This file was forked by the Kallithea project in July 2014.

Original author and date, and relevant copyright and licensing information is below:

:created_on: Jan 26, 2010

:author: marcink

:copyright: (c) 2013 RhodeCode GmbH, and others.

:license: GPLv3, see LICENSE.md for more details.

"""

import logging

import os

import sys

import traceback

from os.path import dirname

from shutil import rmtree

from time import mktime

from whoosh.index import create_in, exists_in, open_dir

from whoosh.qparser import QueryParser

from kallithea.config.conf import INDEX_EXTENSIONS, INDEX_FILENAMES

from kallithea.lib.indexers import CHGSET_IDX_NAME, CHGSETS_SCHEMA, IDX_NAME, SCHEMA

from kallithea.lib.utils2 import safe_str

from kallithea.model.db import Repository

from kallithea.model.scm import ScmModel

# Add location of top level folder to sys.path

project_path = dirname(dirname(dirname(dirname(os.path.realpath(__file__)))))

sys.path.append(project_path)

log = logging.getLogger('whoosh_indexer')

class WhooshIndexingDaemon(object):

    """

    Daemon for atomic indexing jobs

    """

    def __init__(self, indexname=IDX_NAME, index_location=None,

                 repo_location=None, repo_list=None,

                 repo_update_list=None):

        self.indexname = indexname

            writer_is_dirty = False

            try:

                indexed_total = 0

                repo_name = None

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

                    log.debug('Updating changeset index for repo %s', repo_name)

                    # 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("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("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<<')