from __future__ import absolute_import from collections import defaultdict from itertools import chain import logging import os from pip._vendor import pkg_resources from pip._vendor import requests from pip.compat import expanduser from pip.download import (is_file_url, is_dir_url, is_vcs_url, url_to_path, unpack_url) from pip.exceptions import (InstallationError, BestVersionAlreadyInstalled, DistributionNotFound, PreviousBuildDirError, HashError, HashErrors, HashUnpinned, DirectoryUrlHashUnsupported, VcsHashUnsupported) from pip.req.req_install import InstallRequirement from pip.utils import ( display_path, dist_in_usersite, ensure_dir, normalize_path) from pip.utils.hashes import MissingHashes from pip.utils.logging import indent_log from pip.vcs import vcs from pip.wheel import Wheel logger = logging.getLogger(__name__) class Requirements(object): def __init__(self): self._keys = [] self._dict = {} def keys(self): return self._keys def values(self): return [self._dict[key] for key in self._keys] def __contains__(self, item): return item in self._keys def __setitem__(self, key, value): if key not in self._keys: self._keys.append(key) self._dict[key] = value def __getitem__(self, key): return self._dict[key] def __repr__(self): values = ['%s: %s' % (repr(k), repr(self[k])) for k in self.keys()] return 'Requirements({%s})' % ', '.join(values) class DistAbstraction(object): """Abstracts out the wheel vs non-wheel prepare_files logic. The requirements for anything installable are as follows: - we must be able to determine the requirement name (or we can't correctly handle the non-upgrade case). - we must be able to generate a list of run-time dependencies without installing any additional packages (or we would have to either burn time by doing temporary isolated installs or alternatively violate pips 'don't start installing unless all requirements are available' rule - neither of which are desirable). - for packages with setup requirements, we must also be able to determine their requirements without installing additional packages (for the same reason as run-time dependencies) - we must be able to create a Distribution object exposing the above metadata. """ def __init__(self, req_to_install): self.req_to_install = req_to_install def dist(self, finder): """Return a setuptools Dist object.""" raise NotImplementedError(self.dist) def prep_for_dist(self): """Ensure that we can get a Dist for this requirement.""" raise NotImplementedError(self.dist) def make_abstract_dist(req_to_install): """Factory to make an abstract dist object. Preconditions: Either an editable req with a source_dir, or satisfied_by or a wheel link, or a non-editable req with a source_dir. :return: A concrete DistAbstraction. """ if req_to_install.editable: return IsSDist(req_to_install) elif req_to_install.link and req_to_install.link.is_wheel: return IsWheel(req_to_install) else: return IsSDist(req_to_install) class IsWheel(DistAbstraction): def dist(self, finder): return list(pkg_resources.find_distributions( self.req_to_install.source_dir))[0] def prep_for_dist(self): # FIXME:https://github.com/pypa/pip/issues/1112 pass class IsSDist(DistAbstraction): def dist(self, finder): dist = self.req_to_install.get_dist() # FIXME: shouldn't be globally added: if dist.has_metadata('dependency_links.txt'): finder.add_dependency_links( dist.get_metadata_lines('dependency_links.txt') ) return dist def prep_for_dist(self): self.req_to_install.run_egg_info() self.req_to_install.assert_source_matches_version() class Installed(DistAbstraction): def dist(self, finder): return self.req_to_install.satisfied_by def prep_for_dist(self): pass class RequirementSet(object): def __init__(self, build_dir, src_dir, download_dir, upgrade=False, ignore_installed=False, as_egg=False, target_dir=None, ignore_dependencies=False, force_reinstall=False, use_user_site=False, session=None, pycompile=True, isolated=False, wheel_download_dir=None, wheel_cache=None, require_hashes=False): """Create a RequirementSet. :param wheel_download_dir: Where still-packed .whl files should be written to. If None they are written to the download_dir parameter. Separate to download_dir to permit only keeping wheel archives for pip wheel. :param download_dir: Where still packed archives should be written to. If None they are not saved, and are deleted immediately after unpacking. :param wheel_cache: The pip wheel cache, for passing to InstallRequirement. """ if session is None: raise TypeError( "RequirementSet() missing 1 required keyword argument: " "'session'" ) self.build_dir = build_dir self.src_dir = src_dir # XXX: download_dir and wheel_download_dir overlap