def setUp(self): # Setup logging self.input_path_python = os.path.dirname( os.path.realpath(__file__)) + '/data/python_project' self.input_path_java = os.path.dirname( os.path.realpath(__file__)) + '/data/java_project2' self.out_java = os.path.dirname( os.path.realpath(__file__)) + '/data/out_java' # Clear database first (we need a small hack here, as mongomocks drop_database does not work) Project.drop_collection() VCSSystem.drop_collection() Commit.drop_collection() File.drop_collection() self.project_id = Project(name="zookeeper").save().id self.vcs_id = VCSSystem(url="http://test.de", project_id=self.project_id, repository_type="test").save().id self.commit_id = Commit(revision_hash="2342", vcs_system_id=self.vcs_id).save() self.file1 = File(path="contribs/CekiGulcu/AppenderTable.java", vcs_system_id=self.vcs_id).save() self.file2 = File(path="contribs/LeosLiterak/TempFileAppender.java", vcs_system_id=self.vcs_id).save() self.file3 = File( path="src/main/java/org/apache/log4j/AsyncAppender.java", vcs_system_id=self.vcs_id).save() shutil.rmtree(self.out_java, ignore_errors=True) shutil.rmtree(self.input_path_java, ignore_errors=True) os.makedirs(self.out_java) # Copying up fake files that are generated by SourceMeter self.class_csv = os.path.dirname( os.path.realpath(__file__)) + '/data/csv_data/zookeeper-Class.csv' self.package_csv = os.path.dirname(os.path.realpath( __file__)) + '/data/csv_data/zookeeper-Package.csv' self.component_csv = os.path.dirname(os.path.realpath( __file__)) + '/data/csv_data/zookeeper-Component.csv' shutil.copy(self.class_csv, self.out_java) shutil.copy(self.package_csv, self.out_java) shutil.copy(self.component_csv, self.out_java) # Create Files and directories os.makedirs(self.input_path_java + '/contribs/CekiGulcu') os.makedirs(self.input_path_java + '/contribs/LeosLiterak') os.makedirs(self.input_path_java + '/src/main/java/org/apache/log4j') Path(self.input_path_java + '/contribs/CekiGulcu/AppenderTable.java').touch() Path(self.input_path_java + '/contribs/LeosLiterak/TempFileAppender.java').touch() Path(self.input_path_java + '/src/main/java/org/apache/log4j/AsyncAppender.java').touch()
def main(args): # timing start = timeit.default_timer() if args.log_level and hasattr(logging, args.log_level): log.setLevel(getattr(logging, args.log_level)) uri = create_mongodb_uri_string(args.db_user, args.db_password, args.db_hostname, args.db_port, args.db_authentication, args.ssl) connect(args.db_database, host=uri) # Get the id of the project for which the code entities shall be merged try: project_id = Project.objects(name=args.project_name).get().id except DoesNotExist: log.error('Project %s not found!' % args.project_name) sys.exit(1) vcs = VCSSystem.objects(project_id=project_id).get() log.info("Starting commit labeling") # import every approach defined or all if args.approaches == 'all': # just list every module in the package and import it basepath = os.path.dirname(os.path.abspath(__file__)) for app in os.listdir(os.path.join(basepath, 'approaches/')): if app.endswith('.py') and app != '__init__.py': __import__('approaches.{}'.format(app[:-3])) else: # if we have a list of approaches import only those for app in args.approaches.split(','): __import__('approaches.{}'.format(app)) # add specific configs labelshark = LabelSHARK() commit_count = Commit.objects(vcs_system_id=vcs.id).count() for i,commit in enumerate(Commit.objects(vcs_system_id=vcs.id).only('id', 'revision_hash', 'vcs_system_id', 'message', 'linked_issue_ids', 'parents', 'fixed_issue_ids', 'szz_issue_ids').timeout(False)): if i%100 == 0: log.info("%i/%i commits finished", i, commit_count) labelshark.set_commit(commit) labels = labelshark.get_labels() #log.info('commit: {}, labels: {}'.format(commit.revision_hash, labels)) # save the labels if labels: tmp = {'set__labels__{}'.format(k): v for k, v in labels} Commit.objects(id=commit.id).upsert_one(**tmp) end = timeit.default_timer() - start log.info("Finished commit labeling in {:.5f}s".format(end))
