ForgeImporters/forgeimporters/base.py (365 lines of code) (raw):

# Licensed to the Apache Software Foundation (ASF) under one # or more contributor license agreements. See the NOTICE file # distributed with this work for additional information # regarding copyright ownership. The ASF licenses this file # to you under the Apache License, Version 2.0 (the # "License"); you may not use this file except in compliance # with the License. You may obtain a copy of the License at # # http://www.apache.org/licenses/LICENSE-2.0 # # Unless required by applicable law or agreed to in writing, # software distributed under the License is distributed on an # "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY # KIND, either express or implied. See the License for the # specific language governing permissions and limitations # under the License. import os import errno import logging from io import BytesIO import urllib.request import urllib.parse from collections import defaultdict import traceback from urllib.parse import urlparse from urllib.parse import unquote from datetime import datetime from bs4 import BeautifulSoup from tg import expose, validate, flash, redirect, config from tg.decorators import with_trailing_slash from tg import app_globals as g from tg import tmpl_context as c from tg import request from formencode import validators as fev, schema from webob import exc from allura.lib.decorators import require_post from allura.lib.decorators import task from allura.lib.security import require_access from allura.lib.plugin import ProjectRegistrationProvider, AdminExtension from allura.lib.utils import guess_mime_type from allura.lib import helpers as h from allura.lib import exceptions from allura.lib import validators as v from allura.app import SitemapEntry from allura import model as M from paste.deploy.converters import aslist from ming.utils import LazyProperty from allura.controllers import BaseController log = logging.getLogger(__name__) class ProjectImportForm(schema.Schema): def __init__(self, source): super().__init__() provider = ProjectRegistrationProvider.get() self.add_field('tools', ToolsValidator(source)) self.add_field('project_shortname', provider.shortname_validator) self.allow_extra_fields = True neighborhood = fev.NotEmpty() project_name = v.UnicodeString(not_empty=True, max=40) class ToolImportForm(schema.Schema): def __init__(self, tool_class): super().__init__() self.add_field('mount_point', v.MountPointValidator(tool_class)) mount_label = v.UnicodeString() class ImportErrorHandler: def __init__(self, importer, project_name, project): self.importer = importer self.project_name = project_name self.project = project def __enter__(self): return self def __exit__(self, exc_type, exc_val, exc_tb): if hasattr(self.importer, 'clear_pending'): self.importer.clear_pending(self.project) if exc_type: g.post_event('import_tool_task_failed', error=str(exc_val), traceback=traceback.format_exc(), importer_source=self.importer.source, importer_tool_label=self.importer.tool_label, project_name=self.project_name, ) def success(self, app): with h.push_config(c, project=self.project, app=app): g.post_event('import_tool_task_succeeded', self.importer.source, self.importer.tool_label, ) def object_from_path(path): """Given a dotted path, import and return the object at that path. """ module_name, obj_name = path.rsplit('.', 1) module = __import__(module_name, fromlist=[obj_name]) return getattr(module, obj_name) @task(notifications_disabled=True) def import_tool(importer_path, project_name=None, mount_point=None, mount_label=None, **kw): importer = object_from_path(importer_path)() with ImportErrorHandler(importer, project_name, c.project) as handler, \ M.session.substitute_extensions(M.artifact_orm_session, [M.session.BatchIndexer]): try: M.artifact_orm_session._get().skip_last_updated = True app = importer.import_tool( c.project, c.user, project_name=project_name, mount_point=mount_point, mount_label=mount_label, **kw) # manually update project's last_updated field at the end of the # import instead of it being updated automatically by each artifact # since long-running task can cause stale project data to be saved M.Project.query.update( {'_id': c.project._id}, {'$set': {'last_updated': datetime.utcnow()}}) finally: M.artifact_orm_session._get().skip_last_updated = False M.artifact_orm_session.flush() M.session.BatchIndexer.flush() if app: with h.notifications_disabled(c.project, disabled=False): g.director.create_activity(c.user, "imported", app.config, related_nodes=[c.project], tags=['import']) handler.success(app) class ProjectExtractor: """Base class for project extractors. Subclasses should use :meth:`urlopen` to make HTTP requests, as it provides a custom User-Agent and