Beispiel #1
0
    def collect_used_dir_cache(self,
                               remote=None,
                               force=False,
                               jobs=None,
                               filter_info=None):
        """Get a list of `info`s related to the given directory.

        - Pull the directory entry from the remote cache if it was changed.

        Example:

            Given the following commands:

            $ echo "foo" > directory/foo
            $ echo "bar" > directory/bar
            $ dvc add directory

            It will return a NamedCache like:

            nc = NamedCache()
            nc.add(self.scheme, 'c157a79031e1', 'directory/foo')
            nc.add(self.scheme, 'd3b07384d113', 'directory/bar')
        """

        cache = NamedCache()

        try:
            self.get_dir_cache(jobs=jobs, remote=remote)
        except DvcException:
            logger.debug(f"failed to pull cache for '{self}'")

        try:
            objects.check(self.odb, self.odb.get(self.hash_info))
        except (FileNotFoundError, ObjectFormatError):
            msg = ("Missing cache for directory '{}'. "
                   "Cache for files inside will be lost. "
                   "Would you like to continue? Use '-f' to force.")
            if not force and not prompt.confirm(msg.format(self.path_info)):
                raise CollectCacheError(
                    "unable to fully collect used cache"
                    " without cache for directory '{}'".format(self))
            return cache

        path = str(self.path_info)
        filter_path = str(filter_info) if filter_info else None
        for entry_key, entry_obj in self.obj:
            entry_path = os.path.join(path, *entry_key)
            if (not filter_path or entry_path == filter_path
                    or entry_path.startswith(filter_path + os.sep)):
                cache.add(self.scheme, entry_obj.hash_info.value, entry_path)

        return cache
Beispiel #2
0
    def _collect_used_dir_cache(self, remote=None, force=False, jobs=None):
        """Get a list of `info`s related to the given directory.

        - Pull the directory entry from the remote cache if it was changed.

        Example:

            Given the following commands:

            $ echo "foo" > directory/foo
            $ echo "bar" > directory/bar
            $ dvc add directory

            It will return a NamedCache like:

            nc = NamedCache()
            nc.add(self.scheme, 'c157a79031e1', 'directory/foo')
            nc.add(self.scheme, 'd3b07384d113', 'directory/bar')
        """

        cache = NamedCache()

        if self.cache.changed_cache_file(self.checksum):
            try:
                self.repo.cloud.pull(
                    NamedCache.make("local", self.checksum, str(self)),
                    jobs=jobs,
                    remote=remote,
                    show_checksums=False,
                )
            except DvcException:
                logger.debug("failed to pull cache for '{}'".format(self))

        if self.cache.changed_cache_file(self.checksum):
            msg = ("Missing cache for directory '{}'. "
                   "Cache for files inside will be lost. "
                   "Would you like to continue? Use '-f' to force.")
            if not force and not prompt.confirm(msg.format(self.path_info)):
                raise CollectCacheError(
                    "unable to fully collect used cache"
                    " without cache for directory '{}'".format(self))
            else:
                return cache

        for entry in self.dir_cache:
            checksum = entry[self.remote.PARAM_CHECKSUM]
            path_info = self.path_info / entry[self.remote.PARAM_RELPATH]
            cache.add(self.scheme, checksum, str(path_info))

        return cache