return cs

    def get_paths(self, repo):

        """

        recursive walk in root dir and return a set of all path in that dir

        based on repository walk function

        """

        index_paths_ = set()

        try:

            cs = self._get_index_changeset(repo)

            for _topnode, _dirs, files in cs.walk('/'):

                for f in files:

                    index_paths_.add(jn(safe_str(repo.path), safe_str(f.path)))

        except RepositoryError:

            log.debug(traceback.format_exc())

            pass

        return index_paths_

    def get_node(self, repo, path):

        """

        gets a filenode based on given full path.It operates on string for

        hg git compatability.

        :param repo: scm repo instance

        :param path: full path including root location

        :return: FileNode

        """

        root_path = safe_str(repo.path)+'/'

        parts = safe_str(path).partition(root_path)

        cs = self._get_index_changeset(repo)

        node = cs.get_node(parts[-1])

        return node

    def get_node_mtime(self, node):

        return mktime(node.last_changeset.date.timetuple())

    def add_doc(self, writer, path, repo, repo_name):

        """

        Adding doc to writer this function itself fetches data from

        the instance of vcs backend

        """

        node = self.get_node(repo, path)

        indexed = indexed_w_content = 0

        # we just index the content of chosen files, and skip binary files

        if node.extension in INDEX_EXTENSIONS and not node.is_binary:

            u_content = node.content

            if not isinstance(u_content, unicode):

                log.warning('  >> %s Could not get this content as unicode '

                            'replacing with empty content' % path)

                u_content = u''

            else:

                log.debug('    >> %s [WITH CONTENT]' % path)

                indexed_w_content += 1

        else:

            log.debug('    >> %s' % path)

            # just index file name without it's content

            u_content = u''

            indexed += 1

        p = safe_unicode(path)

        writer.add_document(

            fileid=p,

            owner=unicode(repo.contact),

            repository=safe_unicode(repo_name),

            path=p,

            content=u_content,

            modtime=self.get_node_mtime(node),

            extension=node.extension

        )

        return indexed, indexed_w_content

    def index_changesets(self, writer, repo_name, repo, start_rev=None):

        """

        Add all changeset in the vcs repo starting at start_rev

        to the index writer

        :param writer: the whoosh index writer to add to

        :param repo_name: name of the repository from whence the

          changeset originates including the repository group

        :param repo: the vcs repository instance to index changesets for,

          the presumption is the repo has changesets to index

        :param start_rev=None: the full sha id to start indexing from

          if start_rev is None then index from the first changeset in

          the repo

        """

        if start_rev is None:

            start_rev = repo[0].raw_id

        log.debug('indexing changesets in %s starting at rev: %s' %

                  (repo_name, start_rev))

        indexed = 0

            writer.add_document(

                raw_id=unicode(cs.raw_id),

                owner=unicode(repo.contact),

                date=cs._timestamp,

                repository=safe_unicode(repo_name),

                author=cs.author,

                message=cs.message,

                last=cs.last,

                added=u' '.join([safe_unicode(node.path) for node in cs.added]).lower(),

                removed=u' '.join([safe_unicode(node.path) for node in cs.removed]).lower(),

                changed=u' '.join([safe_unicode(node.path) for node in cs.changed]).lower(),

                parents=u' '.join([cs.raw_id for cs in cs.parents]),

            )

            indexed += 1

        log.debug('indexed %d changesets for repo %s' % (indexed, repo_name))

        return indexed

    def index_files(self, file_idx_writer, repo_name, repo):

        """

        Index files for given repo_name

        :param file_idx_writer: the whoosh index writer to add to

        :param repo_name: name of the repository we're indexing

        :param repo: instance of vcs repo

        """

        i_cnt = iwc_cnt = 0

        log.debug('building index for %s @revision:%s' % (repo.path,

                                                self._get_index_revision(repo)))

        for idx_path in self.get_paths(repo):

            i, iwc = self.add_doc(file_idx_writer, idx_path, repo, repo_name)

            i_cnt += i

            iwc_cnt += iwc

        log.debug('added %s files %s with content for repo %s' %

                  (i_cnt + iwc_cnt, iwc_cnt, repo.path))

        return i_cnt, iwc_cnt

    def update_changeset_index(self):

        idx = open_dir(self.index_location, indexname=CHGSET_IDX_NAME)

        with idx.searcher() as searcher:

            writer = idx.writer()

            writer_is_dirty = False

            try:

                indexed_total = 0

                repo_name = None

                for repo_name, repo in self.repo_paths.items():

                    # skip indexing if there aren't any revs in the repo

                    num_of_revs = len(repo)

                    if num_of_revs < 1:

                        continue

                    qp = QueryParser('repository', schema=CHGSETS_SCHEMA)

                    q = qp.parse(u"last:t AND %s" % repo_name)

                    results = searcher.search(q)

                    # default to scanning the entire repo

                    last_rev = 0

                    start_id = None

                    if len(results) > 0:

                        # assuming that there is only one result, if not this

                        # may require a full re-index.

                        start_id = results[0]['raw_id']

                        last_rev = repo.get_changeset(revision=start_id).revision

                    # there are new changesets to index or a new repo to index

                    if last_rev == 0 or num_of_revs > last_rev + 1:

                        # delete the docs in the index for the previous

                        # last changeset(s)

                        for hit in results:

                            q = qp.parse(u"last:t AND %s AND raw_id:%s" %

                                            (repo_name, hit['raw_id']))

                            writer.delete_by_query(q)

                        # index from the previous last changeset + all new ones

                        indexed_total += self.index_changesets(writer,

                                                repo_name, repo, start_id)

                        writer_is_dirty = True

                log.debug('indexed %s changesets for repo %s' % (

                          indexed_total, repo_name)

                )

            finally:

                if writer_is_dirty:

                    log.debug('>> COMMITING CHANGES TO CHANGESET INDEX<<')

                    writer.commit(merge=True)

                    log.debug('>>> FINISHED REBUILDING CHANGESET INDEX <<<')

                else:

                    log.debug('>> NOTHING TO COMMIT TO CHANGESET INDEX<<')

    def update_file_index(self):

        log.debug((u'STARTING INCREMENTAL INDEXING UPDATE FOR EXTENSIONS %s '

                   'AND REPOS %s') % (INDEX_EXTENSIONS, self.repo_paths.keys()))