automatically retries timed-out requests. """ PAGE_MAP = {} def __init__(self, project_name, page_name_or_url=None, **kw): self.project_name = project_name self._page_cache = {} self.url = None self.page = None if page_name_or_url: self.get_page(page_name_or_url, **kw) @staticmethod def urlopen(url, retries=3, codes=(408, 500, 502, 503, 504), timeout=120, unredirected_hdrs=None, **kw): req = urllib.request.Request(url, **kw) # noqa: S310 if unredirected_hdrs: for key, val in unredirected_hdrs.items(): req.add_unredirected_header(key, val) req.add_header('User-Agent', 'Allura Data Importer (https://allura.apache.org/)') return h.urlopen(req, retries=retries, codes=codes, timeout=timeout) def get_page(self, page_name_or_url, parser=None, **kw): """Return a Beautiful soup object for the given page name or url. If a page name is provided, the associated url is looked up in :attr:`PAGE_MAP`. If provided, the class or callable passed in :param:`parser` will be used to transform the result of the `urlopen` before returning it. Otherwise, the class's :meth:`parse_page` will be used. Results are cached so that subsequent calls for the same page name or url will return the cached result rather than making another HTTP request. """ if page_name_or_url in self.PAGE_MAP: self.url = self.get_page_url(page_name_or_url, **kw) else: self.url = page_name_or_url if self.url in self._page_cache: self.page = self._page_cache[self.url] else: if parser is None: parser = self.parse_page self.page = self._page_cache[self.url] = \ parser(self.urlopen(self.url)) return self.page def get_page_url(self, page_name, **kw): """Return the url associated with ``page_name``. Raises KeyError if ``page_name`` is not in :attr:`PAGE_MAP`. """ return self.PAGE_MAP[page_name].format( project_name=urllib.parse.quote(self.project_name), **kw) def parse_page(self, page): """Transforms the result of a `urlopen` call before returning it from :meth:`get_page`. The default implementation create a :class:`BeautifulSoup` object from the html. Subclasses can override to change the behavior or handle other types of content (like JSON). The parser can also be overridden via the `parser` parameter to :meth:`get_page` :param page: A file-like object return from :meth:`urlopen` """ return BeautifulSoup(page, convertEntities=BeautifulSoup.HTML_ENTITIES) class ProjectImporter(BaseController): """ Base class for project importers. Subclasses are required to implement the :meth:`index()` and :meth:`process()` views described below. """ source = None tool_label = 'Project Info' process_validator = None index_template = None def __init__(self, neighborhood, *a, **kw): self.neighborhood = neighborhood def _check_security(self): with h.login_overlay(exceptions=['process']): require_access(self.neighborhood, 'register') @LazyProperty def tool_importers(self): """ List of all tool importers that import from the same source as this project importer. """ tools = {} for ep in h.iter_entry_points('allura.importers'): epv = ep.load() if epv.source == self.source: tools[ep.name] = epv() return tools @with_trailing_slash @expose() def index(self, **kw): """ Override and expose this view to present the project import form. The template used by this view should extend the base template in:: jinja:forgeimporters:templates/project_base.html This will list the available tool importers. Other project fields (e.g., project_name) should go in the project_fields block. """ return {'importer': self, 'tg_template': self.index_template} @require_post() @expose() @validate(process_validator, error_handler=index) def process(self, **kw): """ Override and expose this to handle a project import. This should at a minimum create the stub project with the appropriate tools installed and redirect to the new project, presumably with a message indicating that some data will not be available immediately. """ try: with h.push_config(config, **{'project.verify_phone': 'false'}): c.project = self.neighborhood.register_project( kw['project_shortname'], project_name=kw['project_name']) except exceptions.ProjectOverlimitError: flash("You have exceeded the maximum number of projects you are allowed to create", 'error') redirect('.') except exceptions.ProjectRatelimitError: flash("Project creation rate limit exceeded. Please try again later.", 'error') redirect('.') except Exception: log.error('error registering project: %s', kw['project_shortname'], exc_info=True) flash('Internal Error. Please try again later.', 'error') redirect('.') self.after_project_create(c.project, **kw) tools = aslist(kw.get('tools')) for importer_name in tools: ToolImporter.by_name(importer_name).post(**kw) M.AuditLog.log('import project from %s' % self.source) flash('Welcome to your new project on %s! ' 'Your project