key'] # limit the data to a single article and filter out all article fields # to save bandwidth params = {'limit': 1, 'fields': 'id'} headers = {'apikey': api_key} try: response = requests.get(url, params=params, headers=headers, timeout=30) except requests.exceptions.ConnectionError as err: raise IngestApiError.apiConnectionError(exception=err) if not response.ok: if response.status_code == 404: raise IngestApiError.apiNotFoundError( Exception(response.reason), provider) else: raise IngestApiError.apiGeneralError( Exception(response.reason), provider) def _update(self, provider, update): config = provider.get('config', {}) json_items = self._fetch_data(config, provider) parsed_items = [] for item in json_items: try: parser = self.get_feed_parser(provider, item) parsed_items.append(parser.parse(item)) except Exception as ex: raise ParserError.parseMessageError(ex, provider, data=item) return parsed_items def _fetch_data(self, config, provider): url = config['url'] api_key = config['api_key'] last_update = provider.get( 'last_updated', utcfromtimestamp(0)).strftime('%Y-%m-%dT%H:%M:%S') # Results are pagified so we'll read this many at a time offset_jump = 10 params = {'start': last_update, 'limit': offset_jump} headers = {'apikey': api_key} items = [] offset = 0 while True: params['offset'] = offset try: response = requests.get(url, params=params, headers=headers, timeout=30) except requests.exceptions.ConnectionError as err: raise IngestApiError.apiConnectionError(exception=err) if response.ok: # The total number of results are given to us in json, get them # via a regex to read the field so we don't have to convert the # whole thing to json pointlessly item_ident = re.search('\"total\": *[0-9]*', response.text).group() results_str = re.search('[0-9]+', item_ident).group() if results_str is None: raise IngestApiError.apiGeneralError( Exception(response.text), provider) num_results = int(results_str) if num_results > 0: items.append(response.text) if offset >= num_results: return items offset += offset_jump else: if re.match('Error: No API Key provided', response.text): raise IngestApiError.apiAuthError(Exception(response.text), provider) elif response.status_code == 404: raise IngestApiError.apiNotFoundError( Exception(response.reason), provider) else: raise IngestApiError.apiGeneralError( Exception(response.reason), provider) return items
class RSSFeedingService(HTTPFeedingServiceBase): """ Feeding service for providing feeds received in RSS 2.0 format. (NOTE: it should also work with other syndicated feeds formats, too, since the underlying parser supports them, but for our needs RSS 2.0 is assumed) """ NAME = "rss" ERRORS = [ IngestApiError.apiAuthError().get_error_description(), IngestApiError.apiNotFoundError().get_error_description(), IngestApiError.apiGeneralError().get_error_description(), ParserError.parseMessageError().get_error_description(), ] label = "RSS/Atom" fields = ([{ "id": "url", "type": "text", "label": "Host", "placeholder": "RSS Feed URL", "required": True, "errors": { 4001: "Connection timed out.", 4006: "URL not found.", 4009: "Can't connect to host.", 1001: "Can't parse the RSS.", }, }] + HTTPFeedingServiceBase.AUTH_REQ_FIELDS + [{ "id": "field_aliases", "type": "mapping", "label": "Content Field Aliases", "add_mapping_label": "Add alias", "remove_mapping_label": "Remove", "empty_label": "No field aliases defined.", "first_field_options": { "label": "Content Field Name", "values": [ "body_text", "guid", "published_parsed", "summary", "title", "updated_parsed" ], }, "second_field_options": { "label": "Field Alias", "placeholder": "Enter field alias" }, }]) HTTP_AUTH = None field_groups = { "auth_data": { "label": "Authentication Info", "fields": ["username", "password"] } } ItemField = namedtuple("ItemField", ["name", "name_in_data", "type"]) item_fields = [ ItemField("guid", "guid", str), ItemField("uri", "guid", str), ItemField("firstcreated", "published_parsed", datetime), ItemField("versioncreated", "updated_parsed", datetime), ItemField("headline", "title", str), ItemField("abstract", "summary", str), ItemField("body_html", "body_text", str), ItemField("byline", "author", str), ] """A list of fields that items created from the ingest data should contain. Each list item is a named tuple with the following three attribues: * name - the name of the field (attribute) in the resulting ingest item * name_in_data - the expected name of the data field in the retrieved ingest data (this can be overriden