A daemon will read from task table and run tasks

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, safe_unicode

from kallithea.lib.vcs.exceptions import ChangesetError, NodeDoesNotExistError, RepositoryError

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

        self.index_location = index_location

        if not index_location:

            raise Exception('You have to provide index location')

        self.repo_location = repo_location

        if not repo_location:

            raise Exception('You have to provide repositories location')

        self.repo_paths = ScmModel().repo_scan(self.repo_location)

        # filter repo list

        if repo_list:

            # Fix non-ascii repo names to unicode

            repo_list = set(safe_unicode(repo_name) for repo_name in repo_list)

            self.filtered_repo_paths = {}

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

                if repo_name in repo_list:

                    self.filtered_repo_paths[repo_name] = repo

            self.repo_paths = self.filtered_repo_paths

        # filter update repo list

        self.filtered_repo_update_paths = {}

        if repo_update_list:

            self.filtered_repo_update_paths = {}

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

                if repo_name in repo_update_list:

                    self.filtered_repo_update_paths[repo_name] = repo

            self.repo_paths = self.filtered_repo_update_paths

        self.initial = True

        if not os.path.isdir(self.index_location):

            os.makedirs(self.index_location)

            log.info('Cannot run incremental index since it does not '

                     'yet exist - running full build')

        elif not exists_in(self.index_location, IDX_NAME):

            log.info('Running full index build, as the file content '

                     'index does not exist')

        elif not exists_in(self.index_location, CHGSET_IDX_NAME):

            log.info('Running full index build, as the changeset '

                     'index does not exist')

        else:

            self.initial = False

    def _get_index_revision(self, repo):

        landing_rev = 'tip'

        if db_repo:

            _rev_type, _rev = db_repo.landing_rev

            landing_rev = _rev

        return landing_rev

    def _get_index_changeset(self, repo, index_rev=None):

        if not index_rev:

            index_rev = self._get_index_revision(repo)

        cs = repo.get_changeset(index_rev)

        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(os.path.join(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, index_rev=None):

        """

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

        hg git compatibility.

        :param repo: scm repo instance

        :param path: full path including root location

        :return: FileNode

        """

        # FIXME: paths should be normalized ... or even better: don't include repo.path

        path = safe_str(path)

        repo_path = safe_str(repo.path)

        assert path.startswith(repo_path)

        assert path[len(repo_path)] in (os.path.sep, os.path.altsep)

        node_path = path[len(repo_path) + 1:]

        cs = self._get_index_changeset(repo, index_rev=index_rev)

        node = cs.get_node(node_path)

        return node

    def is_indexable_node(self, node):

        """

        Just index the content of chosen files, skipping binary files

        """

        return (node.extension in INDEX_EXTENSIONS or node.name in INDEX_FILENAMES) and \

               not node.is_binary

    def get_node_mtime(self, node):

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

    def add_doc(self, writer, path, repo, repo_name, index_rev=None):

        """

        Adding doc to writer this function itself fetches data from

        the instance of vcs backend

        """

        try:

            node = self.get_node(repo, path, index_rev)

        except (ChangesetError, NodeDoesNotExistError):

            log.debug("    >> %s - not found in %s %s", path, repo, index_rev)

            return 0, 0

        indexed = indexed_w_content = 0

        if self.is_indexable_node(node):

            u_content = node.content

            if not isinstance(u_content, unicode):

                log.warning('    >> %s - no text content', path)

                u_content = u''

            else:

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

                indexed_w_content += 1

        else:

            log.debug('    >> %s - not indexable', 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_rawname=safe_unicode(repo_name),

            repository=safe_unicode(repo_name),

            path=p,

            content=u_content,

            modtime=self.get_node_mtime(node),

            extension=node.extension

        )

    Base for all available scm backends

    :created_on: Apr 8, 2010

    :copyright: (c) 2010-2011 by Marcin Kuzminski, Lukasz Balcerzak.

