diff --git a/docs/uri-scheme-api-person.rst b/docs/uri-scheme-api-person.rst deleted file mode 100644 index cec4e9de..00000000 --- a/docs/uri-scheme-api-person.rst +++ /dev/null @@ -1,4 +0,0 @@ -Person ------- - -.. autosimple:: swh.web.api.views.person.api_person diff --git a/docs/uri-scheme-api-revision.rst b/docs/uri-scheme-api-revision.rst index 3e176ee4..63521d5d 100644 --- a/docs/uri-scheme-api-revision.rst +++ b/docs/uri-scheme-api-revision.rst @@ -1,12 +1,9 @@ Revision -------- .. autosimple:: swh.web.api.views.revision.api_revision .. autosimple:: swh.web.api.views.revision.api_revision_directory .. autosimple:: swh.web.api.views.revision.api_revision_log -.. autosimple:: swh.web.api.views.revision.api_revision_with_origin - -.. autosimple:: swh.web.api.views.revision.api_revision_log_by diff --git a/docs/uri-scheme-api.rst b/docs/uri-scheme-api.rst index d451f879..f435743d 100644 --- a/docs/uri-scheme-api.rst +++ b/docs/uri-scheme-api.rst @@ -1,24 +1,22 @@ .. _swh-web-api-urls: swh-web API URLs ================ .. include:: uri-scheme-api-content.rst .. include:: uri-scheme-api-directory.rst .. include:: uri-scheme-api-identifiers.rst .. include:: uri-scheme-api-origin.rst -.. include:: uri-scheme-api-person.rst - .. include:: uri-scheme-api-release.rst .. include:: uri-scheme-api-revision.rst .. include:: uri-scheme-api-snapshot.rst .. include:: uri-scheme-api-stat.rst .. include:: uri-scheme-api-vault.rst diff --git a/swh/web/browse/utils.py b/swh/web/browse/utils.py index 038cfed4..a57b9a1d 100644 --- a/swh/web/browse/utils.py +++ b/swh/web/browse/utils.py @@ -1,1098 +1,1100 @@ # Copyright (C) 2017-2019 The Software Heritage developers # See the AUTHORS file at the top-level directory of this distribution # License: GNU Affero General Public License version 3, or any later version # See top-level LICENSE file for more information import base64 import magic import pypandoc import stat import textwrap from collections import defaultdict from threading import Lock from django.core.cache import cache from django.utils.safestring import mark_safe from django.utils.html import escape from swh.model.identifiers import persistent_identifier from swh.web.common import highlightjs, service from swh.web.common.exc import NotFoundExc, http_status_code_message from swh.web.common.origin_visits import get_origin_visit from swh.web.common.utils import ( reverse, format_utc_iso_date, get_swh_persistent_id, swh_object_icons ) from swh.web.config import get_config def get_directory_entries(sha1_git): """Function that retrieves the content of a directory from the archive. The directories entries are first sorted in lexicographical order. Sub-directories and regular files are then extracted. Args: sha1_git: sha1_git identifier of the directory Returns: A tuple whose first member corresponds to the sub-directories list and second member the regular files list Raises: NotFoundExc if the directory is not found """ cache_entry_id = 'directory_entries_%s' % sha1_git cache_entry = cache.get(cache_entry_id) if cache_entry: return cache_entry entries = list(service.lookup_directory(sha1_git)) for e in entries: e['perms'] = stat.filemode(e['perms']) if e['type'] == 'rev': # modify dir entry name to explicitly show it points # to a revision e['name'] = '%s @ %s' % (e['name'], e['target'][:7]) dirs = [e for e in entries if e['type'] in ('dir', 'rev')] files = [e for e in entries if e['type'] == 'file'] dirs = sorted(dirs, key=lambda d: d['name']) files = sorted(files, key=lambda f: f['name']) cache.set(cache_entry_id, (dirs, files)) return dirs, files _lock = Lock() def get_mimetype_and_encoding_for_content(content): """Function that returns the mime type and the encoding associated to a content buffer using the magic module under the hood. Args: content (bytes): a content buffer Returns: A tuple (mimetype, encoding), for instance ('text/plain', 'us-ascii'), associated to the provided content. """ # https://pypi.org/project/python-magic/ # packaged as python3-magic in debian buster if hasattr(magic, 'from_buffer'): m = magic.Magic(mime=True, mime_encoding=True) mime_encoding = m.from_buffer(content) mime_type, encoding = mime_encoding.split(';') encoding = encoding.replace(' charset=', '') # https://pypi.org/project/file-magic/ # packaged as python3-magic in debian stretch else: # TODO: Remove that code when production environment is upgraded # to debian buster # calls to the file-magic API are not thread-safe so they must # be protected with a Lock to guarantee they will succeed _lock.acquire() magic_result = magic.detect_from_content(content) _lock.release() mime_type = magic_result.mime_type encoding = magic_result.encoding return mime_type, encoding # maximum authorized content size in bytes for HTML display # with code highlighting content_display_max_size = get_config()['content_display_max_size'] snapshot_content_max_size = get_config()['snapshot_content_max_size'] def _re_encode_content(mimetype, encoding, content_data): # encode textual content to utf-8 if needed if mimetype.startswith('text/'): # probably a malformed UTF-8 content, re-encode it # by replacing invalid chars with a substitution one if encoding == 'unknown-8bit': content_data = content_data.decode('utf-8', 'replace')\ .encode('utf-8') elif encoding not in ['utf-8', 'binary']: content_data = content_data.decode(encoding, 'replace')\ .encode('utf-8') elif mimetype.startswith('application/octet-stream'): # file may detect a text content as binary # so try to decode it for display encodings = ['us-ascii'] encodings += ['iso-8859-%s' % i for i in range(1, 17)] for encoding in encodings: try: content_data = content_data.decode(encoding)\ .encode('utf-8') except Exception: pass else: # ensure display in content view mimetype = 'text/plain' break return mimetype, content_data def request_content(query_string, max_size=content_display_max_size, raise_if_unavailable=True, re_encode=True): """Function that retrieves a content from the archive. Raw bytes content is first retrieved, then the content mime type. If the mime type is not stored in the archive, it will be