semantically and may # be combined if we're willing to have non-wheel archives present in # the wheelhouse output by 'pip wheel'. self.download_dir = download_dir self.upgrade = upgrade self.ignore_installed = ignore_installed self.force_reinstall = force_reinstall self.requirements = Requirements() # Mapping of alias: real_name self.requirement_aliases = {} self.unnamed_requirements = [] self.ignore_dependencies = ignore_dependencies self.successfully_downloaded = [] self.successfully_installed = [] self.reqs_to_cleanup = [] self.as_egg = as_egg self.use_user_site = use_user_site self.target_dir = target_dir # set from --target option self.session = session self.pycompile = pycompile self.isolated = isolated if wheel_download_dir: wheel_download_dir = normalize_path(wheel_download_dir) self.wheel_download_dir = wheel_download_dir self._wheel_cache = wheel_cache self.require_hashes = require_hashes # Maps from install_req -> dependencies_of_install_req self._dependencies = defaultdict(list) def __str__(self): reqs = [req for req in self.requirements.values() if not req.comes_from] reqs.sort(key=lambda req: req.name.lower()) return ' '.join([str(req.req) for req in reqs]) def __repr__(self): reqs = [req for req in self.requirements.values()] reqs.sort(key=lambda req: req.name.lower()) reqs_str = ', '.join([str(req.req) for req in reqs]) return ('<%s object; %d requirement(s): %s>' % (self.__class__.__name__, len(reqs), reqs_str)) def add_requirement(self, install_req, parent_req_name=None): """Add install_req as a requirement to install. :param parent_req_name: The name of the requirement that needed this added. The name is used because when multiple unnamed requirements resolve to the same name, we could otherwise end up with dependency links that point outside the Requirements set. parent_req must already be added. Note that None implies that this is a user supplied requirement, vs an inferred one. :return: Additional requirements to scan. That is either [] if the requirement is not applicable, or [install_req] if the requirement is applicable and has just been added. """ name = install_req.name if not install_req.match_markers(): logger.warning("Ignoring %s: markers %r don't match your " "environment", install_req.name, install_req.markers) return [] # This check has to come after we filter requirements with the # environment markers. if install_req.link and install_req.link.is_wheel: wheel = Wheel(install_req.link.filename) if not wheel.supported(): raise InstallationError( "%s is not a supported wheel on this platform." % wheel.filename ) install_req.as_egg = self.as_egg install_req.use_user_site = self.use_user_site install_req.target_dir = self.target_dir install_req.pycompile = self.pycompile if not name: # url or path requirement w/o an egg fragment self.unnamed_requirements.append(install_req) return [install_req] else: try: existing_req = self.get_requirement(name) except KeyError: existing_req = None if (parent_req_name is None and existing_req and not existing_req.constraint and existing_req.extras == install_req.extras and not existing_req.req.specifier == install_req.req.specifier): raise InstallationError( 'Double requirement given: %s (already in %s, name=%r)' % (install_req, existing_req, name)) if not existing_req: # Add requirement self.requirements[name] = install_req # FIXME: what about other normalizations? E.g., _ vs. -? if name.lower() != name: self.requirement_aliases[name.lower()] = name result = [install_req] else: # Assume there's no need to scan, and that we've already # encountered this for scanning. result = [] if not install_req.constraint and existing_req.constraint: if (install_req.link and not (existing_req.link and install_req.link.path == existing_req.link.path)): self.reqs_to_cleanup.append(install_req) raise InstallationError( "Could not satisfy constraints for '%s': " "installation from path or url cannot be " "constrained to a version" % name) # If we're now installing a constraint, mark the existing # object for real installation. existing_req.constraint = False existing_req.extras = tuple( sorted(set(existing_req.extras).union( set(install_req.extras)))) logger.debug("Setting %s extras to: %s", existing_req, existing_req.extras) # And now we need to scan this. result = [existing_req] # Canonicalise to the already-added object for the backref # check below. install_req = existing_req if parent_req_name: parent_req = self.get_requirement(parent_req_name) self._dependencies[parent_req].append(install_req) return result def has_requirement(self, project_name): name = project_name.lower() if (name in self.requirements and not