diff --git a/swh/lister/bitbucket/lister.py b/swh/lister/bitbucket/lister.py index 5378983..e067148 100644 --- a/swh/lister/bitbucket/lister.py +++ b/swh/lister/bitbucket/lister.py @@ -1,87 +1,87 @@ # Copyright (C) 2017-2019 The Software Heritage developers # See the AUTHORS file at the top-level directory of this distribution # License: GNU General Public License version 3, or any later version # See top-level LICENSE file for more information import logging import iso8601 from datetime import datetime, timezone -from typing import Any, Dict, List, Optional, Union +from typing import Any, Dict, List, Optional from urllib import parse from requests import Response from swh.lister.bitbucket.models import BitBucketModel from swh.lister.core.indexing_lister import IndexingHttpLister logger = logging.getLogger(__name__) class BitBucketLister(IndexingHttpLister): PATH_TEMPLATE = '/repositories?after=%s' MODEL = BitBucketModel LISTER_NAME = 'bitbucket' DEFAULT_URL = 'https://api.bitbucket.org/2.0' instance = 'bitbucket' default_min_bound = datetime.fromtimestamp(0, timezone.utc) # type: Any def __init__(self, url: str = None, override_config=None, per_page: int = 100) -> None: super().__init__(url=url, override_config=override_config) per_page = self.config.get('per_page', per_page) self.PATH_TEMPLATE = '%s&pagelen=%s' % ( self.PATH_TEMPLATE, per_page) def get_model_from_repo(self, repo: Dict) -> Dict[str, Any]: return { 'uid': repo['uuid'], 'indexable': iso8601.parse_date(repo['created_on']), 'name': repo['name'], 'full_name': repo['full_name'], 'html_url': repo['links']['html']['href'], 'origin_url': repo['links']['clone'][0]['href'], 'origin_type': repo['scm'], } def get_next_target_from_response(self, response: Response - ) -> Union[None, datetime]: + ) -> Optional[datetime]: """This will read the 'next' link from the api response if any and return it as a datetime. Args: response (Response): requests' response from api call Returns: next date as a datetime """ body = response.json() next_ = body.get('next') if next_ is not None: next_ = parse.urlparse(next_) return iso8601.parse_date(parse.parse_qs(next_.query)['after'][0]) return None def transport_response_simplified(self, response: Response ) -> List[Dict[str, Any]]: repos = response.json()['values'] return [self.get_model_from_repo(repo) for repo in repos] - def request_uri(self, identifier: datetime) -> str: + def request_uri(self, identifier: datetime) -> str: # type: ignore identifier_str = parse.quote(identifier.isoformat()) return super().request_uri(identifier_str or '1970-01-01') def is_within_bounds(self, inner: int, lower: Optional[int] = None, upper: Optional[int] = None) -> bool: # values are expected to be datetimes if lower is None and upper is None: ret = True elif lower is None: ret = inner <= upper # type: ignore elif upper is None: ret = inner >= lower else: ret = lower <= inner <= upper return ret diff --git a/swh/lister/cgit/lister.py b/swh/lister/cgit/lister.py index 1f5545c..bf37ea3 100644 --- a/swh/lister/cgit/lister.py +++ b/swh/lister/cgit/lister.py @@ -1,150 +1,150 @@ # Copyright (C) 2019 the Software Heritage developers # License: GNU General Public License version 3, or any later version # See top-level LICENSE file for more information import re import logging from urllib.parse import urlparse, urljoin from bs4 import BeautifulSoup from requests import Session -# from requests.structures import CaseInsensitiveDict + from requests.adapters import HTTPAdapter -from typing import Any, Dict, Generator, Union +from typing import Any, Dict, Generator, Optional from .models import CGitModel from swh.core.utils import grouper from swh.lister import USER_AGENT from swh.lister.core.lister_base import ListerBase logger = logging.getLogger(__name__) class CGitLister(ListerBase): """Lister class for CGit repositories. This lister will retrieve the list of published git repositories by parsing the HTML page(s) of the index retrieved at `url`. For each found git repository, a query is made at the given url found in this index to gather published "Clone" URLs to be used as origin URL for that git repo. If several "Clone" urls are provided, prefer the http/https one, if any, otherwise fall bak to the first one. A loader task is created for each git repository:: Task: Type: load-git Policy: recurring Args: Example:: Task: Type: load-git Policy: recurring Args: 'https://git.savannah.gnu.org/git/elisp-es.git' """ MODEL = CGitModel DEFAULT_URL = 'https://git.savannah.gnu.org/cgit/' LISTER_NAME = 'cgit' url_prefix_present = True def __init__(self, url=None, instance=None, override_config=None): """Lister class for CGit repositories. Args: - url : main URL of the CGit instance, i.e. url of the index + url (str): main URL of the CGit instance, i.e. url of the index of published git repositories on this instance. - instance : Name of cgit instance. Defaults to url's hostname + instance (str): Name of cgit instance. Defaults to url's hostname if unset. """ super().