data will be imported and should show up here shortly.' % config['site_name']) redirect(c.project.script_name + 'admin/overview') @expose('json:') @validate(process_validator) def check_names(self, **kw): """ Ajax form validation. """ return request.validation.errors def after_project_create(self, project, **kw): """ Called after project is created. Useful for doing extra processing on the project before individual tool imports happen. :param project: The newly created project. :param kw: The keyword arguments that were posted to the controller method that created the project. """ pass class ToolImportControllerMeta(type): def __call__(cls, importer, *args, **kw): """ Decorate the `create` post handler with a validator that references the appropriate App for this controller's importer. """ if hasattr(cls, 'create') and not getattr(cls.create.decoration, 'validations', None): index_meth = getattr(cls.index, '__func__', cls.index) cls.create = validate(cls.import_form(aslist(importer.target_app)[0]), error_handler=index_meth)(cls.create) return type.__call__(cls, importer, *args, **kw) class ToolImportController(BaseController, metaclass=ToolImportControllerMeta): """ Base class for ToolImporter controllers. """ def __init__(self, importer): """ :param importer: :class:`ToolImporter` instance to which this controller belongs. """ self.importer = importer @property def target_app(self): return aslist(self.importer.target_app)[0] class ToolImporterMeta(type): def __init__(cls, name, bases, attrs): if not (hasattr(cls, 'target_app_ep_names') or hasattr(cls, 'target_app')): raise AttributeError(f"{name} must define either `target_app` or `target_app_ep_names`") type.__init__(cls, name, bases, attrs) def __call__(cls, *args, **kw): """ Right before the first instance of cls is created, get the list of target_app classes from ep names. Can't do this at cls create/init time b/c g.entry_points is not guaranteed to be loaded at that point. """ if not getattr(cls, 'target_app', None): cls.target_app = [g.entry_points['tool'][ep_name] for ep_name in aslist(cls.target_app_ep_names) if ep_name in g.entry_points['tool']] return type.__call__(cls, *args, **kw) class ToolImporter(metaclass=ToolImporterMeta): """ Base class for tool importers. Subclasses are required to implement :meth:`import_tool()` described below and define the following attributes: .. py:attribute:: target_app_ep_names A string or list of strings which are entry point names of the tool(s) to which this class imports. E.g.:: target_app_ep_names = ['git', 'hg'] .. py:attribute:: target_app A reference or list of references to the tool(s) that this imports to. This attribute is not required if `target_app_ep_names` is defined (which is preferable). E.g.:: target_app = [forgegit.ForgeGitApp, forgehg.ForgeHgApp] .. py:attribute:: source A string indicating where this imports from. This must match the `source` value of the :class:`ProjectImporter` for this importer to be discovered during full-project imports. E.g.:: source = 'GitHub' .. py:attribute:: controller The controller for this importer, to handle single tool imports. """ target_app = None # app or list of apps source = None # string description of source, must match project importer controller = None @staticmethod def by_name(name): """ Return a ToolImporter subclass instance given its entry-point name. """ for ep in h.iter_entry_points('allura.importers', name): return ep.load()() @staticmethod def by_app(app): """ Return a ToolImporter subclass instance given its target_app class. """ importers = {} for ep in h.iter_entry_points('allura.importers'): importer = ep.load() if app in aslist(importer.target_app): importers[ep.name] = importer() return importers @property def classname(self): return self.__class__.__name__ def enforce_limit(self, project): """ Enforce rate limiting of tool imports on a given project. Returns False if limit is met / exceeded. Otherwise, increments the count of pending / in-progress imports and returns True. """ limit = config.get('tool_import.rate_limit', 1) pending_key = 'tool_data.%s.pending' % self.classname modified_project = M.Project.query.find_one_and_update( { '_id': project._id, '$or': [ {pending_key: None}, {pending_key: {'$lt': limit}}, ], }, update={'$inc': {pending_key: 1}}, return_document=True, ) return modified_project is not None def clear_pending(self, project): """ Decrement the pending counter for this importer on the given project, to indicate that an import is complete. """ pending_key = 'tool_data.