def _classify_hunk_documentation(self, commit_issue_df): commit = Commit.objects(id=commit_issue_df._id[0]) commit_df = map_mongo_to_pandas(commit) commit_df.rename(columns={'_id': 'commit_id'}, inplace=True) # read changed files file_action_files = self._get_actions_files(commit_df) file_action_files = file_action_files.loc[:, ("commit_id", "path", "file_id", "file_action_id")] file_action_files = file_action_files[ file_action_files.path.str.contains( "(\.java$)|(\.c$)|(\.cpp$)|(\.py$)|(\.h$)|(\.cc$)") == True] commit_files = commit_df.merge(file_action_files, how="left") hunks = Hunk.objects( file_action_id__in=commit_files.file_action_id.values.tolist()) hunks_df = map_mongo_to_pandas(hunks) commit_files_hunk = commit_files.merge(hunks_df, how="left", left_on="file_action_id", right_on="file_action_id") # check if hunks contain actual code rather than comments and blank lines for candidate test commit_files_hunk["c_documentation"] = 0 commit_files_hunk = commit_files_hunk.apply( self._find_documentation_hunk, axis=1) commit_files_hunk = commit_files_hunk.loc[:, ("commit_id", "c_documentation")] commit_files_hunk = commit_files_hunk.groupby( "commit_id").sum().reset_index() commit_files_hunk.loc[commit_files_hunk.c_documentation != 0, ("c_documentation")] = 1 commit_files_hunk = commit_files_hunk.loc[:, ("commit_id", "c_documentation")] return commit_files_hunk
def initialize(self, config, repository_url, repository_type): """Initializes the mongostore by connecting to the mongodb, creating the project in the project collection \ and setting up processes (see: :class:`pyvcsshark.datastores.mongostore.CommitStorageProcess`, which read commits out of the commitqueue, process them and store them into the mongodb. :param config: all configuration :param repository_url: url of the repository, which is to be analyzed :param repository_type: type of the repository, which is to be analyzed (e.g. "git") """ logger.setLevel(config.debug_level) logger.info("Initializing MongoStore...") # Create queue for multiprocessing self.commit_queue = multiprocessing.JoinableQueue() # We define, that the user we authenticate with is in the admin database logger.info("Connecting to MongoDB...") uri = create_mongodb_uri_string(config.db_user, config.db_password, config.db_hostname, config.db_port, config.db_authentication, config.ssl_enabled) connect(config.db_database, host=uri, connect=False) # Get project_id try: project_id = Project.objects(name=config.project_name).get().id except DoesNotExist: logger.error('Project with name "%s" does not exist in database!' % config.project_name) sys.exit(1) # Check if vcssystem already exist, and use upsert vcs_system_id = VCSSystem.objects(url=repository_url).upsert_one( url=repository_url, repository_type=repository_type, last_updated=datetime.datetime.today(), project_id=project_id).id # Get the last commit by date of the project (if there is any) last_commit = Commit.objects(vcs_system_id=vcs_system_id)\ .only('committer_date').order_by('-committer_date').first() if last_commit is not None: last_commit_date = last_commit.committer_date else: last_commit_date = None # Start worker, they will wait till something comes into the queue and then process it for i in range(self.NUMBER_OF_PROCESSES): name = "StorageProcess-%d" % i process = CommitStorageProcess(self.commit_queue, vcs_system_id, last_commit_date, config, name) process.daemon = True process.start() logger.info("Starting storage Process...")