self.requirements[name].constraint or name in self.requirement_aliases and not self.requirements[self.requirement_aliases[name]].constraint): return True return False @property def has_requirements(self): return list(req for req in self.requirements.values() if not req.constraint) or self.unnamed_requirements @property def is_download(self): if self.download_dir: self.download_dir = expanduser(self.download_dir) if os.path.exists(self.download_dir): return True else: logger.critical('Could not find download directory') raise InstallationError( "Could not find or access download directory '%s'" % display_path(self.download_dir)) return False def get_requirement(self, project_name): for name in project_name, project_name.lower(): if name in self.requirements: return self.requirements[name] if name in self.requirement_aliases: return self.requirements[self.requirement_aliases[name]] raise KeyError("No project with the name %r" % project_name) def uninstall(self, auto_confirm=False): for req in self.requirements.values(): if req.constraint: continue req.uninstall(auto_confirm=auto_confirm) req.commit_uninstall() def prepare_files(self, finder): """ Prepare process. Create temp directories, download and/or unpack files. """ # make the wheelhouse if self.wheel_download_dir: ensure_dir(self.wheel_download_dir) # If any top-level requirement has a hash specified, enter # hash-checking mode, which requires hashes from all. root_reqs = self.unnamed_requirements + self.requirements.values() require_hashes = (self.require_hashes or any(req.has_hash_options for req in root_reqs)) if require_hashes and self.as_egg: raise InstallationError( '--egg is not allowed with --require-hashes mode, since it ' 'delegates dependency resolution to setuptools and could thus ' 'result in installation of unhashed packages.') # Actually prepare the files, and collect any exceptions. Most hash # exceptions cannot be checked ahead of time, because # req.populate_link() needs to be called before we can make decisions # based on link type. discovered_reqs = [] hash_errors = HashErrors() for req in chain(root_reqs, discovered_reqs): try: discovered_reqs.extend(self._prepare_file( finder, req, require_hashes=require_hashes, ignore_dependencies=self.ignore_dependencies)) except HashError as exc: exc.req = req hash_errors.append(exc) if hash_errors: raise hash_errors def _check_skip_installed(self, req_to_install, finder): """Check if req_to_install should be skipped. This will check if the req is installed, and whether we should upgrade or reinstall it, taking into account all the relevant user options. After calling this req_to_install will only have satisfied_by set to None if the req_to_install is to be upgraded/reinstalled etc. Any other value will be a dist recording the current thing installed that satisfies the requirement. Note that for vcs urls and the like we can't assess skipping in this routine - we simply identify that we need to pull the thing down, then later on it is pulled down and introspected to assess upgrade/ reinstalls etc. :return: A text reason for why it was skipped, or None. """ # Check whether to upgrade/reinstall this req or not. req_to_install.check_if_exists() if req_to_install.satisfied_by: skip_reason = 'satisfied (use --upgrade to upgrade)' if self.upgrade: best_installed = False # For link based requirements we have to pull the # tree down and inspect to assess the version #, so # its handled way down. if not (self.force_reinstall or req_to_install.link): try: finder.find_requirement(req_to_install, self.upgrade) except BestVersionAlreadyInstalled: skip_reason = 'up-to-date' best_installed = True except DistributionNotFound: # No distribution found, so we squash the # error - it will be raised later when we # re-try later to do the install. # Why don't we just raise here? pass if not best_installed: # don't uninstall conflict if user install and # conflict is not user install if not (self.use_user_site and not dist_in_usersite(req_to_install.satisfied_by)): req_to_install.conflicts_with = \ req_to_install.satisfied_by req_to_install.satisfied_by = None return skip_reason else: return None def _prepare_file(self, finder, req_to_install, require_hashes=False, ignore_dependencies=False): """Prepare a single requirements file. :return: A list of additional InstallRequirements to also install. """ # Tell user what we are doing for this requirement: # obtain (editable), skipping, processing (local url), collecting # (remote url or package name) if req_to_install.constraint or req_to_install.prepared: return [] req_to_install.prepared = True # ###################### # # # print log messages # # # ###################### # if req_to_install.editable: logger.info('Obtaining %s', req_to_install) else: # satisfied_by is only evaluated by calling _check_skip_installed, # so it must be None here. assert req_to_install.satisfied_by is None if not self.ignore_installed: skip_reason = self._check_skip_installed( req_to_install, finder) if req_to_install.satisfied_by: assert skip_reason is not None, ( '_check_skip_installed returned None but ' 'req_to_install.satisfied_by is set to %r' % (req_to_install.satisfied_by,)) logger.info( 'Requirement already %s: %s', skip_reason, req_to_install) else: if (req_to_install.link and req_to_install.link.scheme == 'file'): path = url_to_path(req_to_install.link.url) logger.info('Processing %s', display_path(path)) else: logger.info('Collecting %s', req_to_install) with indent_log(): # ################################ # # # vcs update or unpack archive # # # ################################ # if req_to_install.editable: if require_hashes: raise InstallationError( 'The editable requirement %s cannot be installed when ' 'requiring hashes, because there is no single file to ' 'hash.' % req_to_install) req_to_install.ensure_has_source_dir(self.src_dir) req_to_install.update_editable(not self.is_download) abstract_dist = make_abstract_dist(req_to_install) abstract_dist.prep_for_dist() if self.is_download: req_to_install.archive(self.download_dir) elif req_to_install.satisfied_by: if require_hashes: logger.debug( 'Since it is already installed, we are trusting this ' 'package without checking its hash. To ensure a ' 'completely repeatable environment, install into an ' 'empty virtualenv.') abstract_dist = Installed(req_to_install) else: # @@ if filesystem packages are not marked # editable in a req, a non deterministic error # occurs when the script attempts to unpack the # build directory req_to_install.ensure_has_source_dir(self.build_dir) # If a checkout exists, it's unwise to keep going. version # inconsistencies are logged later, but do not fail the # installation. # FIXME: this won't upgrade when there's an existing # package unpacked in `req_to_install.source_dir` if os.path.exists( os.path.join(req_to_install.source_dir, 'setup.py')): raise PreviousBuildDirError( "pip can't proceed with requirements '%s' due to a" " pre-existing build directory (%s). This is " "likely due to a previous installation that failed" ". pip is being responsible and not assuming it " "can delete this. Please delete it and try again." % (req_to_install, req_to_install.source_dir) ) req_to_install.populate_link( finder, self.upgrade, require_hashes) # We can't hit this spot and have populate_link return None. # req_to_install.satisfied_by is None here (because we're # guarded) and upgrade has no impact except when satisfied_by # is not None. # Then inside find_requirement existing_applicable -> False # If no new versions are found, DistributionNotFound is raised, # otherwise a result is guaranteed. assert req_to_install.link link = req_to_install.link # Now that we have the real link, we can tell what kind of # requirements we have and raise some more informative errors # than otherwise. (For example, we can raise VcsHashUnsupported # for a VCS URL rather than HashMissing.) if require_hashes: # We could check these first 2 conditions inside # unpack_url and save repetition of conditions, but then # we would report less-useful error messages for # unhashable requirements, complaining that there's no # hash provided. if is_vcs_url(link): raise VcsHashUnsupported() elif is_file_url(link) and is_dir_url(link): raise DirectoryUrlHashUnsupported() if (not req_to_install.original_link and not req_to_install.is_pinned): # Unpinned packages are asking for trouble when a new # version is uploaded. This isn't a security check, but # it saves users a surprising hash mismatch in the # future. # # file:/// URLs aren't pinnable, so don't complain # about them not being pinned. raise HashUnpinned() hashes = req_to_install.hashes( trust_internet=not require_hashes) if require_hashes and not hashes: # Known-good hashes are missing for this requirement, so # shim it with a facade object that will provoke hash # computation and then raise a HashMissing exception # showing the user what the hash should be. hashes = MissingHashes() try: download_dir = self.download_dir # We always delete unpacked sdists after pip ran. autodelete_unpacked = True if req_to_install.link.is_wheel \ and self.wheel_download_dir: # when doing 'pip wheel` we download wheels to a # dedicated dir. download_dir = self.wheel_download_dir if req_to_install.link.is_wheel: if download_dir: # When downloading, we only unpack wheels