computed using Python magic module. Args: query_string: a string of the form "[ALGO_HASH:]HASH" where optional ALGO_HASH can be either ``sha1``, ``sha1_git``, ``sha256``, or ``blake2s256`` (default to ``sha1``) and HASH the hexadecimal representation of the hash value max_size: the maximum size for a content to retrieve (default to 1MB, no size limit if None) Returns: A tuple whose first member corresponds to the content raw bytes and second member the content mime type Raises: NotFoundExc if the content is not found """ content_data = service.lookup_content(query_string) filetype = None language = None license = None # requests to the indexer db may fail so properly handle # those cases in order to avoid content display errors try: filetype = service.lookup_content_filetype(query_string) language = service.lookup_content_language(query_string) license = service.lookup_content_license(query_string) except Exception: pass mimetype = 'unknown' encoding = 'unknown' if filetype: mimetype = filetype['mimetype'] encoding = filetype['encoding'] # workaround when encountering corrupted data due to implicit # conversion from bytea to text in the indexer db (see T818) # TODO: Remove that code when all data have been correctly converted if mimetype.startswith('\\'): filetype = None content_data['error_code'] = 200 content_data['error_message'] = '' content_data['error_description'] = '' if not max_size or content_data['length'] < max_size: try: content_raw = service.lookup_content_raw(query_string) except Exception as e: if raise_if_unavailable: raise e else: content_data['raw_data'] = None content_data['error_code'] = 404 content_data['error_description'] = \ 'The bytes of the content are currently not available in the archive.' # noqa content_data['error_message'] = \ http_status_code_message[content_data['error_code']] else: content_data['raw_data'] = content_raw['data'] if not filetype: mimetype, encoding = \ get_mimetype_and_encoding_for_content(content_data['raw_data']) # noqa if re_encode: mimetype, raw_data = _re_encode_content( mimetype, encoding, content_data['raw_data']) content_data['raw_data'] = raw_data else: content_data['raw_data'] = None content_data['mimetype'] = mimetype content_data['encoding'] = encoding if language: content_data['language'] = language['lang'] else: content_data['language'] = 'not detected' if license: content_data['licenses'] = ', '.join(license['facts'][0]['licenses']) else: content_data['licenses'] = 'not detected' return content_data _browsers_supported_image_mimes = set(['image/gif', 'image/png', 'image/jpeg', 'image/bmp', 'image/webp', 'image/svg', 'image/svg+xml']) def prepare_content_for_display(content_data, mime_type, path): """Function that prepares a content for HTML display. The function tries to associate a programming language to a content in order to perform syntax highlighting client-side using highlightjs. The language is determined using either the content filename or its mime type. If the mime type corresponds to an image format supported by web browsers, the content will be encoded in base64 for displaying the image. Args: content_data (bytes): raw bytes of the content mime_type (string): mime type of the content path (string): path of the content including filename Returns: A dict containing the content bytes (possibly different from the one provided as parameter if it is an image) under the key 'content_data and the corresponding highlightjs language class under the key 'language'. """ language = highlightjs.get_hljs_language_from_filename(path) if not language: language = highlightjs.get_hljs_language_from_mime_type(mime_type) if not language: language = 'nohighlight' elif mime_type.startswith('application/'): mime_type = mime_type.replace('application/', 'text/') if mime_type.startswith('image/'): if mime_type in _browsers_supported_image_mimes: content_data = base64.b64encode(content_data) content_data = content_data.decode('utf-8') else: content_data = None if mime_type.startswith('image/svg'): mime_type = 'image/svg+xml' return {'content_data': content_data, 'language': language, 'mimetype': mime_type} def process_snapshot_branches(snapshot): """ Process a dictionary describing snapshot branches: extract those targeting revisions and releases, put them in two different lists, then sort those lists in lexicographical order of the branches' names. Args: snapshot_branches (dict): A dict describing the branches of a snapshot as returned for instance by :func:`swh.web.common.service.lookup_snapshot` Returns: tuple: A tuple whose first member is the sorted list of branches targeting revisions and second member the sorted list of branches targeting releases """ snapshot_branches = snapshot['branches'] branches = {} branch_aliases = {} releases = {} revision_to_branch = defaultdict(set) revision_to_release = defaultdict(set) release_to_branch = defaultdict(set) for branch_name, target in snapshot_branches.items(): if not target: # FIXME: display branches with an unknown target anyway continue target_id = target['target'] target_type = target['target_type'] if target_type == 'revision': branches[branch_name] = { 'name': branch_name, 'revision': target_id, } revision_to_branch[target_id].add(branch_name) elif target_type == 'release': release_to_branch[target_id].add(branch_name) elif target_type == 'alias': branch_aliases[branch_name] = target_id # FIXME: handle pointers to other object types def _enrich_release_branch(branch, release): releases[branch] = { 'name': release['name'], 'branch_name': branch, 'date': format_utc_iso_date(release['date']), 'id': release['id'], 'message': release['message'], 'target_type': release['target_type'], 'target': release['target'], } def _enrich_revision_branch(branch, revision): branches[branch].update({ 'revision': revision['id'], 'directory': revision['directory'], 'date': format_utc_iso_date(revision['date']), 'message': revision['message'] }) releases_info = service.lookup_release_multiple( release_to_branch.keys() ) for release in releases_info: branches_to_update = release_to_branch[release['id']] for branch in branches_to_update: _enrich_release_branch(branch, release) if release['target_type'] == 'revision': revision_to_release[release['target']].update( branches_to_update ) revisions = service.lookup_revision_multiple( set(revision_to_branch.keys()) | set(revision_to_release.keys()) ) for revision in revisions: if not revision: continue for branch in revision_to_branch[revision['id']]: _enrich_revision_branch(branch, revision) for release in revision_to_release[revision['id']]: releases[release]['directory'] = revision['directory'] for branch_alias, branch_target in branch_aliases.items(): if branch_target in branches: branches[branch_alias] = dict(branches[branch_target]) else: snp = service.lookup_snapshot(snapshot['id'], branches_from=branch_target, branches_count=1) if snp and branch_target in snp['branches']: if snp['branches'][branch_target] is None: continue target_type = snp['branches'][branch_target]['target_type'] target = snp['branches'][branch_target]['target'] if target_type == 'revision': branches[branch_alias] = snp['branches'][branch_target] revision = service.lookup_revision(target) _enrich_revision_branch(branch_alias, revision) elif target_type == 'release': release = service.lookup_release(target) _enrich_release_branch(branch_alias, release) if branch_alias in branches: branches[branch_alias]['name'] = branch_alias ret_branches = list(sorted(branches.values(), key=lambda b: b['name'])) ret_releases = list(sorted(releases.values(), key=lambda b: b['name'])) return ret_branches, ret_releases def get_snapshot_content(snapshot_id): """Returns the lists of branches and releases associated to a swh snapshot. That list is put in cache in order to speedup the navigation in the swh-web/browse ui. .. warning:: At most 1000 branches contained in the snapshot will be returned for performance reasons. Args: snapshot_id (str): hexadecimal representation of the snapshot identifier Returns: A tuple with two members. The first one is a list of dict describing the snapshot branches. The second one is a list of dict describing the snapshot releases. Raises: NotFoundExc if the snapshot does not exist """ cache_entry_id = 'swh_snapshot_%s' % snapshot_id cache_entry = cache.get(cache_entry_id) if cache_entry: return cache_entry['branches'], cache_entry['releases'] branches = [] releases = [] if snapshot_id: snapshot = service.lookup_snapshot( snapshot_id, branches_count=snapshot_content_max_size) branches, releases = process_snapshot_branches(snapshot) cache.set(cache_entry_id, { 'branches': branches, 'releases': releases, }) return branches, releases def get_origin_visit_snapshot(origin_info, visit_ts=None, visit_id=None, snapshot_id=None): """Returns the lists of branches and releases associated to a swh origin for a given visit. The visit is expressed by a timestamp. In the latter case, the closest visit from the provided timestamp will be used. If no visit parameter is provided, it returns the list of branches found for the latest visit. That list is put in cache in order to speedup the navigation in the swh-web/browse ui. .. warning:: At most 1000 branches contained in the snapshot will be returned for performance reasons. Args: origin_info (dict): a dict filled with origin information (id, url, type) visit_ts (int or str): an ISO date string or Unix timestamp to parse visit_id (int): optional visit id for disambiguation in case several visits have the same timestamp Returns: A tuple with two members. The first one is a list of dict describing the origin branches for the given visit. The second one is a list of dict describing the origin releases for the given visit. Raises: NotFoundExc if the origin or its visit are not found """ visit_info = get_origin_visit(origin_info, visit_ts, visit_id, snapshot_id) return get_snapshot_content(visit_info['snapshot']) def gen_link(url, link_text=None, link_attrs=None): """ Utility function for generating an HTML link to insert in Django templates. Args: url (str): an url link_text (str): optional text for the produced link, if not provided the url will be used link_attrs (dict): optional attributes (e.g. class) to add to the link Returns: An HTML link in the form 'link_text' """ attrs = ' ' if link_attrs: for k, v in link_attrs.items(): attrs += '%s="%s" ' % (k, v) if not link_text: link_text = url link = '%s' \ % (attrs, escape(url), escape(link_text)) return mark_safe(link) def _snapshot_context_query_params(snapshot_context): query_params = None if snapshot_context and snapshot_context['origin_info']: origin_info = snapshot_context['origin_info'] query_params = {'origin': origin_info['url']} if 'timestamp' in snapshot_context['url_args']: query_params['timestamp'] = \ snapshot_context['url_args']['timestamp'] if 'visit_id' in snapshot_context['query_params']: query_params['visit_id'] = \ snapshot_context['query_params']['visit_id'] elif snapshot_context: query_params = {'snapshot_id': snapshot_context['snapshot_id']} return query_params def gen_revision_url(revision_id, snapshot_context=None): """ Utility function for generating an url to a revision. Args: revision_id (str): a revision id snapshot_context (dict): if provided, generate snapshot-dependent browsing url Returns: str: The url to browse the revision """ query_params = _snapshot_context_query_params(snapshot_context) return reverse('browse-revision', url_args={'sha1_git': revision_id}, query_params=query_params) def gen_revision_link(revision_id, shorten_id=False, snapshot_context=None, link_text='Browse', link_attrs={'class': 'btn btn-default btn-sm', 'role': 'button'}): """ Utility function for generating a link to a revision HTML view to insert in Django templates. Args: revision_id (str): a revision id shorten_id (boolean): whether to shorten the revision id to 7 characters for the link text snapshot_context (dict): if provided, generate snapshot-dependent browsing link link_text (str): optional text for the generated link (the revision id will be used by default) link_attrs (dict): optional attributes (e.g. class) to add to the link Returns: str: An HTML link in the form 'revision_id' """ if not revision_id: return None revision_url = gen_revision_url(revision_id, snapshot_context) if shorten_id: return gen_link(revision_url, revision_id[:7], link_attrs) else: if not link_text: link_text = revision_id return gen_link(revision_url, link_text, link_attrs) def gen_directory_link(sha1_git, snapshot_context=None, link_text='Browse', link_attrs={'class': 'btn btn-default btn-sm', 'role': 'button'}): """ Utility function for generating a link to a directory HTML view to insert in Django templates. Args: sha1_git (str): directory identifier link_text (str): optional text for the generated link (the directory id will be used by default) link_attrs (dict): optional attributes (e.g. class) to add to the link Returns: An HTML link in the form 'link_text' """ if not sha1_git: return None query_params = _snapshot_context_query_params(snapshot_context) directory_url = reverse('browse-directory', url_args={'sha1_git': sha1_git}, query_params=query_params) if not link_text: link_text = sha1_git return gen_link(directory_url, link_text, link_attrs) def gen_snapshot_link(snapshot_id, snapshot_context=None, link_text='Browse', link_attrs={'class': 'btn btn-default btn-sm', 'role': 'button'}): """ Utility function for generating a link to a snapshot HTML view to insert in Django templates. Args: snapshot_id (str): snapshot identifier link_text (str): optional text for the generated link (the snapshot id will be used by default) link_attrs (dict): optional attributes (e.g. class) to add to the link Returns: An HTML link in the form 'link_text' """ query_params = _snapshot_context_query_params(snapshot_context) snapshot_url = reverse('browse-snapshot', url_args={'snapshot_id': snapshot_id}, query_params=query_params) if not link_text: link_text = snapshot_id return gen_link(snapshot_url, link_text, link_attrs) def gen_content_link(sha1_git, snapshot_context=None, link_text='Browse', link_attrs={'class': 'btn btn-default btn-sm', 'role': 'button'}): """ Utility function for generating a link to a content HTML view to insert in Django templates. Args: sha1_git (str): content identifier link_text (str): optional text for the generated link (the content sha1_git will be used by default) link_attrs (dict): optional attributes (e.g. class) to add to the link Returns: An HTML link in the form 'link_text' """ if not sha1_git: return None query_params = _snapshot_context_query_params(snapshot_context) content_url = reverse('browse-content', url_args={'query_string': 'sha1_git:' + sha1_git}, query_params=query_params) if not link_text: link_text = sha1_git return gen_link(content_url, link_text, link_attrs) def get_revision_log_url(revision_id, snapshot_context=None): """ Utility function for getting the URL for a revision log HTML view (possibly in the context of an origin). Args: revision_id (str): revision identifier the history heads to snapshot_context (dict): if provided, generate snapshot-dependent browsing link Returns: The revision log view URL """ query_params = {'revision': revision_id} if snapshot_context and snapshot_context['origin_info']: origin_info = snapshot_context['origin_info'] url_args = {'origin_url': origin_info['url']} if 'timestamp' in snapshot_context['url_args']: url_args['timestamp'] = \ snapshot_context['url_args']['timestamp'] if 'visit_id' in snapshot_context['query_params']: query_params['visit_id'] = \ snapshot_context['query_params']['visit_id'] revision_log_url = reverse('browse-origin-log', url_args=url_args, query_params=query_params) elif snapshot_context: url_args = {'snapshot_id': snapshot_context['snapshot_id']} revision_log_url = reverse('browse-snapshot-log', url_args=url_args, query_params=query_params) else: revision_log_url = reverse('browse-revision-log', url_args={'sha1_git': revision_id}) return revision_log_url def gen_revision_log_link(revision_id, snapshot_context=None, link_text='Browse', link_attrs={'class': 'btn btn-default btn-sm', 'role': 'button'}): """ Utility function for generating a link to a revision log HTML view (possibly in the context of an origin) to insert in Django templates. Args: revision_id (str): revision identifier the history heads to snapshot_context (dict): if provided, generate snapshot-dependent browsing link link_text (str): optional text to use for the generated link (the revision id will be used by default) link_attrs (dict): optional attributes (e.g. class) to add to the link Returns: An HTML link in the form 'link_text' """ if not revision_id: return None revision_log_url = get_revision_log_url(revision_id, snapshot_context) if not link_text: link_text = revision_id return gen_link(revision_log_url, link_text, link_attrs) def gen_person_mail_link(person, link_text=None): """ Utility function for generating a mail link to a person to insert in Django templates. Args: person (dict): dictionary containing person data (*name*, *email*, *fullname*) link_text (str): optional text to use for the generated mail link (the person name will be used by default) Returns: str: A mail link to the person or the person name if no email is present in person data """ person_name = person['name'] or person['fullname'] or 'None' if link_text is None: link_text = person_name person_email = person['email'] if person['email'] else None if person_email is None and '@' in person_name and ' ' not in person_name: person_email = person_name if person_email: return gen_link(url='mailto:%s' % person_email, link_text=link_text) else: return person_name def gen_release_link(sha1_git, snapshot_context=None, link_text='Browse', link_attrs={'class': 'btn btn-default btn-sm', 'role': 'button'}): """ Utility function for generating a link to a release HTML view to insert in Django templates. Args: sha1_git (str): release identifier link_text (str): optional text for the generated link (the release id will be used by default) link_attrs (dict): optional attributes (e.g. class) to add to the link Returns: An HTML link in the form 'link_text' """ query_params = _snapshot_context_query_params(snapshot_context) release_url = reverse('browse-release', url_args={'sha1_git': sha1_git}, query_params=query_params) if not link_text: link_text = sha1_git return gen_link(release_url, link_text, link_attrs) def format_log_entries(revision_log, per_page, snapshot_context=None): """ Utility functions that process raw revision log data for HTML display. Its purpose is to: * add links to relevant browse views * format date in human readable format * truncate the message log Args: revision_log (list): raw revision log as returned by the swh-web api per_page (int): number of log entries per page snapshot_context (dict): if provided, generate snapshot-dependent browsing link """ revision_log_data = [] for i, rev in enumerate(revision_log): if i == per_page: break author_name = 'None' author_fullname = 'None' committer_fullname = 'None' if rev['author']: author_name = gen_person_mail_link(rev['author']) author_fullname = rev['author']['fullname'] if rev['committer']: committer_fullname = rev['committer']['fullname'] author_date = format_utc_iso_date(rev['date']) committer_date = format_utc_iso_date(rev['committer_date']) tooltip = 'revision %s\n' % rev['id'] tooltip += 'author: %s\n' % author_fullname tooltip += 'author date: %s\n' % author_date tooltip += 'committer: %s\n' % committer_fullname tooltip += 'committer date: %s\n\n' % committer_date if rev['message']: tooltip += textwrap.indent(rev['message'], ' '*4) revision_log_data.append({ 'author': author_name, 'id': rev['id'][:7], 'message': rev['message'], 'date': author_date, 'commit_date': committer_date, 'url': gen_revision_url(rev['id'], snapshot_context), 'tooltip': tooltip }) return revision_log_data def get_snapshot_context(snapshot_id=None, origin_url=None, timestamp=None, visit_id=None): """ Utility function to compute relevant information when navigating the archive in a snapshot context. The snapshot is either referenced by its id or it will be retrieved from an origin visit. Args: snapshot_id (str): hexadecimal representation of a snapshot identifier, all other parameters will be ignored if it is provided origin_url (str): the origin_url (e.g. https://github.com/(user)/(repo)/) timestamp (str): a datetime string for retrieving the closest visit of the origin visit_id (int): optional visit id for disambiguation in case of several visits with the same timestamp Returns: A dict with the following entries: * origin_info: dict containing origin information * visit_info: dict containing visit information * branches: the list of branches for the origin found during the visit * releases: the list of releases for the origin found during the visit * origin_browse_url: the url to browse the origin * origin_branches_url: the url to browse the origin branches * origin_releases_url': the url to browse the origin releases * origin_visit_url: the url to browse the snapshot of the origin found during the visit * url_args: dict containing url arguments to use when browsing in the context of the origin and its visit Raises: NotFoundExc: if no snapshot is found for the visit of an origin. """ origin_info = None visit_info = None url_args = None query_params = {} branches = [] releases = [] browse_url = None visit_url = None branches_url = None releases_url = None swh_type = 'snapshot' if origin_url: swh_type = 'origin' origin_info = service.lookup_origin({'url': origin_url}) visit_info = get_origin_visit(origin_info, timestamp, visit_id, snapshot_id) fmt_date = format_utc_iso_date(visit_info['date']) visit_info['fmt_date'] = fmt_date snapshot_id = visit_info['snapshot'] if not snapshot_id: raise NotFoundExc('No snapshot associated to the visit of origin ' '%s on %s' % (escape(origin_url), fmt_date)) # provided timestamp is not necessarily equals to the one # of the retrieved visit, so get the exact one in order # use it in the urls generated below if timestamp: timestamp = visit_info['date'] branches, releases = \ get_origin_visit_snapshot(origin_info, timestamp, visit_id, snapshot_id) url_args = {'origin_url': origin_info['url']} query_params = {'visit_id': visit_id} browse_url = reverse('browse-origin-visits', url_args=url_args) if timestamp: url_args['timestamp'] = format_utc_iso_date(timestamp, '%Y-%m-%dT%H:%M:%S') visit_url = reverse('browse-origin-directory', url_args=url_args, query_params=query_params) visit_info['url'] = visit_url branches_url = reverse('browse-origin-branches', url_args=url_args, query_params=query_params) releases_url = reverse('browse-origin-releases', url_args=url_args, query_params=query_params) elif snapshot_id: branches, releases = get_snapshot_content(snapshot_id) url_args = {'snapshot_id': snapshot_id} browse_url = reverse('browse-snapshot', url_args=url_args) branches_url = reverse('browse-snapshot-branches', url_args=url_args) releases_url = reverse('browse-snapshot-releases', url_args=url_args) releases = list(reversed(releases)) snapshot_size = service.lookup_snapshot_size(snapshot_id) is_empty = sum(snapshot_size.values()) == 0 swh_snp_id = persistent_identifier('snapshot', snapshot_id) return { 'swh_type': swh_type, 'swh_object_id': swh_snp_id, 'snapshot_id': snapshot_id, 'snapshot_size': snapshot_size, 'is_empty': is_empty, 'origin_info': origin_info, 'visit_info': visit_info, 'branches': branches, 'releases': releases, 'branch': None, 'release': None, 'browse_url': browse_url, 'branches_url': branches_url, 'releases_url': releases_url, 'url_args': url_args, 'query_params': query_params } # list of common readme names ordered by preference # (lower indices have higher priority) _common_readme_names = [ "readme.markdown", "readme.md", "readme.rst", "readme.txt", "readme" ] def get_readme_to_display(readmes): """ Process a list of readme files found in a directory in order to find the adequate one to display. Args: readmes: a list of dict where keys are readme file names and values are readme sha1s Returns: A tuple (readme_name, readme_sha1) """ readme_name = None readme_url = None readme_sha1 = None readme_html = None lc_readmes = {k.lower(): {'orig_name': k, 'sha1': v} for k, v in readmes.items()} # look for readme names according to the preference order # defined by the _common_readme_names list for common_readme_name in _common_readme_names: if common_readme_name in