def run(self): """ Endless loop for the processes, which consists of several steps: 1. Get a object of class :class:`pyvcsshark.dbmodels.models.CommitModel` from the queue 2. Check if this commit was stored before and if it is so: update branches and tags (if they have changed) 3. Store author and committer in mongodb 4. Store Tags in mongodb 5. Create a list of branches, where the commit belongs to 6. Save the different file actions, which were done in this commit in the mongodb 7. Save the commit itself .. NOTE:: The committer date is used to check if a commit was already stored before. Meaning: We get the \ last commit out of the database and check if the committer date of the commits we process are > than the \ committer date of the last commit. .. WARNING:: We only look for changed tags and branches here for already processed commits! """ while True: commit = self.queue.get() logger.debug("Process %s is processing commit with hash %s." % (self.proc_name, commit.id)) # Try to get the commit try: mongo_commit = Commit.objects(vcs_system_id=self.vcs_system_id, revision_hash=commit.id).get() except DoesNotExist: mongo_commit = Commit(vcs_system_id=self.vcs_system_id, revision_hash=commit.id).save() self.set_whole_commit(mongo_commit, commit) # Save Revision object mongo_commit.save() logger.debug( "Process %s saved commit with hash %s. Queue size: %d" % (self.proc_name, commit.id, self.queue.qsize())) self.queue.task_done()
def get_commit_id(self, vcs_system_id): """ Gets the commit id for the corresponding projectid and revision :param vcs_system_id: id of the vcs system. :class:`bson.objectid.ObjectId` :return: commit_id (:class:`bson.objectid.ObjectId`) """ try: return Commit.objects(vcs_system_id=vcs_system_id, revision_hash=self.revision_hash).get().id except DoesNotExist: logger.error( "Commit with vcs_system_id %s and revision %s does not exist" % (vcs_system_id, self.revision_hash)) sys.exit(1)
def _classify_metric_documentation(self, commit_issue_df): commit = Commit.objects(id=commit_issue_df._id[0]) commit_df = map_mongo_to_pandas(commit) commit_hash_df = commit_df.apply(lambda x: pd.Series(x['parents']), axis=1).stack().reset_index(level=1, drop=True) commit_hash_df.name = 'parents' commit_hash_df = commit_df.drop("parents", axis=1).join(commit_hash_df) commit_hash_df = commit_hash_df.loc[:, ("_id", "revision_hash")] commit_hash_df.columns = ['commit_id', 'revision_hash'] # read previous revisions parent_commits = Commit.objects( revision_hash__in=commit_hash_df.revision_hash.values.tolist()) prev_df = map_mongo_to_pandas(parent_commits) prev_df = prev_df.loc[:, ("_id", "revision_hash")] prev_df.columns = ["parent_commit_id", "parent_revision_hash"] current_previous_commits_df = commit_hash_df.merge( prev_df, left_on="revision_hash", right_on="parent_revision_hash") current_previous_commits_df.drop("revision_hash", axis=1, inplace=True) # read changed files file_action_files = self._get_actions_files( current_previous_commits_df) file_action_files = file_action_files.loc[:, ("commit_id", "path", "file_id")] file_action_files = file_action_files[ file_action_files.path.str.contains("(\.java)") == True] commit_prev_files = current_previous_commits_df.merge( file_action_files) current_commits = [{ 'commit_id': commit_id } for commit_id in commit_prev_files.commit_id.values.tolist()] parent_commits = [{ 'commit_id': commit_id } for commit_id in commit_prev_files.parent_commit_id.values.tolist()] file_ids = [{ 'file_id': file_id } for file_id in commit_prev_files.file_id.values.tolist()] # supported types by sourcemeter for java code_entity_types = [{ 'ce_type': 'annotation' }, { 'ce_type': 'class' }, { 'ce_type': 'enum' }, { 'ce_type': 'interface' }, { 'ce_type': 'method' }] current_commits_files = [] for i in range(len(current_commits)): current_commits_files_dict = dict( list(current_commits[i].items()) + list(file_ids[i].items())) current_commits_files_dict['$or'] = code_entity_types current_commits_files.append(current_commits_files_dict) parent_commits_files = [] for i in range(len(parent_commits)): parent_commits_files_dict = dict( list(parent_commits[i].items()) + list(file_ids[i].items())) parent_commits_files_dict['$or'] = code_entity_types parent_commits_files.append(parent_commits_files_dict) # read metrics realted to commit current_code_entity_states = CodeEntityState.objects( __raw__={'$or': current_commits_files}) current_metrics = map_mongo_to_pandas(current_code_entity_states) current_metrics = current_metrics.loc[:, ("commit_id", "long_name", "metrics", "file_id")] current_metrics.rename(columns={'commit_id': 'commit_id_x'}, inplace=True) parent_code_entity_states = CodeEntityState.objects( __raw__={'$or': parent_commits_files}) parent_metrics = map_mongo_to_pandas(parent_code_entity_states) parent_metrics = parent_metrics.loc[:, ("commit_id", "long_name", "metrics", "file_id")] parent_metrics.rename(columns={'commit_id': 'commit_id_y'}, inplace=True) commit_prev_files_cpy = commit_prev_files.copy() commit_metrices_current = pd.merge(commit_prev_files, current_metrics, how='left', left_on=['commit_id', 'file_id'], right_on=['commit_id_x', 'file_id']) commit_metrices_current = commit_metrices_current.loc[:, ( "commit_id", "parent_commit_id", "file_id", "long_name", "metrics")] commit_metrices_parent = pd.merge( commit_prev_files_cpy, parent_metrics, how='left', left_on=['parent_commit_id', 'file_id'], right_on=['commit_id_y', 'file_id']) commit_metrices_parent = commit_metrices_parent.loc[:, ( "commit_id", "parent_commit_id", "file_id", "long_name", "metrics")] commit_metrics = pd.merge( commit_metrices_current, commit_metrices_parent, how='outer', left_on=['commit_id', 'parent_commit_id', 'file_id', 'long_name'], right_on=['commit_id', 'parent_commit_id', 'file_id', 'long_name']) commit_metrics["delta_DLOC"] = 0 commit_metrics = commit_metrics.apply(self._calculate_dloc_metric, axis=1) commit_metrics = commit_metrics.loc[:, ("commit_id", "delta_DLOC")] commit_metrics = commit_metrics.groupby( "commit_id").sum().reset_index() commit_metrics.loc[commit_metrics.delta_DLOC != 0, ("delta_DLOC")] = 1 return commit_metrics
def _find_boundary_date(self, issues, version_dates, affected_versions): """Find suspect boundary date. latest issue information but earliest date in commit (between created_at and affected versions) - latest creation date of linked bugs - earliest affected version """ tags = git_tag_filter(self._project_name, discard_patch=False, correct_broken_tags=True) issue_dates = [] affected_version_dates = [] for issue in issues: if not issue.created_at: self._log.warn( 'no reporting date for issue {} id({}), ignoring it'. format(issue.external_id, issue.id)) continue # direct link match, broken dates are already filtered in pycoshark so we do not need to do that here for av in issue.affects_versions: for tag in tags: if av.lower() == tag['original'].lower(): rev = tag['revision'] if 'corrected_revision' in tag.keys(): rev = tag['corrected_revision'] c = Commit.objects( vcs_system_id=self._vcs_id, revision_hash=rev).only('committer_date').get() affected_version_dates.append(c.committer_date) self._log.debug( 'found direct link between tag: {} and affected version: {} using ' .format(tag['original'], av)) for av in get_affected_versions(issue, self._project_name, self._jira_key): avt = tuple(av) if avt in version_dates.keys(): for version_date in version_dates[avt]: if version_date not in affected_version_dates: affected_version_dates.append(version_date) else: self._log.warn( 'affected version {} not found in git tags, skipping'. format(avt)) issue_dates.append(issue.created_at) # find latest bug report suspect_boundary_date = max(issue_dates) # latest bug report