diff --git a/swh/lister/gitlab/lister.py b/swh/lister/gitlab/lister.py index 2cff743..b588083 100644 --- a/swh/lister/gitlab/lister.py +++ b/swh/lister/gitlab/lister.py @@ -1,125 +1,122 @@ # Copyright (C) 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 random -import re import time from .. import utils from ..core.paging_lister import PageByPageHttpLister from .models import GitLabModel class GitLabLister(PageByPageHttpLister): # Template path expecting an integer that represents the page id PATH_TEMPLATE = '/projects?page=%d&order_by=id' - API_URL_INDEX_RE = re.compile(r'^.*/projects.*page=(\d+).*') MODEL = GitLabModel LISTER_NAME = 'gitlab' def __init__(self, api_baseurl=None, instance=None, override_config=None, sort='asc'): super().__init__(api_baseurl=api_baseurl, override_config=override_config) self.instance = instance self.PATH_TEMPLATE = '%s&sort=%s' % (self.PATH_TEMPLATE, sort) @property def ADDITIONAL_CONFIG(self): """Override additional config as the 'credentials' structure change between the ancestor classes and this class. cf. request_params method below """ default_config = super().ADDITIONAL_CONFIG # 'credentials' is a dict of (instance, {username, password}) dict default_config['credentials'] = ('dict', {}) return default_config def request_params(self, identifier): """Get the full parameters passed to requests given the transport_request identifier. For the gitlab lister, the 'credentials' entries is configured per instance. For example: - credentials: - gitlab.com: - username: user0 password: - username: user1 password: - ... - other-gitlab-instance: ... """ params = { 'headers': self.request_headers() or {} } # Retrieve the credentials per instance creds = self.config['credentials'] if creds: creds_lister = creds[self.instance] auth = random.choice(creds_lister) if creds else None if auth: params['auth'] = (auth['username'], auth['password']) return params def get_model_from_repo(self, repo): return { 'instance': self.instance, 'uid': repo['id'], 'indexable': repo['id'], 'name': repo['name'], 'full_name': repo['path_with_namespace'], 'html_url': repo['web_url'], 'origin_url': repo['http_url_to_repo'], 'origin_type': 'git', 'description': repo['description'], } def transport_quota_check(self, response): """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_next_target_from_response(self, response): """Determine the next page identifier. """ - if 'next' in response.links: - next_url = response.links['next']['url'] - return int(self.API_URL_INDEX_RE.match(next_url).group(1)) - return None + _next = utils.get(response.headers, ['X-Next-Page', 'x-next-page']) + if _next: + return int(_next) def get_pages_information(self): """Determine pages information. """ response = self.transport_head(identifier=1) h = response.headers total = utils.get(h, ['X-Total', 'x-total']) total_pages = utils.get(h, ['X-Total-Pages', 'x-total-pages']) per_page = utils.get(h, ['X-Per-Page', 'x-per-page']) if total is not None: total = int(total) if total_pages is not None: total_pages = int(total_pages) if per_page is not None: per_page = int(per_page) return total, total_pages, per_page def transport_response_simplified(self, response): repos = response.json() return [self.get_model_from_repo(repo) for repo in repos] diff --git a/swh/lister/gitlab/tests/test_gitlab_lister.py b/swh/lister/gitlab/tests/test_gitlab_lister.py index 2d3a416..9d17330 100644 --- a/swh/lister/gitlab/tests/test_gitlab_lister.py +++ b/swh/lister/gitlab/tests/test_gitlab_lister.py @@ -1,46 +1,38 @@ # 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 re import unittest from datetime import datetime, timedelta from swh.lister.gitlab.lister import GitLabLister from swh.lister.core.tests.test_lister import HttpListerTesterBase class GitLabListerTester(HttpListerTesterBase, unittest.TestCase): Lister = GitLabLister test_re = re.compile(r'^.*/projects.*page=(\d+).*') lister_subdir = 'gitlab' good_api_response_file = 'api_response.json' bad_api_response_file = 'api_empty_response.json' first_index = 1 entries_per_page = 10 def response_headers(self, request): headers = {'RateLimit-Remaining': '1'} if self.request_index(request) == str(self.first_index): headers.update({ - 'Link': ';' - ' rel="next",' - ';' - ' rel="first"' - }) - else: - headers.update({ - 'Link': ';' - ' rel="first"' + 'x-next-page': '3', }) return headers def mock_rate_quota(self, n, request, context): self.rate_limit += 1 context.status_code = 403 context.headers['RateLimit-Remaining'] = '0' one_second = int((datetime.now() + timedelta(seconds=1.5)).timestamp()) context.headers['RateLimit-Reset'] = str(one_second) return '{"error":"dummy"}'