to get # metadata. autodelete_unpacked = True else: # When installing a wheel, we use the unpacked # wheel. autodelete_unpacked = False unpack_url( req_to_install.link, req_to_install.source_dir, download_dir, autodelete_unpacked, session=self.session, hashes=hashes) except requests.HTTPError as exc: logger.critical( 'Could not install requirement %s because ' 'of error %s', req_to_install, exc, ) raise InstallationError( 'Could not install requirement %s because ' 'of HTTP error %s for URL %s' % (req_to_install, exc, req_to_install.link) ) abstract_dist = make_abstract_dist(req_to_install) abstract_dist.prep_for_dist() if self.is_download: # Make a .zip of the source_dir we already created. if req_to_install.link.scheme in vcs.all_schemes: req_to_install.archive(self.download_dir) # req_to_install.req is only avail after unpack for URL # pkgs repeat check_if_exists to uninstall-on-upgrade # (#14) if not self.ignore_installed: req_to_install.check_if_exists() if req_to_install.satisfied_by: if self.upgrade or self.ignore_installed: # don't uninstall conflict if user install and # conflict is not user install if not (self.use_user_site and not dist_in_usersite( req_to_install.satisfied_by)): req_to_install.conflicts_with = \ req_to_install.satisfied_by req_to_install.satisfied_by = None else: logger.info( 'Requirement already satisfied (use ' '--upgrade to upgrade): %s', req_to_install, ) # ###################### # # # parse dependencies # # # ###################### # dist = abstract_dist.dist(finder) more_reqs = [] def add_req(subreq): sub_install_req = InstallRequirement( str(subreq), req_to_install, isolated=self.isolated, wheel_cache=self._wheel_cache, ) more_reqs.extend(self.add_requirement( sub_install_req, req_to_install.name)) # We add req_to_install before its dependencies, so that we # can refer to it when adding dependencies. if not self.has_requirement(req_to_install.name): # 'unnamed' requirements will get added here self.add_requirement(req_to_install, None) if not ignore_dependencies: if (req_to_install.extras): logger.debug( "Installing extra requirements: %r", ','.join(req_to_install.extras), ) missing_requested = sorted( set(req_to_install.extras) - set(dist.extras) ) for missing in missing_requested: logger.warning( '%s does not provide the extra \'%s\'', dist, missing ) available_requested = sorted( set(dist.extras) & set(req_to_install.extras) ) for subreq in dist.requires(available_requested): add_req(subreq) # cleanup tmp src self.reqs_to_cleanup.append(req_to_install) if not req_to_install.editable and not req_to_install.satisfied_by: # XXX: --no-install leads this to report 'Successfully # downloaded' for only non-editable reqs, even though we took # action on them. self.successfully_downloaded.append(req_to_install) return more_reqs def cleanup_files(self): """Clean up files, remove builds.""" logger.debug('Cleaning up...') with indent_log(): for req in self.reqs_to_cleanup: req.remove_temporary_source() def _to_install(self): """Create the installation order. The installation order is topological - requirements are installed before the requiring thing. We break cycles at an arbitrary point, and make no other guarantees. """ # The current implementation, which we may change at any point # installs the user specified things in the order given, except when # dependencies must come earlier to achieve topological order. order = [] ordered_reqs = set() def schedule(req): if req.satisfied_by or req in ordered_reqs: return if req.constraint: return ordered_reqs.add(req) for dep in self._dependencies[req]: schedule(dep) order.append(req) for install_req in self.requirements.values(): schedule(install_req) return order def install(self, install_options, global_options=(), *args, **kwargs): """ Install everything in this set (after having downloaded and unpacked the packages) """ to_install = self._to_install() if to_install: logger.info( 'Installing collected packages: %s', ', '.join([req.name for req in to_install]), ) with indent_log(): for requirement in to_install: if requirement.conflicts_with: logger.info( 'Found existing installation: %s', requirement.conflicts_with, ) with indent_log(): requirement.uninstall(auto_confirm=True) try: requirement.install( install_options, global_options, *args, **kwargs ) except: # if install did not succeed, rollback previous uninstall if (requirement.conflicts_with and not requirement.install_succeeded): requirement.rollback_uninstall() raise else: if (requirement.conflicts_with and requirement.install_succeeded): requirement.commit_uninstall() requirement.remove_temporary_source() self.successfully_installed = to_install