class Git2DbMain():
"""
This class handles the import of Git data
"""
NUM_PROCESSES = 5
def __init__(self, db_name, project_name, repo_name, git_repo_path,
before_date, import_type, references, num_processes, config,
log_root_path):
"""
:type db_name: str
:param db_name: the name of an existing DB
:type project_name: str
:param project_name: the name of an existing project in the DB
:type repo_name: str
:param repo_name: the name of the Git repository to import
:type git_repo_path: str
:param git_repo_path: the local path of the Git repository
:type before_date: str
:param before_date: import data before date (YYYY-mm-dd)
:type import_type: int
:param import_type: 1 does not import patches, 2 imports patches but not at line level, 3 imports patches with line detail
:type references: list str
:param references: list of references to import
:type num_processes: int
:param num_processes: number of processes to import the data (default 5)
:type config: dict
:param config: the DB configuration file
:type log_root_path: str
:param log_root_path: the log path
"""
self._log_path = log_root_path + "import-git-" + db_name + "-" + project_name + "-" + repo_name
self._git_repo_path = git_repo_path
self._project_name = project_name
self._db_name = db_name
self._repo_name = repo_name
self._before_date = before_date
self._import_type = import_type
self._references = references
if num_processes:
self._num_processes = num_processes
else:
self._num_processes = Git2DbMain.NUM_PROCESSES
config.update({'database': db_name})
self._config = config
self._logging_util = LoggingUtil()
self._logger = None
self._fileHandler = None
self._querier = None
self._dao = None
def _get_existing_references(self, repo_id):
#retrieves already imported references
existing_refs = []
cursor = self._dao.get_cursor()
query = "SELECT ref.name " \
"FROM reference ref JOIN repository r ON ref.repo_id = r.id " \
"WHERE r.id = %s"
arguments = [repo_id]
self._dao.execute(cursor, query, arguments)
row = self._dao.fetchone(cursor)
while row:
existing_refs.append(row[0])
row = self._dao.fetchone(cursor)
self._dao.close_cursor(cursor)
return existing_refs
def _get_info_contribution(self, repo_id):
#processes Git data
existing_refs = self._get_existing_references(repo_id)
queue_references = multiprocessing.JoinableQueue()
results = multiprocessing.Queue()
# Start consumers
multiprocessing_util.start_consumers(self._num_processes,
queue_references, results)
for reference in self._querier.get_references():
if self._references:
if reference[0] in self._references:
git_ref_extractor = Git2DbReference(
self._db_name, repo_id, self._git_repo_path,
self._before_date, self._import_type, reference[0], "",
self._config, self._log_path)
queue_references.put(git_ref_extractor)
else:
if reference[0] not in existing_refs:
git_ref_extractor = Git2DbReference(
self._db_name, repo_id, self._git_repo_path,
self._before_date, self._import_type, reference[0], "",
self._config, self._log_path)
queue_references.put(git_ref_extractor)
# Add end-of-queue markers
multiprocessing_util.add_poison_pills(self._num_processes,
queue_references)
# Wait for all of the tasks to finish
queue_references.join()
def extract(self):
"""
extracts Git data and stores it in the DB
"""
try:
self._logger = self._logging_util.get_logger(self._log_path)
self._fileHandler = self._logging_util.get_file_handler(
self._logger, self._log_path, "info")
self._logger.info("Git2DbMain started")
start_time = datetime.now()
self._querier = GitQuerier(self._git_repo_path, self._logger)
self._dao = GitDao(self._config, self._logger)
project_id = self._dao.select_project_id(self._project_name)
self._dao.insert_repo(project_id, self._repo_name)
repo_id = self._dao.select_repo_id(self._repo_name)
#info contribution does not need a connection to the db
self._get_info_contribution(repo_id)
self._dao.restart_connection()
self._dao.fix_commit_parent_table(repo_id)
end_time = datetime.now()
minutes_and_seconds = self._logging_util.calculate_execution_time(
end_time, start_time)
self._logger.info("Git2DbMain finished after " +
str(minutes_and_seconds[0]) + " minutes and " +
str(round(minutes_and_seconds[1], 1)) + " secs")
self._logging_util.remove_file_handler_logger(
self._logger, self._fileHandler)
except Exception:
