Changeset - c7c5825299fe
[Not reviewed]
beta
0 1 0
Marcin Kuzminski - 13 years ago 2012-09-15 21:17:23
marcin@python-works.com
fixed logging messages on whoosh indexer
1 file changed with 5 insertions and 5 deletions:
0 comments (0 inline, 0 general)
rhodecode/lib/indexers/daemon.py
Show inline comments
 
@@ -235,180 +235,180 @@ class WhooshIndexingDaemon(object):
 

	
 
        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('>> COMMITTED CHANGES TO CHANGESET INDEX<<')
 
                    log.debug('>>> FINISHED REBUILDING CHANGESET INDEX <<<')
 
                else:
 
                    writer.cancel
 
                    log.debug('>> NOTHING TO COMMIT<<')
 
                    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()))
 

	
 
        idx = open_dir(self.index_location, indexname=self.indexname)
 
        # The set of all paths in the index
 
        indexed_paths = set()
 
        # The set of all paths we need to re-index
 
        to_index = set()
 

	
 
        writer = idx.writer()
 
        writer_is_dirty = False
 
        try:
 
            with idx.reader() as reader:
 

	
 
                # Loop over the stored fields in the index
 
                for fields in reader.all_stored_fields():
 
                    indexed_path = fields['path']
 
                    indexed_repo_path = fields['repository']
 
                    indexed_paths.add(indexed_path)
 

	
 
                    if not indexed_repo_path in self.filtered_repo_update_paths:
 
                        continue
 

	
 
                    repo = self.repo_paths[indexed_repo_path]
 

	
 
                    try:
 
                        node = self.get_node(repo, indexed_path)
 
                        # Check if this file was changed since it was indexed
 
                        indexed_time = fields['modtime']
 
                        mtime = self.get_node_mtime(node)
 
                        if mtime > indexed_time:
 
                            # The file has changed, delete it and add it to
 
                            # the list of files to reindex
 
                            log.debug(
 
                                'adding to reindex list %s mtime: %s vs %s' % (
 
                                    indexed_path, mtime, indexed_time)
 
                            )
 
                            writer.delete_by_term('fileid', indexed_path)
 
                            writer_is_dirty = True
 

	
 
                            to_index.add(indexed_path)
 
                    except (ChangesetError, NodeDoesNotExistError):
 
                        # This file was deleted since it was indexed
 
                        log.debug('removing from index %s' % indexed_path)
 
                        writer.delete_by_term('path', indexed_path)
 
                        writer_is_dirty = True
 

	
 
            # Loop over the files in the filesystem
 
            # Assume we have a function that gathers the filenames of the
 
            # documents to be indexed
 
            ri_cnt_total = 0  # indexed
 
            riwc_cnt_total = 0  # indexed with content
 
            for repo_name, repo in self.repo_paths.items():
 
                # skip indexing if there aren't any revisions
 
                if len(repo) < 1:
 
                    continue
 
                ri_cnt = 0   # indexed
 
                riwc_cnt = 0  # indexed with content
 
                for path in self.get_paths(repo):
 
                    path = safe_unicode(path)
 
                    if path in to_index or path not in indexed_paths:
 

	
 
                        # This is either a file that's changed, or a new file
 
                        # that wasn't indexed before. So index it!
 
                        i, iwc = self.add_doc(writer, path, repo, repo_name)
 
                        writer_is_dirty = True
 
                        log.debug('re indexing %s' % path)
 
                        ri_cnt += i
 
                        ri_cnt_total += 1
 
                        riwc_cnt += iwc
 
                        riwc_cnt_total += iwc
 
                log.debug('added %s files %s with content for repo %s' % (
 
                             ri_cnt + riwc_cnt, riwc_cnt, repo.path)
 
                )
 
            log.debug('indexed %s files in total and %s with content' % (
 
                        ri_cnt_total, riwc_cnt_total)
 
            )
 
        finally:
 
            if writer_is_dirty:
 
                log.debug('>> COMMITING CHANGES <<')
 
                log.debug('>> COMMITING CHANGES TO FILE INDEX <<')
 
                writer.commit(merge=True)
 
                log.debug('>>> FINISHED REBUILDING INDEX <<<')
 
                log.debug('>>> FINISHED REBUILDING FILE INDEX <<<')
 
            else:
 
                log.debug('>> NOTHING TO COMMIT<<')
 
                log.debug('>> NOTHING TO COMMIT TO FILE INDEX <<')
 
                writer.cancel()
 

	
 
    def build_indexes(self):
 
        if os.path.exists(self.index_location):
 
            log.debug('removing previous index')
 
            rmtree(self.index_location)
 

	
 
        if not os.path.exists(self.index_location):
 
            os.mkdir(self.index_location)
 

	
 
        chgset_idx = create_in(self.index_location, CHGSETS_SCHEMA,
 
                               indexname=CHGSET_IDX_NAME)
 
        chgset_idx_writer = chgset_idx.writer()
 

	
 
        file_idx = create_in(self.index_location, SCHEMA, indexname=IDX_NAME)
 
        file_idx_writer = file_idx.writer()
 
        log.debug('BUILDING INDEX FOR EXTENSIONS %s '
 
                  'AND REPOS %s' % (INDEX_EXTENSIONS, self.repo_paths.keys()))
 

	
 
        for repo_name, repo in self.repo_paths.items():
 
            # skip indexing if there aren't any revisions
 
            if len(repo) < 1:
 
                continue
 

	
 
            self.index_files(file_idx_writer, repo_name, repo)
 
            self.index_changesets(chgset_idx_writer, repo_name, repo)
 

	
 
        log.debug('>> COMMITING CHANGES <<')
 
        file_idx_writer.commit(merge=True)
 
        chgset_idx_writer.commit(merge=True)
 
        log.debug('>>> FINISHED BUILDING INDEX <<<')
 

	
 
    def update_indexes(self):
 
        self.update_file_index()
 
        self.update_changeset_index()
 

	
 
    def run(self, full_index=False):
 
        """Run daemon"""
 
        if full_index or self.initial:
 
            self.build_indexes()
 
        else:
 
            self.update_indexes()
0 comments (0 inline, 0 general)