self._log.debug('latest bug report date is {} of {}'.format( suspect_boundary_date, issue_dates)) # return earliest affected version, only if we want if affected_versions and affected_version_dates: min_affected_date = min(affected_version_dates) self._log.debug( 'affected versions earliest date is {} while max bug report date is {}' .format(min_affected_date, suspect_boundary_date)) suspect_boundary_date = min(min_affected_date, suspect_boundary_date) self._log.debug( 'suspect boundary dates is {} from issue dates: {} and affected_versions: {}, use affected versions? {}' .format(suspect_boundary_date, issue_dates, affected_version_dates, affected_versions)) return suspect_boundary_date
def initialize(self, config, repository_url, repository_type): """Initializes the mongostore by connecting to the mongodb, creating the project in the project collection \ and setting up processes (see: :class:`pyvcsshark.datastores.mongostore.CommitStorageProcess`, which read commits out of the commitqueue, process them and store them into the mongodb. :param config: all configuration :param repository_url: url of the repository, which is to be analyzed :param repository_type: type of the repository, which is to be analyzed (e.g. "git") """ logger.setLevel(config.debug_level) logger.info("Initializing MongoStore...") # Create queue for multiprocessing self.commit_queue = multiprocessing.JoinableQueue() # we need an extra queue for branches because all commits need to be finished before we can process branches self.branch_queue = multiprocessing.JoinableQueue() self.config = config self.cores_per_job = config.cores_per_job # We define, that the user we authenticate with is in the admin database logger.info("Connecting to MongoDB...") uri = create_mongodb_uri_string(config.db_user, config.db_password, config.db_hostname, config.db_port, config.db_authentication, config.ssl_enabled) connect(config.db_database, host=uri, connect=False) # Get project_id try: project_id = Project.objects(name=config.project_name).get().id except DoesNotExist: logger.error('Project with name "%s" does not exist in database!' % config.project_name) sys.exit(1) # Check if vcssystem already exist, and use upsert vcs_system = VCSSystem.objects(url=repository_url).upsert_one( url=repository_url, repository_type=repository_type, last_updated=datetime.datetime.today(), project_id=project_id) self.vcs_system_id = vcs_system.id # Tar.gz name based on project name tar_gz_name = '{}.tar.gz'.format(config.project_name) # Tar.gz of repository folder with tarfile.open(tar_gz_name, "w:gz") as tar: tar.add(config.path, arcname=config.project_name) # Add repository to gridfs if not existent if vcs_system.repository_file.grid_id is None: logger.info('Copying project to gridfs...') # Store in gridfs with open(tar_gz_name, 'rb') as tar_file: vcs_system.repository_file.put(tar_file, content_type='application/gzip', filename=tar_gz_name) vcs_system.save() else: # replace file if not existent logger.info('Replacing project file in gridfs...') with open(tar_gz_name, 'rb') as tar_file: vcs_system.repository_file.replace( tar_file, content_type='application/gzip', filename=tar_gz_name) vcs_system.save() # Delete tar.gz file os.remove(tar_gz_name) # Get the last commit by date of the project (if there is any) last_commit = Commit.objects(vcs_system_id=self.vcs_system_id)\ .only('committer_date').order_by('-committer_date').first() if last_commit is not None: last_commit_date = last_commit.committer_date else: last_commit_date = None # Start worker, they will wait till something comes into the queue and then process it for i in range(self.cores_per_job): name = "StorageProcess-%d" % i process = CommitStorageProcess(self.commit_queue, self.vcs_system_id, last_commit_date, self.config, name) process.daemon = True process.start() logger.info("Starting storage Process...")