by providing a field name alias) * type - field's data type """ IMG_MIME_TYPES = ( "image/gif", "image/jpeg", "image/png", "image/tiff", ) """ Supported MIME types for ingesting external images referenced by the RSS entries. """ IMG_FILE_SUFFIXES = (".gif", ".jpeg", ".jpg", ".png", ".tif", ".tiff") """ Supported image filename extensions for ingesting (used for the <media:thumbnail> tags - they lack the "type" attribute). """ def prepare_href(self, url, mimetype=None): """Prepare a link to an external resource (e.g. an image file). It can be directly used by the ingest machinery for fetching it. If provider requires authentication, basic HTTP authentication info is added to the given url, otherwise it is returned unmodified. :param str url: the original URL as extracted from an RSS entry :return: prepared URL :rtype: str """ if self.auth_info: userinfo_part = "{}:{}@".format( urlquote(self.auth_info["username"]), urlquote(self.auth_info["password"])) scheme, netloc, path, query, fragment = urlsplit(url) netloc = userinfo_part + netloc url = urlunsplit((scheme, netloc, path, query, fragment)) return url def _test(self, provider): """Test connection.""" self.provider = provider xml = self._fetch_data() data = feedparser.parse(xml) if data.bozo: raise ParserError.parseMessageError(data.bozo_exception, provider) def _update(self, provider, update): """ Check data provider for data updates and returns new items (if any). :param provider: data provider instance :return: a list containing a list of new content items :rtype: list :raises IngestApiError: if data retrieval error occurs :raises ParserError: if retrieved RSS data cannot be parsed """ xml_data = self._fetch_data() try: data = feedparser.parse(xml_data) except Exception as ex: raise ParserError.parseMessageError(ex, provider, data=xml_data) # If provider last updated time is not available, set it to 1.1.1970 # so that it will be recognized as "not up to date". # Also convert it to a naive datetime object (removing tzinfo is fine, # because it is in UTC anyway) t_provider_updated = provider.get(LAST_ITEM_UPDATE, utcfromtimestamp(0)) t_provider_updated = t_provider_updated.replace(tzinfo=None) new_items = [] field_aliases = self.config.get("field_aliases") for entry in data.entries: try: t_entry_updated = utcfromtimestamp(timegm( entry.updated_parsed)) if t_entry_updated <= t_provider_updated: continue except (AttributeError, TypeError): # missing updated info, so better ingest it pass item = self._create_item(entry, field_aliases, provider.get("source", None)) self.localize_timestamps(item) # If the RSS entry references any images, create picture items from # them and create a package referencing them and the entry itself. # If there are no image references, treat entry as a simple text # item, even if it might reference other media types, e.g. videos. image_urls = self._extract_image_links(entry) if image_urls: image_items = self._create_image_items(image_urls, item) new_items.extend(image_items) new_items.append(item) item = self._create_package(item, image_items) new_items.append(item) return [new_items] def _fetch_data(self): """Fetch the latest feed data. :return: fetched RSS data :rtype: str :raises IngestApiError: if fetching data fails for any reason (e.g. authentication error, resource not found, etc.) """ url = self.config["url"] response = self.get_url(url) return response.content def _extract_image_links(self, rss_entry): """Extract URLs of all images referenced by the given RSS entry. Images can be referenced via `<enclosure>`, `<media:thumbnail>` or `<media:content>` RSS tag and must be listed among the allowed image types. All other links to external media are ignored. Duplicate URLs are omitted from the result. :param rss_entry: parsed RSS item (entry) :type rss_entry: :py:class:`feedparser.FeedParserDict` :return: a list of all unique image URLs found (as strings) """ img_links = set() for link in getattr(rss_entry, "links", []): if link.get("type") in self.IMG_MIME_TYPES: img_links.add(link["href"]) for item in getattr(rss_entry, "media_thumbnail", []): url = item.get("url", "") if url.endswith(self.IMG_FILE_SUFFIXES): img_links.add(url) for item in getattr(rss_entry, "media_content", []): if item.get("type") in self.IMG_MIME_TYPES: img_links.add(item["url"]) return list(img_links) def _create_item(self, data, field_aliases=None, source="source"): """Create a new content item from RSS feed data. :param dict data: parsed data of a single feed entry :param