__init__(override_config=override_config) if url is None: url = self.config.get('url', self.DEFAULT_URL) self.url = url if not instance: instance = urlparse(url).hostname self.instance = instance self.session = Session() self.session.mount(self.url, HTTPAdapter(max_retries=3)) self.session.headers = { 'User-Agent': USER_AGENT, } def run(self) -> Dict[str, str]: status = 'uneventful' total = 0 for repos in grouper(self.get_repos(), 10): models = list(filter(None, (self.build_model(repo) for repo in repos))) injected_repos = self.inject_repo_data_into_db(models) self.schedule_missing_tasks(models, injected_repos) self.db_session.commit() total += len(injected_repos) logger.debug('Scheduled %s tasks for %s', total, self.url) status = 'eventful' return {'status': status} - def get_repos(self) -> Generator: + def get_repos(self) -> Generator[str, None, None]: """Generate git 'project' URLs found on the current CGit server """ next_page = self.url while next_page: bs_idx = self.get_and_parse(next_page) for tr in bs_idx.find( 'div', {"class": "content"}).find_all( "tr", {"class": ""}): yield urljoin(self.url, tr.find('a')['href']) try: pager = bs_idx.find('ul', {'class': 'pager'}) current_page = pager.find('a', {'class': 'current'}) if current_page: next_page = current_page.parent.next_sibling.a['href'] next_page = urljoin(self.url, next_page) except (AttributeError, KeyError): # no pager, or no next page next_page = None - def build_model(self, repo_url: str) -> Union[None, Dict[str, Any]]: + def build_model(self, repo_url: str) -> Optional[Dict[str, Any]]: """Given the URL of a git repo project page on a CGit server, return the repo description (dict) suitable for insertion in the db. """ bs = self.get_and_parse(repo_url) urls = [x['href'] for x in bs.find_all('a', {'rel': 'vcs-git'})] if not urls: return None # look for the http/https url, if any, and use it as origin_url for url in urls: if urlparse(url).scheme in ('http', 'https'): origin_url = url break else: # otherwise, choose the first one origin_url = urls[0] return {'uid': repo_url, 'name': bs.find('a', title=re.compile('.+'))['title'], 'origin_type': 'git', 'instance': self.instance, 'origin_url': origin_url, } def get_and_parse(self, url: str) -> BeautifulSoup: "Get the given url and parse the retrieved HTML using BeautifulSoup" return BeautifulSoup(self.session.get(url).text, features='html.parser') diff --git a/swh/lister/core/indexing_lister.py b/swh/lister/core/indexing_lister.py index 8611e7f..d13933d 100644 --- a/swh/lister/core/indexing_lister.py +++ b/swh/lister/core/indexing_lister.py @@ -1,269 +1,271 @@ # Copyright (C) 2015-2019 the Software Heritage developers # License: GNU General Public License version 3, or any later version # See top-level LICENSE file for more information import abc import logging from itertools import count import dateutil from sqlalchemy import func from .lister_transports import ListerHttpTransport from .lister_base import ListerBase from requests import Response -from typing import Any, Dict, List, Tuple, Optional +from typing import Any, Dict, List, Tuple, Optional, Union +from datetime import datetime logger = logging.getLogger(__name__) class IndexingLister(ListerBase): """Lister* intermediate class for any service that follows the pattern: - The service must report at least one stable unique identifier, known herein as the UID value, for every listed repository. - If the service splits the list of repositories into sublists, it must report at least one stable and sorted index identifier for every listed repository, known herein as the indexable value, which can be used as part of the service endpoint query to request a sublist beginning from that index. This might be the UID if the UID is monotonic. - Client sends a request to list repositories starting from a given index. - Client receives structured (json/xml/etc) response with information about a sequential series of repositories starting from that index and, if necessary/available, some indication of the URL or index for fetching the next series of repository data. See :class:`swh.lister.core.lister_base.ListerBase` for more details. This class cannot be instantiated. To create a new Lister for a source code listing service that follows the model described above, you must subclass this class and provide the required overrides in addition to any unmet implementation/override requirements of this class's base. (see parent class and member docstrings for details) Required Overrides:: def get_next_target_from_response """ flush_packet_db = 20 """Number of iterations in-between write flushes of lister repositories to db (see fn:`run`). """ default_min_bound = '' """Default initialization value for the minimum boundary index to use when undefined (see fn:`run`). """ @abc.abstractmethod - def get_next_target_from_response(self, response: Response): + def get_next_target_from_response( + self, response: Response + ) -> Union[Optional[datetime], Optional[str], Optional[int]]: """Find the next server endpoint identifier given the entire response. Implementation of this method depends on the server API spec and the shape of the network response object returned by the transport_request method. Args: response (transport response): response page from the server Returns: index of next page, possibly extracted from a next href url """ pass # You probably don't need to override anything below this line. def filter_before_inject( self, models_list: List[Dict[str, Any]]) -> List[Dict[str, Any]]: """Overrides ListerBase.filter_before_inject Bounds query results by this Lister's set max_index. """ models_list = [ m for m in models_list if self.is_within_bounds(m['indexable'], None, self.max_index) ] return models_list def db_query_range(self, start, end): """Look in the db for a range of repositories with indexable values in the range [start, end] Args: start (model indexable type): start of desired indexable range end (model indexable type): end of desired indexable range Returns: a list of sqlalchemy.ext.declarative.declarative_base objects with indexable values within the given range """ retlist = self.db_session.query(self.MODEL) if start is not None: retlist = retlist.filter(self.MODEL.indexable >= start) if end is not None: retlist = retlist.filter(self.MODEL.indexable <= end) return retlist def db_partition_indices( self, partition_size: int ) -> List[Tuple[Optional[int], Optional[int]]]: """Describe an index-space compartmentalization of the db table in equal sized chunks. This is used to describe min&max bounds for parallelizing fetch tasks. Args: partition_size (int): desired size to make each partition Returns: a list of tuples (begin, end) of indexable value that declare approximately equal-sized ranges of existing repos """ n = max(self.db_num_entries(), 10) partition_size = min(partition_size, n) n_partitions = n // partition_size min_index = self.db_first_index() max_index = self.db_last_index() if min_index is None or max_index is None: # Nothing to list return [] if isinstance(min_index, str): def format_bound(bound): return bound.isoformat() min_index = dateutil.parser.parse(min_index) max_index = dateutil.parser.parse(max_index) elif isinstance(max_index - min_index, int): def format_bound(bound): return int(bound) else: def format_bound(bound): return bound partition_width = (max_index - min_index) / n_partitions # Generate n_partitions + 1 bounds for n_partitions partitons bounds = [ format_bound(min_index + i * partition_width) for i in range(n_partitions + 1) ] # Trim duplicate bounds bounds.append(None) bounds = [cur for cur, next in zip(bounds[:-1], bounds[1:]) if cur != next] # Remove bounds for lowest and highest partition bounds[0] = bounds[-1] = None return list(zip(bounds[:-1], bounds[1:])) def db_first_index(self): """Look in the db for the smallest indexable value Returns: the smallest indexable value of all repos in the db """ t = self.db_session.query(func.min(self.MODEL.indexable)).first() if t: return t[0] return None def db_last_index(self): """Look in the db for the largest indexable value Returns: the largest indexable value of all repos in the db """ t = self.db_session.query(func.max(self.MODEL.indexable)).first() if t: return t[0] return None - def disable_deleted_repo_tasks( - self, start, end, keep_these): + def disable_deleted_repo_tasks(self, start, end, keep_these): """Disable tasks for repos that no longer exist between start and end. Args: start: beginning of range to disable end: end of range to disable keep_these (uid list): do not disable repos with uids in this list """ if end is None: end = self.db_last_index() if not