project_id = Project.objects(name=name).get().id except DoesNotExist: print('unknown project:', name) sys.exit(1) cur_vcs_system = VCSSystem.objects(project_id=project_id).get().id cur_issue_system = IssueSystem.objects(project_id=project_id).get().id # 1) fetch commits print('fetching commit ids') issue_ids = ['LLOC'] last_commit = None commit_bug_map = {} for commit in Commit.objects(vcs_system_id=cur_vcs_system, committer_date__gte=date_start, committer_date__lt=date_end, branches=master_branch)\ .only('id', 'committer_date', 'revision_hash', 'linked_issue_ids', 'message', 'parents'): linked_bugs = [] if commit.linked_issue_ids is not None and len( commit.linked_issue_ids) > 0: for issue in Issue.objects(id__in=commit.linked_issue_ids): if issue.external_id in excluded_issues: continue resolved = False fixed = False if issue.issue_type and issue.issue_type.lower() == 'bug': if issue.status in ['resolved', 'closed']: resolved = True fixed |= issue.resolution.lower() != 'duplicated'
def start(self, cfg): """ Executes the linkSHARK. :param cfg: configuration object that is used """ self._log.setLevel(cfg.get_debug_level()) start_time = timeit.default_timer() uri = create_mongodb_uri_string(cfg.user, cfg.password, cfg.host, cfg.port, cfg.authentication_db, cfg.ssl_enabled) connect(cfg.database, host=uri) # Get the id of the project for which the code entities shall be merged try: project_id = Project.objects(name=cfg.project_name).get().id except DoesNotExist: self._log.error('Project %s not found!' % cfg.project_name) sys.exit(1) vcs_system = VCSSystem.objects(project_id=project_id).get() self._itss = [] self._log.info('found the following issue tracking systems:') for its in IssueSystem.objects(project_id=project_id).order_by('url'): self._log.info(its.url) self._itss.append(its) if len(cfg.correct_key) > 0: correct_keys_per_its = cfg.correct_key.split(';') if len(correct_keys_per_its) != len(self._itss): self._log_critical( '--correct-key must correct keys for all issue tracking systems if specified' ) sys.exit(1) for i, correct_key in enumerate(correct_keys_per_its): self._correct_key[self._itss[i].url] = correct_key if len(cfg.broken_keys) > 0: broken_keys_per_its = cfg.broken_keys.split(';') if len(broken_keys_per_its) != len(self._itss): self._log_critical( '--broken-keys must correct keys for all issue tracking systems if specified. If there are no keys to correct for one of the ITS just use the name of the correct key twice itself' ) sys.exit(1) for i, broken_keys in enumerate(broken_keys_per_its): self._broken_keys[self._itss[i].url] = broken_keys.split(',') self._log.info("Starting issue linking") commit_count = Commit.objects(vcs_system_id=vcs_system.id).count() issue_map = {} for i, issue_system in enumerate(self._itss): project_id_string = correct_keys_per_its[i] for issue in Issue.objects(issue_system_id=issue_system.id): if issue.external_id.startswith(project_id_string): try: issue_number = [ int(s) for s in issue.external_id.split('-') if s.isdigit() ][0] except IndexError: self._log.error( "index error because SZZ currently only support JIRA, may not link all issues correctly:", issue.external_id) continue if issue_number not in issue_map: issue_map[issue_number] = [issue] else: issue_map[issue_number].append(issue) for i, commit in enumerate( Commit.objects(vcs_system_id=vcs_system.id).only( 'id', 'revision_hash', 'vcs_system_id', 'message', 'author_id', 'committer_id')): if i % 100 == 0: self._log.info("%i/%i commits finished", i, commit_count) issue_links = self._get_issue_links(commit) if len(issue_links) > 0: commit.linked_issue_ids = issue_links commit.save() szz_links = self._get_szz_issue_links(commit, issue_map) if len(szz_links) > 0: commit.szz_issue_ids = szz_links commit.save() elapsed = timeit.default_timer() - start_time self._log.info("Execution time: %0.5f s" % elapsed)