lc_readmes: readme_name = lc_readmes[common_readme_name]['orig_name'] readme_sha1 = lc_readmes[common_readme_name]['sha1'] readme_url = reverse('browse-content-raw', url_args={'query_string': readme_sha1}, query_params={'re_encode': 'true'}) break # otherwise pick the first readme like file if any if not readme_name and len(readmes.items()) > 0: readme_name = next(iter(readmes)) readme_sha1 = readmes[readme_name] readme_url = reverse('browse-content-raw', url_args={'query_string': readme_sha1}, query_params={'re_encode': 'true'}) # convert rst README to html server side as there is # no viable solution to perform that task client side if readme_name and readme_name.endswith('.rst'): cache_entry_id = 'readme_%s' % readme_sha1 cache_entry = cache.get(cache_entry_id) if cache_entry: readme_html = cache_entry else: try: rst_doc = request_content(readme_sha1) readme_html = pypandoc.convert_text(rst_doc['raw_data'], 'html', format='rst') cache.set(cache_entry_id, readme_html) except Exception: readme_html = 'Readme bytes are not available' return readme_name, readme_url, readme_html def get_swh_persistent_ids(swh_objects, snapshot_context=None): """ Returns a list of dict containing info related to persistent identifiers of swh objects. Args: swh_objects (list): a list of dict with the following keys: + * type: swh object type (content/directory/release/revision/snapshot) * id: swh object id + snapshot_context (dict): optional parameter describing the snapshot in which the object has been found Returns: list: a list of dict with the following keys: * object_type: the swh object type (content/directory/release/revision/snapshot) * object_icon: the swh object icon to use in HTML views * swh_id: the computed swh object persistent identifier * swh_id_url: the url resolving the persistent identifier * show_options: boolean indicating if the persistent id options must be displayed in persistent ids HTML view """ swh_ids = [] for swh_object in swh_objects: if not swh_object['id']: continue swh_id = get_swh_persistent_id(swh_object['type'], swh_object['id']) show_options = swh_object['type'] == 'content' or \ (snapshot_context and snapshot_context['origin_info'] is not None) object_icon = swh_object_icons[swh_object['type']] swh_ids.append({ 'object_type': swh_object['type'], 'object_icon': object_icon, 'swh_id': swh_id, 'swh_id_url': reverse('browse-swh-id', url_args={'swh_id': swh_id}), 'show_options': show_options }) return swh_ids diff --git a/swh/web/common/origin_save.py b/swh/web/common/origin_save.py index 28eaee57..e5231393 100644 --- a/swh/web/common/origin_save.py +++ b/swh/web/common/origin_save.py @@ -1,532 +1,535 @@ # Copyright (C) 2018-2019 The Software Heritage developers # See the AUTHORS file at the top-level directory of this distribution # License: GNU Affero General Public License version 3, or any later version # See top-level LICENSE file for more information import json import logging from bisect import bisect_right from datetime import datetime, timezone, timedelta import requests from django.core.exceptions import ObjectDoesNotExist from django.core.exceptions import ValidationError from django.core.validators import URLValidator from django.utils.html import escape from swh.web import config from swh.web.common import service from swh.web.common.exc import BadInputExc, ForbiddenExc, NotFoundExc from swh.web.common.models import ( SaveUnauthorizedOrigin, SaveAuthorizedOrigin, SaveOriginRequest, SAVE_REQUEST_ACCEPTED, SAVE_REQUEST_REJECTED, SAVE_REQUEST_PENDING, SAVE_TASK_NOT_YET_SCHEDULED, SAVE_TASK_SCHEDULED, SAVE_TASK_SUCCEED, SAVE_TASK_FAILED, SAVE_TASK_RUNNING ) from swh.web.common.origin_visits import get_origin_visits from swh.web.common.utils import parse_timestamp from swh.scheduler.utils import create_oneshot_task_dict scheduler = config.scheduler() logger = logging.getLogger(__name__) def get_origin_save_authorized_urls(): """ Get the list of origin url prefixes authorized to be immediately loaded into the archive (whitelist). Returns: list: The list of authorized origin url prefix """ return [origin.url for origin in SaveAuthorizedOrigin.objects.all()] def get_origin_save_unauthorized_urls(): """ Get the list of origin url prefixes forbidden to be loaded into the archive (blacklist). Returns: list: the list of unauthorized origin url prefix """ return [origin.url for origin in SaveUnauthorizedOrigin.objects.all()] def can_save_origin(origin_url): """ Check if a software origin can be saved into the archive. Based on the origin url, the save request will be either: * immediately accepted if the url is whitelisted * rejected if the url is blacklisted * put in pending state for manual review otherwise Args: origin_url (str): the software origin url to check Returns: str: the origin save request status, either **accepted**, **rejected** or **pending** """ # origin url may be blacklisted for url_prefix in get_origin_save_unauthorized_urls(): if origin_url.startswith(url_prefix): return SAVE_REQUEST_REJECTED # if the origin url is in the white list, it can be immediately saved for url_prefix in get_origin_save_authorized_urls(): if origin_url.startswith(url_prefix): return SAVE_REQUEST_ACCEPTED # otherwise, the origin url needs to be manually verified return SAVE_REQUEST_PENDING # map visit type to scheduler task # TODO: do not hardcode the task name here (T1157) _visit_type_task = { 'git': 'load-git', 'hg': 'load-hg', 'svn': 'load-svn' } # map scheduler task status to origin save status _save_task_status = { 'next_run_not_scheduled': SAVE_TASK_NOT_YET_SCHEDULED, 'next_run_scheduled': SAVE_TASK_SCHEDULED, 'completed': SAVE_TASK_SUCCEED, 'disabled': SAVE_TASK_FAILED } def get_savable_visit_types(): return sorted(list(_visit_type_task.keys())) def _check_visit_type_savable(visit_type): """ Get the list of visit types that can be performed through a save request. Returns: list: the list of saveable visit types """ allowed_visit_types = ', '.join(get_savable_visit_types()) if visit_type not in _visit_type_task: raise BadInputExc('Visit of type %s can not be saved! ' 'Allowed types are the following: %s' % (visit_type, allowed_visit_types)) _validate_url = URLValidator(schemes=['http', 'https', 'svn', 'git']) def _check_origin_url_valid(origin_url): try: _validate_url(origin_url) except ValidationError: raise BadInputExc('The provided origin url (%s) is not valid!' % escape(origin_url)) def _get_visit_info_for_save_request(save_request): visit_date = None visit_status = None try: origin = {'url': save_request.origin_url} origin_info = service.lookup_origin(origin) origin_visits = get_origin_visits(origin_info) visit_dates = [parse_timestamp(v['date']) for v in origin_visits] i = bisect_right(visit_dates, save_request.request_date) if i != len(visit_dates): visit_date = visit_dates[i] visit_status = origin_visits[i]['status'] if origin_visits[i]['status'] == 'ongoing': visit_date = None except Exception: pass return visit_date, visit_status def _check_visit_update_status(save_request, save_task_status): visit_date, visit_status = _get_visit_info_for_save_request(save_request) save_request.visit_date = visit_date # visit has been performed, mark the saving task as succeed if visit_date and visit_status is not None: save_task_status = SAVE_TASK_SUCCEED elif visit_status == 'ongoing': save_task_status = SAVE_TASK_RUNNING else: time_now = datetime.now(tz=timezone.utc) time_delta = time_now - save_request.request_date # consider the task as failed if it is still in scheduled state # 30 days after its submission if time_delta.days > 30: save_task_status = SAVE_TASK_FAILED return visit_date, save_task_status def _save_request_dict(save_request, task=None): must_save = False visit_date = save_request.visit_date # save task still in scheduler db if task: save_task_status = _save_task_status[task['status']] # Consider request from which a visit date has already been found # as succeeded to avoid retrieving it again if save_task_status == SAVE_TASK_SCHEDULED and visit_date: save_task_status = SAVE_TASK_SUCCEED if save_task_status in (SAVE_TASK_FAILED, SAVE_TASK_SUCCEED) \ and not visit_date: visit_date, _ = _get_visit_info_for_save_request(save_request) save_request.visit_date = visit_date must_save = True # Check tasks still marked as scheduled / not yet scheduled if save_task_status in (SAVE_TASK_SCHEDULED, SAVE_TASK_NOT_YET_SCHEDULED): visit_date, save_task_status = _check_visit_update_status( save_request, save_task_status) # save task may have been archived else: save_task_status = save_request.loading_task_status if save_task_status in (SAVE_TASK_SCHEDULED, SAVE_TASK_NOT_YET_SCHEDULED): visit_date, save_task_status = _check_visit_update_status( save_request, save_task_status) else: save_task_status = save_request.loading_task_status if save_request.loading_task_status != save_task_status: save_request.loading_task_status = save_task_status must_save = True if must_save: save_request.save() return {'id': save_request.id, 'visit_type': save_request.visit_type, 'origin_url': save_request.origin_url, 'save_request_date': save_request.request_date.isoformat(), 'save_request_status': save_request.status, 'save_task_status': save_task_status, 'visit_date': visit_date.isoformat() if visit_date else None} def create_save_origin_request(visit_type, origin_url): """ Create a loading task to save a software origin into the archive. This function aims to create a software origin loading task trough the use of the swh-scheduler component. First, some checks are performed to see if the visit type and origin url are valid but also if the the save request can be accepted. If those checks passed, the loading task is then created. Otherwise, the save request is put in pending or rejected state. All the submitted save requests are logged into the swh-web database to keep track of them. Args: visit_type (str): the type of visit to perform (currently only ``git`` but ``svn`` and ``hg`` will soon be available) origin_url (str): the url of the origin to save Raises: BadInputExc: the visit type or origin url is invalid ForbiddenExc: the provided origin url is blacklisted Returns: dict: A dict describing the save request with the following keys: * **visit_type**: the type of visit to perform * **origin_url**: the url of the origin * **save_request_date**: the date the request was submitted * **save_request_status**: the request status, either **accepted**, **rejected** or **pending** * **save_task_status**: the origin loading task status, either **not created**, **not yet scheduled**, **scheduled**, **succeed** or **failed** """ _check_visit_type_savable(visit_type) _check_origin_url_valid(origin_url) save_request_status = can_save_origin(origin_url) task = None # if the origin save request is accepted, create a scheduler # task to load it into the archive if save_request_status == SAVE_REQUEST_ACCEPTED: # create a task with high priority kwargs = {'priority': 'high'} # set task parameters according to the visit type if visit_type == 'git': kwargs['repo_url'] = origin_url elif visit_type == 'hg': kwargs['origin_url'] = origin_url elif visit_type == 'svn': kwargs['origin_url'] = origin_url kwargs['svn_url'] = origin_url sor = None # get list of previously sumitted save requests current_sors = \ list(SaveOriginRequest.objects.filter(visit_type=visit_type, origin_url=origin_url)) can_create_task = False # if no save requests previously submitted, create the scheduler task if not current_sors: can_create_task = True else: # get the latest submitted save request sor = current_sors[0] # if it was in pending state, we need to create the scheduler task # and update the save request info in the database if sor.status == SAVE_REQUEST_PENDING: can_create_task = True # a task has already been created to load the origin elif sor.loading_task_id != -1: # get the scheduler task and its status tasks = scheduler.get_tasks([sor.loading_task_id]) task = tasks[0] if tasks else None task_status = _save_request_dict(sor, task)['save_task_status'] # create a new scheduler task only if the previous one has been # already executed if task_status == SAVE_TASK_FAILED or \ task_status == SAVE_TASK_SUCCEED: can_create_task = True sor = None else: can_create_task = False if