self.is_within_bounds(end, None, self.max_index): end = self.max_index deleted_repos = self.winnow_models( self.db_query_range(start, end), self.MODEL.uid, keep_these ) tasks_to_disable = [repo.task_id for repo in deleted_repos if repo.task_id is not None] if tasks_to_disable: self.scheduler.disable_tasks(tasks_to_disable) for repo in deleted_repos: repo.task_id = None def run(self, min_bound=None, max_bound=None): """Main entry function. Sequentially fetches repository data from the service according to the basic outline in the class docstring, continually fetching sublists until either there is no next index reference given or the given next index is greater than the desired max_bound. Args: min_bound (indexable type): optional index to start from max_bound (indexable type): optional index to stop at Returns: nothing """ status = 'uneventful' self.min_index = min_bound self.max_index = max_bound def ingest_indexes(): index = min_bound or self.default_min_bound for i in count(1): response, injected_repos = self.ingest_data(index) if not response and not injected_repos: logger.info('No response from api server, stopping') return next_index = self.get_next_target_from_response(response) # Determine if any repos were deleted, and disable their tasks. keep_these = list(injected_repos.keys()) self.disable_deleted_repo_tasks(index, next_index, keep_these) # termination condition if next_index is None or next_index == index: logger.info('stopping after index %s, no next link found', index) return index = next_index logger.debug('Index: %s', index) yield i for i in ingest_indexes(): if (i % self.flush_packet_db) == 0: logger.debug('Flushing updates at index %s', i) self.db_session.commit() self.db_session = self.mk_session() status = 'eventful' self.db_session.commit() self.db_session = self.mk_session() return {'status': status} class IndexingHttpLister(ListerHttpTransport, IndexingLister): """Convenience class for ensuring right lookup and init order when combining IndexingLister and ListerHttpTransport.""" def __init__(self, url=None, override_config=None): IndexingLister.__init__(self, override_config=override_config) ListerHttpTransport.__init__(self, url=url) diff --git a/swh/lister/core/lister_transports.py b/swh/lister/core/lister_transports.py index 8c7ceb3..f4d6920 100644 --- a/swh/lister/core/lister_transports.py +++ b/swh/lister/core/lister_transports.py @@ -1,235 +1,235 @@ # Copyright (C) 2017-2018 the Software Heritage developers # License: GNU General Public License version 3, or any later version # See top-level LICENSE file for more information import abc import random from datetime import datetime from email.utils import parsedate from pprint import pformat import logging import requests import xmltodict -from typing import Optional, Union, Dict, Any +from typing import Optional, Union, Dict, Any, List from requests import Response from swh.lister import USER_AGENT_TEMPLATE, __version__ from .abstractattribute import AbstractAttribute from .lister_base import FetchError logger = logging.getLogger(__name__) class ListerHttpTransport(abc.ABC): """Use the Requests library for making Lister endpoint requests. To be used in conjunction with ListerBase or a subclass of it. """ DEFAULT_URL = None # type: Optional[str] PATH_TEMPLATE = \ AbstractAttribute( 'string containing a python string format pattern that produces' ' the API endpoint path for listing stored repositories when given' ' an index, e.g., "/repositories?after=%s". To be implemented in' ' the API-specific class inheriting this.' ) # type: Union[AbstractAttribute, Optional[str]] EXPECTED_STATUS_CODES = (200, 429, 403, 404) def request_headers(self) -> Dict[str, Any]: """Returns dictionary of any request headers needed by the server. MAY BE OVERRIDDEN if request headers are needed. """ return { 'User-Agent': USER_AGENT_TEMPLATE % self.lister_version } - def request_instance_credentials(self): + def request_instance_credentials(self) -> List[Dict[str, Any]]: """Returns dictionary of any credentials configuration needed by the forge instance to list. The 'credentials' configuration is expected to be a dict of multiple levels. The first level is the lister's name, the second is the lister's instance name, which value is expected to be a list of credential structures (typically a couple username/password). For example:: credentials: github: # github lister github: # has only one instance (so far) - username: some password: somekey - username: one password: onekey - ... gitlab: # gitlab lister riseup: # has many instances - username: someone password: ... - ... gitlab: - username: someone password: ... - ... Returns: list of credential dicts for the current lister. """ - all_creds = self.config.get('credentials') + all_creds = self.config.get('credentials') # type: ignore if not all_creds: return [] - lister_creds = all_creds.get(self.LISTER_NAME, {}) - creds = lister_creds.get(self.instance, []) + lister_creds = all_creds.get(self.LISTER_NAME, {}) # type: ignore + creds = lister_creds.get(self.instance, []) # type: ignore return creds - def request_uri(self, identifier): + def request_uri(self, identifier: str) -> str: """Get the full request URI given the transport_request identifier. MAY BE OVERRIDDEN if something more complex than the PATH_TEMPLATE is required. """ - path = self.PATH_TEMPLATE % identifier + path = self.PATH_TEMPLATE % identifier # type: ignore return self.url + path - def request_params(self, identifier: int) -> Dict[str, Any]: + def request_params(self, identifier: str) -> Dict[str, Any]: """Get the full parameters passed to requests given the transport_request identifier. This uses credentials if any are provided (see request_instance_credentials). MAY BE OVERRIDDEN if something more complex than the request headers is needed. """ params = {} params['headers'] = self.request_headers() or {} creds = self.request_instance_credentials() if not creds: return params auth = random.choice(creds) if creds else None if auth: params['auth'] = (auth['username'], # type: ignore auth['password']) return params def transport_quota_check(self, response): """Implements ListerBase.transport_quota_check with standard 429 code check for HTTP with Requests library. MAY BE OVERRIDDEN if the server notifies about rate limits in a non-standard way that doesn't use HTTP 429 and the Retry-After response header. ( https://tools.ietf.org/html/rfc6585#section-4 ) """ if response.status_code == 429: # HTTP too many requests retry_after = response.headers.get('Retry-After', self.back_off()) try: # might be seconds return True, float(retry_after) except Exception: # might be http-date at_date = datetime(*parsedate(retry_after)[:6]) from_now = (at_date - datetime.today()).total_seconds() + 5 return True, max(0, from_now) else: # response ok self.reset_backoff() return False, 0 def __init__(self, url=None): if not url: url = self.config.get('url') if not url: url = self.DEFAULT_URL if not url: raise NameError('HTTP Lister Transport requires an url.') self.url = url # eg. 'https://api.github.com' self.session = requests.Session() self.lister_version = __version__ def _transport_action( - self, identifier: int, method: str = 'get') -> Response: + self, identifier: str, method: str = 'get') -> Response: """Permit to ask information to the api prior to actually executing query. """ path = self.request_uri(identifier) params = self.request_params(identifier) logger.debug('path: %s', path) logger.debug('params: %s', params) logger.debug('method: %s', method) try: if method == 'head': response = self.session.head(path, **params) else: response = self.session.get(path, **params) except requests.exceptions.ConnectionError as e: logger.warning('Failed to fetch %s: %s', path, e) raise FetchError(e) else: if response.status_code not in self.EXPECTED_STATUS_CODES: raise FetchError(response) return response - def transport_head(self, identifier: int) -> Response: + def transport_head(self, identifier: str) -> Response: """Retrieve head information on api. """ return self._transport_action(identifier, method='head') - def transport_request(self, identifier: int) -> Response: + def transport_request(self, identifier: str) -> Response: """Implements ListerBase.transport_request for HTTP using Requests. Retrieve get information on api. """ return self._transport_action(identifier) def transport_response_to_string(self, response: Response) -> str: """Implements ListerBase.transport_response_to_string for HTTP given Requests responses. """ s = pformat(response.request.path_url) s += '\n#\n' + pformat(response.request.headers) s += '\n#\n' + pformat(response.status_code) s += '\n#\n' + pformat(response.headers) s += '\n#\n' try: # json? s += pformat(response.json()) except Exception: # not json try: # xml? s += pformat(xmltodict.parse(response.text)) except Exception: # not xml s += pformat(response.text) return s class ListerOnePageApiTransport(ListerHttpTransport): """Leverage requests library to retrieve basic html page and parse result. To be used in conjunction with ListerBase or a subclass of it. """ PAGE = AbstractAttribute( "URL