field_aliases: (optional) field name aliases. Used for content fields that are named differently in retrieved data. :type field_aliases: list of {field_name: alias} dictionaries or None :param str source: the source of provider :return: created content item :rtype: dict """ if field_aliases is None: field_aliases = {} else: field_aliases = merge_dicts(field_aliases) aliased_fields = set(field_aliases.values()) item = dict(type=CONTENT_TYPE.TEXT) # Only consider fields that are not used as an alias (i.e. used to # populate another field) - unless those fields have their own # aliases, too. # The idea is that if e.g. the main text field is aliased to use the # parsed data's summary field, that summary should not be used to # populate the field it was originally meant for. fields_to_consider = (f for f in self.item_fields if (f.name_in_data not in aliased_fields) or ( f.name_in_data in aliased_fields and f.name_in_data in field_aliases)) utc_now = datetime.utcnow() for field in fields_to_consider: data_field_name = field_aliases.get(field.name_in_data, field.name_in_data) field_value = data.get(data_field_name) if (field.type is datetime) and field_value: field_value = utcfromtimestamp(timegm(field_value)) field_value = utc_now if field_value > utc_now else field_value item[field.name] = field_value # Some feeds use <content:encoded> tag for storing the main content, # and that tag is parsed differently. If the body_html has not been # found in its default data field and is not aliased, try to # populate it using the aforementioned content field as a fallback. if field.name == "body_html" and not field_value and field.name_in_data not in field_aliases: try: item["body_html"] = data.content[0].value except Exception: pass # content either non-existent or parsed differently if not data.get("guidislink") and data.get("link"): item["uri"] = data["link"] scheme, netloc, path, query, fragment = urlsplit(item["uri"]) if data.get("guid"): item["guid"] = generate_tag(domain=netloc, id=data.get("guid")) else: item["guid"] = generate_tag_from_url(data["link"]) if item.get("uri", None): if not item.get("body_html", None): item["body_html"] = "" item[ "body_html"] = '<p><a href="%s" target="_blank">%s</a></p>' % ( item["uri"], source) + item["body_html"] item["dateline"] = { "source": source, "date": item.get("firstcreated", item.get("versioncreated")) } if not item.get("versioncreated") and item.get("firstcreated"): item["versioncreated"] = item["firstcreated"] return item def _create_image_items(self, image_links, text_item): """Create a list of picture items that represent the external images located on given URLs. Each created item's `firstcreated` and `versioncreated` fields are set to the same value as the values of these fields in `text_item`. :param iterable image_links: list of image URLs :param dict text_item: the "main" text item the images are related to :return: list of created image items (as dicts) """ image_items = [] for image_url in image_links: img_item = { "guid": generate_tag_from_url(image_url), ITEM_TYPE: CONTENT_TYPE.PICTURE, "firstcreated": text_item.get("firstcreated"), "versioncreated": text_item.get("versioncreated"), "renditions": { "baseImage": { "href": image_url } }, } image_items.append(img_item) return image_items def _create_package(self, text_item, image_items): """Create a new content package from given content items. The package's `main` group contains only the references to given items, not the items themselves. In the list of references, the reference to the text item preceeds the references to image items. Package's `firstcreated` and `versioncreated` fields are set to values of these fields in `text_item`, and the `headline` is copied as well. :param dict text_item: item representing the text content :param list image_items: list of items (dicts) representing the images related to the text content :return: the created content package :rtype: dict """ package = { ITEM_TYPE: CONTENT_TYPE.COMPOSITE, "guid": "{}:pkg".format(text_item["guid"]), "firstcreated": text_item["firstcreated"], "versioncreated": text_item["versioncreated"], "headline": text_item.get("headline", ""), "groups": [ { "id": "root", "role": "grpRole:NEP", "refs": [{ "idRef": "main" }], }, { "id": "main", "role": "main", "refs": [], }, ], } item_references = package["groups"][1]["refs"] item_references.append({"residRef": text_item["guid"]}) for image in image_items: item_references.append({"residRef": image["guid"]}) return package