def main(args): # timing start = timeit.default_timer() if args.log_level and hasattr(logging, args.log_level): log.setLevel(getattr(logging, args.log_level)) uri = create_mongodb_uri_string(args.db_user, args.db_password, args.db_hostname, args.db_port, args.db_authentication, args.ssl) connect(args.db_database, host=uri) vcs = VCSSystem.objects.get(url=args.url) itss = [] if args.issue_systems == 'all': for its in IssueSystem.objects.filter(project_id=vcs.project_id): itss.append(its) else: for url in args.issue_systems.split(','): its = IssueSystem.objects.get(url=url) itss.append(its) log.info("Starting commit labeling") # import every approach defined or all if args.approaches == 'all': # just list every module in the package and import it basepath = os.path.dirname(os.path.abspath(__file__)) for app in os.listdir(os.path.join(basepath, 'approaches/')): if app.endswith('.py') and app != '__init__.py': __import__('approaches.{}'.format(app[:-3])) else: # if we have a list of approaches import only those for app in args.approaches.split(','): __import__('approaches.{}'.format(app)) # add specific configs config = {'itss': itss, 'args': args} a = LabelSHARK() a.configure(config) if args.linking_approach: log.info('using approach {} for issue links'.format( args.linking_approach)) for commit in Commit.objects.filter(vcs_system_id=vcs.id): a.set_commit(commit) labels = a.get_labels() issue_links = a.get_issue_links() # we get a dict of approach_name => [issue_link_ids] for k, v in issue_links.items(): # log.info('commit: {}, links: {}, from approach: {}'.format(commit.revision_hash, v, k)) if args.linking_approach and k == args.linking_approach: if v: log.info('commit: {}, linked to: {}'.format( commit.revision_hash, ','.join([str(l) for l in v]))) commit.linked_issue_ids = v commit.save() log.info('commit: {}, labels: {}'.format(commit.revision_hash, labels)) # save the labels if labels: tmp = {'set__labels__{}'.format(k): v for k, v in labels} Commit.objects(id=commit.id).upsert_one(**tmp) end = timeit.default_timer() - start log.info("Finished commit labeling in {:.5f}s".format(end))
def start(): """ Compares the commits and code_entity_states of two MongoDBs, whereas the first MongoDB is is condensed with the memeSHARK and the second MongoDB is verbose. """ setup_logging() logger = logging.getLogger("main") logger.info("Starting consistency checker...") parser = argparse.ArgumentParser(description='DB consistency checker.') parser.add_argument('-v', '--version', help='Shows the version', action='version', version='0.1.0') parser.add_argument('-U1', '--db-user1', help='Database user name', default=None) parser.add_argument('-P1', '--db-password1', help='Database user password', default=None) parser.add_argument('-DB1', '--db-database1', help='Database name', default='smartshark') parser.add_argument( '-H1', '--db-hostname1', help='Name of the host, where the database server is running', default='localhost') parser.add_argument('-p1', '--db-port1', help='Port, where the database server is listening', default=27017, type=int) parser.add_argument('-a1', '--db-authentication1', help='Name of the authentication database', default=None) parser.add_argument('--ssl1', help='Enables SSL', default=False, action='store_true') parser.add_argument('-U2', '--db-user2', help='Database user name', default=None) parser.add_argument('-P2', '--db-password2', help='Database user password', default=None) parser.add_argument('-DB2', '--db-database2', help='Database name', default='smartshark_backup') parser.add_argument( '-H2', '--db-hostname2', help='Name of the host, where the database server is running', default='localhost') parser.add_argument('-p2', '--db-port2', help='Port, where the database server is listening', default=27017, type=int) parser.add_argument('-a2', '--db-authentication2', help='Name of the authentication database', default=None) parser.add_argument('--ssl2', help='Enables SSL', default=False, action='store_true') parser.add_argument( '--debug', help='Sets the debug level.', default='DEBUG', choices=['DEBUG', 'INFO', 'WARNING', 'ERROR', 'CRITICAL']) parser.add_argument('--project-name1', help='Name of the project.', default=None) parser.add_argument('--project-name2', help='Name of the project.', default=None) args = parser.parse_args() logger.info(args) logger.info("connecting to database 1 (condensed)...") uri1 = create_mongodb_uri_string(args.db_user1, args.db_password1, args.db_hostname1, args.db_port1, args.db_authentication1, args.ssl1) logger.info(uri1) connect(args.db_database1, host=uri1, alias='default') logger.info("connecting to database 2 (verbose)...") uri2 = create_mongodb_uri_string(args.db_user2, args.db_password2, args.db_hostname2, args.db_port2, args.db_authentication2, args.ssl2) logger.info(uri2) connect(args.db_database2, host=uri2, alias='db-verbose') # fetch all verbose commmits commits_verbose = [] with switch_db(Commit, 'db-verbose') as CommitVerbose: # fetch only commits for selected project try: project_id = Project.objects(name=args.project_name2).get().id except DoesNotExist: logger.error('Project %s not found!' % args.project_name2) sys.exit(1) vcs_systems = VCSSystem.objects(project_id=project_id).get().id logger.info("vcs_system_id: %s", vcs_systems) commit_objects = Commit.objects(vcs_system_id=vcs_systems) for cur_commit_verbose in commit_objects: commits_verbose.append(cur_commit_verbose) with switch_db(VCSSystem, 'default') as VCSSystemCondensed: # fetch only commits for selected project try: project_id = Project.objects(name=args.project_name1).get().id except DoesNotExist: logger.error('Project %s not found!' % args.project_name1) sys.exit(1) vcs_systems_condensed = VCSSystemCondensed.objects( project_id=project_id).get().id # fetch files verbose with switch_db(File, 'db-verbose') as FilesVerbose: files_verbose = {} for cur_file_verbose in FilesVerbose.objects( vcs_system_id=vcs_systems): files_verbose[cur_file_verbose.id] = cur_file_verbose.path with switch_db(File, 'default') as FilesCondensed: files_condensed = {} for cur_file_condensed in FilesCondensed.objects( vcs_system_id=vcs_systems_condensed): files_condensed[cur_file_condensed.id] = cur_file_condensed.path num_commits_verbose = len(commits_verbose) logger.info("num commits verbose: %i", num_commits_verbose) for commit_nr, commit_verbose in enumerate(commits_verbose): logger.info("processing commit %s (%i / %i)", commit_verbose.id, commit_nr + 1, num_commits_verbose) # fetch verbose CES ces_verbose = {} ces_verbose_by_id = {} with switch_db(CodeEntityState, 'db-verbose') as CodeEntityStateVerbose: for cur_ces_verbose in CodeEntityStateVerbose.objects( commit_id=commit_verbose.id): ces_verbose[ cur_ces_verbose.long_name + files_verbose[cur_ces_verbose.file_id]] = cur_ces_verbose ces_verbose_by_id[cur_ces_verbose.id] = cur_ces_verbose # fetch same commit in condensed DB with switch_db(Commit, 'default') as CommitCondensed: try: commit_condensed = CommitCondensed.objects( revision_hash=commit_verbose.revision_hash, vcs_system_id=vcs_systems_condensed).get() except: logger.info("commit %s not found in condensed db", commit_verbose.revision_hash) continue # fetch CES from condensed DB ces_condensed = {} ces_condensed_by_id = {} with switch_db(CodeEntityState, 'default') as CodeEntityStateCondensed: for ces_id in commit_condensed.code_entity_states: cur_ces_condensed = CodeEntityStateCondensed.objects( id=ces_id).get() ces_condensed[cur_ces_condensed.long_name + files_condensed[ cur_ces_condensed.file_id]] = cur_ces_condensed ces_condensed_by_id[cur_ces_condensed.id] = cur_ces_condensed logger.info("num CES verbose : %i", len(ces_verbose.keys())) logger.info("num CES condensed: %i", len(ces_condensed.keys())) ces_unequal = 0 # compare CES for long_name_verbose, cur_ces_verbose in ces_verbose.items(): if long_name_verbose not in ces_condensed: logger.error( "CES with long_name %s not found in condensed DB!", long_name_verbose) ces_unequal += 1 continue cur_ces_condensed = ces_condensed[long_name_verbose] old, new = compare_dicts(cur_ces_verbose, cur_ces_condensed, { 'id', 's_key', 'commit_id', 'ce_parent_id', 'cg_ids', 'file_id' }) if len(new.keys()) > 0 or len(old.keys()) > 0: logger.error( "CES with long_name %s (id verbose: %s /id condensed %s) not equal!", long_name_verbose, cur_ces_verbose.id, cur_ces_condensed.id) logger.error("verbose : %s", old) logger.error("condensed: %s", new) ces_unequal += 1 continue # check if CES parent is equal ces_parent_verbose = ces_verbose_by_id[cur_ces_verbose.id] ces_parent_condensed = ces_condensed_by_id[cur_ces_condensed.id] old, new = compare_dicts\ (ces_parent_verbose, ces_parent_condensed, {'id', 's_key', 'commit_id', 'ce_parent_id', 'cg_ids', 'file_id'}) if len(new.keys()) > 0 or len(old.keys()) > 0: logger.error("ce_parent of CES with long_name %s not equal!", long_name_verbose) logger.error("verbose : %s", old) logger.error("condensed: %s", new) ces_unequal += 1 continue logger.info("num CES from verbose not matched: %i", ces_unequal)