of the API's unique page to retrieve and parse " "for information") # type: Union[AbstractAttribute, str] PATH_TEMPLATE = None # we do not use it def __init__(self, url=None): self.session = requests.Session() self.lister_version = __version__ def request_uri(self, _): """Get the full request URI given the transport_request identifier. """ return self.PAGE diff --git a/swh/lister/gitlab/lister.py b/swh/lister/gitlab/lister.py index 3d8d3d7..d3e45bf 100644 --- a/swh/lister/gitlab/lister.py +++ b/swh/lister/gitlab/lister.py @@ -1,91 +1,91 @@ # Copyright (C) 2018-2019 The Software Heritage developers # See the AUTHORS file at the top-level directory of this distribution # License: GNU General Public License version 3, or any later version # See top-level LICENSE file for more information import time from urllib3.util import parse_url from ..core.page_by_page_lister import PageByPageHttpLister from .models import GitLabModel from typing import Any, Dict, List, Tuple, Union, MutableMapping, Optional from requests import Response class GitLabLister(PageByPageHttpLister): # Template path expecting an integer that represents the page id PATH_TEMPLATE = '/projects?page=%d&order_by=id' DEFAULT_URL = 'https://gitlab.com/api/v4/' MODEL = GitLabModel LISTER_NAME = 'gitlab' def __init__(self, url=None, instance=None, override_config=None, sort='asc', per_page=20): super().__init__(url=url, override_config=override_config) if instance is None: instance = parse_url(self.url).host self.instance = instance self.PATH_TEMPLATE = '%s&sort=%s&per_page=%s' % ( self.PATH_TEMPLATE, sort, per_page) def uid(self, repo: Dict[str, Any]) -> str: return '%s/%s' % (self.instance, repo['path_with_namespace']) def get_model_from_repo(self, repo: Dict[str, Any]) -> Dict[str, Any]: return { 'instance': self.instance, 'uid': self.uid(repo), 'name': repo['name'], 'full_name': repo['path_with_namespace'], 'html_url': repo['web_url'], 'origin_url': repo['http_url_to_repo'], 'origin_type': 'git', } def transport_quota_check(self, response: Response ) -> Tuple[bool, Union[int, float]]: """Deal with rate limit if any. """ # not all gitlab instance have rate limit if 'RateLimit-Remaining' in response.headers: reqs_remaining = int(response.headers['RateLimit-Remaining']) if response.status_code == 403 and reqs_remaining == 0: reset_at = int(response.headers['RateLimit-Reset']) delay = min(reset_at - time.time(), 3600) return True, delay return False, 0 def _get_int(self, headers: MutableMapping[str, Any], key: str) -> Optional[int]: _val = headers.get(key) if _val: return int(_val) return None def get_next_target_from_response( self, response: Response) -> Optional[int]: """Determine the next page identifier. """ return self._get_int(response.headers, 'x-next-page') def get_pages_information(self) -> Tuple[Optional[int], Optional[int], Optional[int]]: """Determine pages information. """ - response = self.transport_head(identifier=1) + response = self.transport_head(identifier=1) # type: ignore if not response.ok: raise ValueError( 'Problem during information fetch: %s' % response.status_code) h = response.headers return (self._get_int(h, 'x-total'), self._get_int(h, 'x-total-pages'), self._get_int(h, 'x-per-page')) def transport_response_simplified(self, response: Response ) -> List[Dict[str, Any]]: repos = response.json() return [self.get_model_from_repo(repo) for repo in repos] diff --git a/swh/lister/phabricator/lister.py b/swh/lister/phabricator/lister.py index a32a6c5..bfb4a95 100644 --- a/swh/lister/phabricator/lister.py +++ b/swh/lister/phabricator/lister.py @@ -1,190 +1,190 @@ # Copyright (C) 2019 the Software Heritage developers # License: GNU General Public License version 3, or any later version # See top-level LICENSE file for more information import logging import random import urllib.parse from collections import defaultdict from sqlalchemy import func from swh.lister.core.indexing_lister import IndexingHttpLister from swh.lister.phabricator.models import PhabricatorModel from typing import Any, Dict, List, Optional from requests import Response logger = logging.getLogger(__name__) class PhabricatorLister(IndexingHttpLister): PATH_TEMPLATE = '?order=oldest&attachments[uris]=1&after=%s' DEFAULT_URL = \ 'https://forge.softwareheritage.org/api/diffusion.repository.search' MODEL = PhabricatorModel LISTER_NAME = 'phabricator' def __init__(self, url=None, instance=None, override_config=None): super().