"""

import datetime

import itertools

from kallithea.lib.vcs.conf import settings

from kallithea.lib.vcs.exceptions import (

    ChangesetError, EmptyRepositoryError, NodeAlreadyAddedError, NodeAlreadyChangedError, NodeAlreadyExistsError, NodeAlreadyRemovedError, NodeDoesNotExistError, NodeNotChangedError, RepositoryError)

from kallithea.lib.vcs.utils import author_email, author_name, safe_unicode

from kallithea.lib.vcs.utils.helpers import get_dict_for_attrs

from kallithea.lib.vcs.utils.lazy import LazyProperty

class BaseRepository(object):

    """

    Base Repository for final backends

    **Attributes**

        ``DEFAULT_BRANCH_NAME``

            name of default branch (i.e. "trunk" for svn, "master" for git etc.

        ``scm``

            alias of scm, i.e. *git* or *hg*

        ``repo``

            object from external api

        ``revisions``

            list of all available revisions' ids, in ascending order

        ``changesets``

            storage dict caching returned changesets

        ``path``

            absolute path to the repository

        ``branches``

            branches as list of changesets

        ``tags``

            tags as list of changesets

    """

    scm = None

    DEFAULT_BRANCH_NAME = None

    EMPTY_CHANGESET = '0' * 40

    def __init__(self, repo_path, create=False, **kwargs):

        """

        Initializes repository. Raises RepositoryError if repository could

        not be find at the given ``repo_path`` or directory at ``repo_path``

        exists and ``create`` is set to True.

        :param repo_path: local path of the repository

        :param create=False: if set to True, would try to create repository.

        :param src_url=None: if set, should be proper url from which repository

          would be cloned; requires ``create`` parameter to be set to True -

          raises RepositoryError if src_url is set and create evaluates to

          False

        """

        raise NotImplementedError

    def __str__(self):

        return '<%s at %s>' % (self.__class__.__name__, self.path)

    def __repr__(self):

        return self.__str__()

    def __len__(self):

        return self.count()

    def __eq__(self, other):

        same_instance = isinstance(other, self.__class__)

        return same_instance and getattr(other, 'path', None) == self.path

    def __ne__(self, other):

        return not self.__eq__(other)

    @LazyProperty

    def alias(self):

        for k, v in settings.BACKENDS.items():

            if v.split('.')[-1] == str(self.__class__.__name__):

                return k

    @LazyProperty

    def name(self):

        """

        Return repository name (without group name)

        """

        raise NotImplementedError

    @LazyProperty

    def owner(self):

        raise NotImplementedError

    @LazyProperty

    def description(self):

        raise NotImplementedError

    @LazyProperty

    def size(self):

        """

        Returns combined size in bytes for all repository files

        """

        size = 0

        try:

            tip = self.get_changeset()

            for topnode, dirs, files in tip.walk('/'):

                for f in files:

                    size += tip.get_file_size(f.path)

        except RepositoryError as e:

            pass

        return size

    def is_valid(self):

        """

        Validates repository.

        """

        raise NotImplementedError

    def is_empty(self):

        return self._empty

    #==========================================================================

    # CHANGESETS

    #==========================================================================

    def get_changeset(self, revision=None):

        """

        Returns instance of ``Changeset`` class. If ``revision`` is None, most

        recent changeset is returned.

        :raises ``EmptyRepositoryError``: if there are no revisions

        """

        raise NotImplementedError

    def __iter__(self):

        """

        Allows Repository objects to be iterated.

        *Requires* implementation of ``__getitem__`` method.

        """

        for revision in self.revisions:

            yield self.get_changeset(revision)

    def get_changesets(self, start=None, end=None, start_date=None,

                       end_date=None, branch_name=None, reverse=False, max_revisions=None):

        """

        Returns iterator of ``BaseChangeset`` objects from start to end,

        both inclusive.

        :param start: None or str

        :param end: None or str

        :param start_date:

        :param end_date:

        :param branch_name:

        :param reversed:

        """

        raise NotImplementedError

    def __getslice__(self, i, j):

        """

        Returns a iterator of sliced repository

        """

        for rev in self.revisions[i:j]:

            yield self.get_changeset(rev)

    def __getitem__(self, key):

        return self.get_changeset(key)

    def count(self):

        return len(self.revisions)

    def tag(self, name, user, revision=None, message=None, date=None, **opts):

        """

        Creates and returns a tag for the given ``revision``.

        :param name: name for new tag

        :param user: full username, i.e.: "Joe Doe <joe.doe@example.com>"

        :param revision: changeset id for which new tag would be created

        :param message: message of the tag's commit

        :param date: date of tag's commit

        :raises TagAlreadyExistError: if tag with same name already exists

        """