%s.pending' % self.classname M.Project.query.find_one_and_update( {'_id': project._id}, update={'$inc': {pending_key: -1}}, return_document=True, ) def import_tool(self, project, user, project_name=None, mount_point=None, mount_label=None, **kw): """ Override this method to perform the tool import. :param project: the Allura project to import to :param project_name: the name of the remote project to import from :param mount_point: the mount point name, to override the default :param mount_label: the mount label name, to override the default """ raise NotImplementedError @property def tool_label(self): """ The label for this tool importer. Defaults to the `tool_label` from the `target_app`. """ return getattr(aslist(self.target_app)[0], 'tool_label', None) @property def tool_option(self): """ The option for this tool importer. Defaults to the `tool_option` from the `target_app`. """ return getattr(aslist(self.target_app)[0], 'tool_option', dict()) @property def tool_description(self): """ The description for this tool importer. Defaults to the `tool_description` from the `target_app`. """ return getattr(aslist(self.target_app)[0], 'tool_description', None) def tool_icon(self, theme, size): return theme.app_icon_url(aslist(self.target_app)[0], size) def post(self, **kw): """Post a task that will call ``import_tool()`` on this instance. """ klass = self.__class__ importer_path = f'{klass.__module__}.{klass.__name__}' import_tool.post(importer_path, **kw) class ToolsValidator(fev.Set): """ Validates the list of tool importers during a project import. This verifies that the tools selected are available and valid for this source. """ def __init__(self, source, *a, **kw): super().__init__(*a, **kw) self.source = source def to_python(self, value, state=None): value = super().to_python(value, state) valid = [] invalid = [] for name in value: importer = ToolImporter.by_name(name) if importer is not None and importer.source == self.source: valid.append(name) else: invalid.append(name) if invalid: pl = 's' if len(invalid) > 1 else '' raise fev.Invalid('Invalid tool%s selected: %s' % (pl, ', '.join(invalid)), value, state) return valid class ProjectToolsImportController: '''List all importers available''' @with_trailing_slash @expose('jinja:forgeimporters:templates/list_all.html') def index(self, *a, **kw): importer_matrix = defaultdict(dict) tools_with_importers = set() hidden = set(aslist(config.get('hidden_importers'), sep=',')) def visible(ep): return ep.name not in hidden for ep in filter(visible, h.iter_entry_points('allura.importers')): # must instantiate to ensure importer.target_app is populated # (see ToolImporterMeta.__call__) importer = ep.load()() for tool in aslist(importer.target_app): tools_with_importers.add(tool.tool_label) importer_matrix[importer.source][tool.tool_label] = ep.name return { 'importer_matrix': importer_matrix, 'tools': tools_with_importers, } @expose() def _lookup(self, name, *remainder): importer = ToolImporter.by_name(name) if importer: return importer.controller(importer), remainder else: raise exc.HTTPNotFound class ImportAdminExtension(AdminExtension): '''Add import link to project admin sidebar''' project_admin_controllers = {'import': ProjectToolsImportController} def update_project_sidebar_menu(self, sidebar_links): base_url = c.project.url() + 'admin/ext/' link = SitemapEntry('Import', base_url + 'import/') sidebar_links.append(link) return sidebar_links def bytesio_parser(page): return { 'content-type': page.info()['content-type'], 'data': BytesIO(page.read()), } class File: def __init__(self, url, filename=None): extractor = ProjectExtractor(None, url, parser=bytesio_parser) self.url = url self.filename = filename or unquote(os.path.basename(urlparse(url).path)) # try to get the mime-type from the filename first, because # some files (e.g., attachements) may have the Content-Type header # forced to encourage the UA to download / save the file self.type = guess_mime_type(self.filename) if self.type == 'application/octet-stream': # however, if that fails, fall back to the given mime-type, # as some files (e.g., project icons) might have no file # extension but return a valid Content-Type header self.type = extractor.page['content-type'] self.file = extractor.page['data'] def get_importer_upload_path(project): shortname = project.shortname if project.is_nbhd_project: shortname = project.url().strip('/') elif project.is_user_project: shortname = project.shortname.split('/')[1] elif not project.is_root: shortname = project.shortname.split('/')[0] upload_path = config['importer_upload_path'].format( nbhd=project.neighborhood.url_prefix.strip('/'), project=shortname, c=c, ) return upload_path def save_importer_upload(project, filename, data): dest_path = get_importer_upload_path(project) dest_file = os.path.join(dest_path, filename) try: os.makedirs(dest_path) except OSError as e: if e.errno != errno.EEXIST: raise with open(dest_file, 'w', encoding='utf-8') as fp: fp.write(data) return dest_file