__init__(url=url, override_config=override_config) if not instance: instance = urllib.parse.urlparse(self.url).hostname self.instance = instance - def request_params(self, identifier: int) -> Dict[str, Any]: + def request_params(self, identifier: str) -> Dict[str, Any]: """Override the default params behavior to retrieve the api token Credentials are stored as: credentials: phabricator: : - username: password: """ creds = self.request_instance_credentials() if not creds: raise ValueError( 'Phabricator forge needs authentication credential to list.') api_token = random.choice(creds)['password'] return {'headers': self.request_headers() or {}, 'params': {'api.token': api_token}} def request_headers(self): """ (Override) Set requests headers to send when querying the Phabricator API """ headers = super().request_headers() headers['Accept'] = 'application/json' return headers def get_model_from_repo( self, repo: Dict[str, Any]) -> Optional[Dict[str, Any]]: url = get_repo_url(repo['attachments']['uris']['uris']) if url is None: return None return { 'uid': url, 'indexable': repo['id'], 'name': repo['fields']['shortName'], 'full_name': repo['fields']['name'], 'html_url': url, 'origin_url': url, 'origin_type': repo['fields']['vcs'], 'instance': self.instance, } def get_next_target_from_response( self, response: Response) -> Optional[int]: body = response.json()['result']['cursor'] if body['after'] and body['after'] != 'null': return int(body['after']) return None def transport_response_simplified( self, response: Response) -> List[Optional[Dict[str, Any]]]: repos = response.json() if repos['result'] is None: raise ValueError( 'Problem during information fetch: %s' % repos['error_code']) repos = repos['result']['data'] return [self.get_model_from_repo(repo) for repo in repos] def filter_before_inject(self, models_list): """ (Overrides) IndexingLister.filter_before_inject Bounds query results by this Lister's set max_index. """ models_list = [m for m in models_list if m is not None] return super().filter_before_inject(models_list) def disable_deleted_repo_tasks( self, index: int, next_index: int, keep_these: str): """ (Overrides) Fix provided index value to avoid: - database query error - erroneously disabling some scheduler tasks """ # First call to the Phabricator API uses an empty 'after' parameter, # so set the index to 0 to avoid database query error if index == '': index = 0 # Next listed repository ids are strictly greater than the 'after' # parameter, so increment the index to avoid disabling the latest # created task when processing a new repositories page returned by # the Phabricator API else: index = index + 1 return super().disable_deleted_repo_tasks(index, next_index, keep_these) def db_first_index(self) -> Optional[int]: """ (Overrides) Filter results by Phabricator instance Returns: the smallest indexable value of all repos in the db """ t = self.db_session.query(func.min(self.MODEL.indexable)) t = t.filter(self.MODEL.instance == self.instance).first() if t: return t[0] return None def db_last_index(self): """ (Overrides) Filter results by Phabricator instance Returns: the largest indexable value of all instance repos in the db """ t = self.db_session.query(func.max(self.MODEL.indexable)) t = t.filter(self.MODEL.instance == self.instance).first() if t: return t[0] def db_query_range(self, start: int, end: int): """ (Overrides) Filter the results by the Phabricator instance to avoid disabling loading tasks for repositories hosted on a different instance. Returns: a list of sqlalchemy.ext.declarative.declarative_base objects with indexable values within the given range for the instance """ retlist = super().db_query_range(start, end) return retlist.filter(self.MODEL.instance == self.instance) def get_repo_url(attachments: List[Dict[str, Any]]) -> Optional[int]: """ Return url for a hosted repository from its uris attachments according to the following priority lists: * protocol: https > http * identifier: shortname > callsign > id """ processed_urls = defaultdict(dict) # type: Dict[str, Any] for uri in attachments: protocol = uri['fields']['builtin']['protocol'] url = uri['fields']['uri']['effective'] identifier = uri['fields']['builtin']['identifier'] if protocol in ('http', 'https'): processed_urls[protocol][identifier] = url elif protocol is None: for protocol in ('https', 'http'): if url.startswith(protocol): processed_urls[protocol]['undefined'] = url break for protocol in ['https', 'http']: for identifier in ['shortname', 'callsign', 'id', 'undefined']: if (protocol in processed_urls and identifier in processed_urls[protocol]): return processed_urls[protocol][identifier] return None