can_create_task: # effectively create the scheduler task task_dict = create_oneshot_task_dict( _visit_type_task[visit_type], **kwargs) task = scheduler.create_tasks([task_dict])[0] # pending save request has been accepted if sor: sor.status = SAVE_REQUEST_ACCEPTED sor.loading_task_id = task['id'] sor.save() else: sor = SaveOriginRequest.objects.create(visit_type=visit_type, origin_url=origin_url, status=save_request_status, # noqa loading_task_id=task['id']) # noqa # save request must be manually reviewed for acceptation elif save_request_status == SAVE_REQUEST_PENDING: # check if there is already such a save request already submitted, # no need to add it to the database in that case try: sor = SaveOriginRequest.objects.get(visit_type=visit_type, origin_url=origin_url, status=save_request_status) # if not add it to the database except ObjectDoesNotExist: sor = SaveOriginRequest.objects.create(visit_type=visit_type, origin_url=origin_url, status=save_request_status) # origin can not be saved as its url is blacklisted, # log the request to the database anyway else: sor = SaveOriginRequest.objects.create(visit_type=visit_type, origin_url=origin_url, status=save_request_status) if save_request_status == SAVE_REQUEST_REJECTED: raise ForbiddenExc('The origin url is blacklisted and will not be ' 'loaded into the archive.') return _save_request_dict(sor, task) def get_save_origin_requests_from_queryset(requests_queryset): """ Get all save requests from a SaveOriginRequest queryset. Args: requests_queryset (django.db.models.QuerySet): input SaveOriginRequest queryset Returns: list: A list of save origin requests dict as described in :func:`swh.web.common.origin_save.create_save_origin_request` """ task_ids = [] for sor in requests_queryset: task_ids.append(sor.loading_task_id) save_requests = [] if task_ids: tasks = scheduler.get_tasks(task_ids) tasks = {task['id']: task for task in tasks} for sor in requests_queryset: sr_dict = _save_request_dict(sor, tasks.get(sor.loading_task_id)) save_requests.append(sr_dict) return save_requests def get_save_origin_requests(visit_type, origin_url): """ Get all save requests for a given software origin. Args: visit_type (str): the type of visit origin_url (str): the url of the origin Raises: BadInputExc: the visit type or origin url is invalid NotFoundExc: no save requests can be found for the given origin Returns: list: A list of save origin requests dict as described in :func:`swh.web.common.origin_save.create_save_origin_request` """ _check_visit_type_savable(visit_type) _check_origin_url_valid(origin_url) sors = SaveOriginRequest.objects.filter(visit_type=visit_type, origin_url=origin_url) if sors.count() == 0: raise NotFoundExc(('No save requests found for visit of type ' '%s on origin with url %s.') % (visit_type, origin_url)) return get_save_origin_requests_from_queryset(sors) def get_save_origin_task_info(save_request_id): """ Get detailed information about an accepted save origin request and its associated loading task. If the associated loading task info is archived and removed from the scheduler database, returns an empty dictionary. Args: save_request_id (int): identifier of a save origin request Returns: dict: A dictionary with the following keys: + - **type**: loading task type - **arguments**: loading task arguments - **id**: loading task database identifier - **backend_id**: loading task celery identifier - **scheduled**: loading task scheduling date - **ended**: loading task termination date - **status**: loading task execution status + Depending on the availability of the task logs in the elasticsearch cluster of Software Heritage, the returned dictionary may also contain the following keys: + - **name**: associated celery task name - **message**: relevant log message from task execution - **duration**: task execution time (only if it succeeded) - **worker**: name of the worker that executed the task """ try: save_request = SaveOriginRequest.objects.get(id=save_request_id) except ObjectDoesNotExist: return {} task = scheduler.get_tasks([save_request.loading_task_id]) task = task[0] if task else None if task is None: return {} task_run = scheduler.get_task_runs([task['id']]) task_run = task_run[0] if task_run else None if task_run is None: return {} task_run['type'] = task['type'] task_run['arguments'] = task['arguments'] task_run['id'] = task_run['task'] del task_run['task'] del task_run['metadata'] del task_run['started'] es_workers_index_url = config.get_config()['es_workers_index_url'] if not es_workers_index_url: return task_run es_workers_index_url += '/_search' if save_request.visit_date: min_ts = save_request.visit_date max_ts = min_ts + timedelta(days=7) else: min_ts = save_request.request_date max_ts = min_ts + timedelta(days=30) min_ts = int(min_ts.timestamp()) * 1000 max_ts = int(max_ts.timestamp()) * 1000 save_task_status = _save_task_status[task['status']] priority = '3' if save_task_status == SAVE_TASK_FAILED else '6' query = { 'bool': { 'must': [ { 'match_phrase': { 'priority': { 'query': priority } } }, { 'match_phrase': { 'swh_task_id': { 'query': task_run['backend_id'] } } }, { 'range': { '@timestamp': { 'gte': min_ts, 'lte': max_ts, 'format': 'epoch_millis' } } } ] } } try: response = requests.post(es_workers_index_url, json={'query': query, 'sort': ['@timestamp']}, timeout=30) results = json.loads(response.text) if results['hits']['total'] >= 1: task_run_info = results['hits']['hits'][-1]['_source'] if 'swh_logging_args_runtime' in task_run_info: duration = task_run_info['swh_logging_args_runtime'] task_run['duration'] = duration if 'message' in task_run_info: task_run['message'] = task_run_info['message'] if 'swh_logging_args_name' in task_run_info: task_run['name'] = task_run_info['swh_logging_args_name'] elif 'swh_task_name' in task_run_info: task_run['name'] = task_run_info['swh_task_name'] if 'hostname' in task_run_info: task_run['worker'] = task_run_info['hostname'] elif 'host' in task_run_info: task_run['worker'] = task_run_info['host'] except Exception as e: logger.warning('Request to Elasticsearch failed\n%s' % str(e)) pass return task_run