self._logger.error("Git2DbMain failed", exc_info=True)
finally:
if self._dao:
self._dao.close_connection()
class Code2DbUpdate():
"""
This class handles the update of code data
"""
NUM_PROCESSES = 5
def __init__(self, db_name, project_name, repo_name, git_repo_path,
extensions, references, num_processes, config, log_root_path):
"""
:type db_name: str
:param db_name: the name of an existing DB
:type project_name: str
:param project_name: the name of an existing project in the DB
:type repo_name: str
:param repo_name: the name of the Git repository to import
:type git_repo_path: str
:param git_repo_path: the local path of the Git repository
:type extensions: list str
:param extensions: file extensions to analyse. Currently extensions supported: ['java', 'py', 'php', 'scala', 'js', 'rb', 'cs', 'cpp', 'c']
:type references: list str
:param references: list of references to analyse
:type num_processes: int
:param num_processes: number of processes to import the data (default 5)
:type config: dict
:param config: the DB configuration file
:type log_root_path: str
:param log_root_path: the log path
"""
self._log_path = log_root_path + "update-code-" + db_name + "-" + project_name + "-" + repo_name
self._git_repo_path = git_repo_path
self._project_name = project_name
self._db_name = db_name
self._repo_name = repo_name
self._extensions = extensions
self._references = references
if num_processes:
self._num_processes = num_processes
else:
self._num_processes = Code2DbUpdate.NUM_PROCESSES
config.update({'database': db_name})
self._config = config
self._logging_util = LoggingUtil()
self._logger = None
self._fileHandler = None
self._querier = None
self._dao = None
def _get_new_commit_file_pairs(self, repo_id):
pairs = []
filter_references = "1 = 1"
if self._references:
filter_references = "r.name IN (" + ",".join(
["'" + e + "'" for e in self._references]) + ")"
filter_extensions = "1 = 1"
if self._extensions:
filter_extensions = "f.ext IN (" + ",".join(
["'" + e + "'" for e in self._extensions]) + ")"
cursor = self._dao.get_cursor()
query = "SELECT existing_pairs.* " \
"FROM ( " \
"SELECT cac.commit_id, cac.file_id FROM code_at_commit cac GROUP BY cac.commit_id, cac.file_id) AS processed_pairs " \
"RIGHT JOIN " \
"(SELECT c.id as commit_id, c.sha, f.id AS file_id, f.name AS file_name, f.ext AS file_ext " \
"FROM commit_in_reference cin JOIN reference r ON r.id = cin.ref_id " \
"JOIN commit c ON c.id = cin.commit_id " \
"JOIN file_modification fm ON fm.commit_id = c.id " \
"JOIN file f ON f.id = fm.file_id " \
"WHERE " + filter_references + " AND " + filter_extensions + " AND cin.repo_id = %s " \
"GROUP BY c.id, f.id) AS existing_pairs " \
"ON processed_pairs.commit_id = existing_pairs.commit_id AND processed_pairs.file_id = existing_pairs.file_id " \
"WHERE processed_pairs.commit_id IS NULL"
arguments = [repo_id]
self._dao.execute(cursor, query, arguments)
row = self._dao.fetchone(cursor)
while row:
pairs.append({
"commit_id": row[0],
"commit_sha": row[1],
"file_id": row[2],
"file_name": row[3],
"file_ext": row[4]
})
row = self._dao.fetchone(cursor)
self._dao.close_cursor(cursor)
return pairs
def _update_existing_references(self, repo_id, import_type):
pairs = self._get_new_commit_file_pairs(repo_id)
intervals = [
i for i in multiprocessing_util.get_tasks_intervals(
pairs, self._num_processes) if len(i) > 0
]
queue_intervals = multiprocessing.JoinableQueue()
results = multiprocessing.Queue()
# Start consumers
multiprocessing_util.start_consumers(self._num_processes, intervals,
results)
for interval in intervals:
issue_extractor = Code2DbCommitFile(self._db_name,
self._git_repo_path, interval,
import_type, self._config,
self._log_path)
queue_intervals.put(issue_extractor)
# Add end-of-queue markers
multiprocessing_util.add_poison_pills(self._num_processes,
queue_intervals)
# Wait for all of the tasks to finish
queue_intervals.join()
def _update_info_code(self, repo_id, import_type):
#updates code data
self._update_existing_references(repo_id, import_type)
def _get_import_type(self, repo_id):
#gets import type
import_type = 0
import_type += self._dao.function_at_commit_is_empty(
repo_id) + self._dao.code_at_commit_is_empty(repo_id)
return import_type
def update(self):
"""
updates the Git data stored in the DB
"""
try:
self._logger = self._logging_util.get_logger(self._log_path)
self._fileHandler = self._logging_util.get_file_handler(
self._logger, self._log_path, "info")
self._logger.info("Code2DbUpdate started")
start_time = datetime.now()
self._querier = GitQuerier(self._git_repo_path, self._logger)
self._dao = GitDao(self._config, self._logger)
project_id = self._dao.select_project_id(self._project_name)
repo_id = self._dao.select_repo_id(self._repo_name)
self._update_info_code(repo_id, self._get_import_type(repo_id))
end_time = datetime.now()
minutes_and_seconds = self._logging_util.calculate_execution_time(
end_time, start_time)
self._logger.info("Code2DbUpdate finished after " +
str(minutes_and_seconds[0]) + " minutes and " +
str(round(minutes_and_seconds[1], 1)) + " secs")
self._logging_util.remove_file_handler_logger(
self._logger, self._fileHandler)
except:
self._logger.error("Code2DbUpdate failed", exc_info=True)
finally:
if self._dao:
self._dao.close_connection()
class Git2DbUpdate():
"""
This class handles the update of Git data
"""
NUM_PROCESSES = 5
def __init__(self, db_name, project_name,
repo_name, git_repo_path, before_date,
num_processes, config, log_root_path):
"""
:type db_name: str
:param db_name: the name of an existing DB
:type project_name: str
:param project_name: the name of an existing project in the DB
:type repo_name: str
:param repo_name: the name of the Git repository to import
:type git_repo_path: str
:param git_repo_path: the local path of the Git repository
:type before_date: str
:param before_date: import data before date (YYYY-mm-dd)
:type num_processes: int
:param num_processes: number of processes to import the data (default 5)
:type config: dict
:param config: the DB configuration file
:type log_root_path: str
:param log_root_path: the log path
"""
self._log_path = log_root_path + "import-git-" + db_name + "-" + project_name + "-" + repo_name
self._git_repo_path = git_repo_path
self._project_name = project_name
self._db_name = db_name
self._repo_name = repo_name
self._before_date = before_date
self._existing_refs = []
if num_processes:
self._num_processes = num_processes
else:
self._num_processes = Git2DbUpdate.NUM_PROCESSES
config.update({'database': db_name})
self._config = config
self._logging_util = LoggingUtil()
self._logger = None
self._fileHandler = None
self._querier = None
self._dao = None
def _update_existing_references(self, repo_id, import_type):
#updates existing references in the DB
cursor = self._dao.get_cursor()
query = "SELECT c.sha, lc.ref_id " \
"FROM commit c " \
"JOIN (SELECT ref_id, max(commit_id) as last_commit_id_in_ref FROM commit_in_reference WHERE repo_id = %s GROUP BY ref_id) as lc " \
"ON c.id = lc.last_commit_id_in_ref"
arguments = [repo_id]
self._dao.execute(cursor, query, arguments)
queue_references = multiprocessing.JoinableQueue()
results = multiprocessing.Queue()
# Start consumers
multiprocessing_util.start_consumers(self._num_processes, queue_references, results)
row = self._dao.fetchone(cursor)
while row:
sha = row[0]
ref_id = row[1]
row = self._dao.fetchone(cursor)
ref_name = self._dao.select_reference_name(repo_id, ref_id)
for reference in self._querier.get_references():
reference_name = reference[0]
if reference_name == ref_name:
self._existing_refs.append(ref_name)
git_ref_extractor = Git2DbReference(self._db_name, repo_id, self._git_repo_path,
self._before_date, import_type, reference[0], sha,
self._config, self._log_path)
queue_references.put(git_ref_extractor)
break
self._dao.close_cursor(cursor)
# Add end-of-queue markers
multiprocessing_util.add_poison_pills(self._num_processes, queue_references)
# Wait for all of the tasks to finish
queue_references.join()
def _update_repo(self, repo_id, import_type):
#updates Git data
self._update_existing_references(repo_id, import_type)
def _get_import_type(self, repo_id):
#gets import type
import_type = 1
import_type += self._dao.line_detail_table_is_empty(repo_id) + self._dao.file_modification_patch_is_empty(repo_id)
return import_type
def update(self):
"""
updates the Git data stored in the DB
"""
try:
self._logger = self._logging_util.get_logger(self._log_path)
self._fileHandler = self._logging_util.get_file_handler(self._logger, self._log_path, "info")
self._logger.info("Git2DbUpdate started")
start_time = datetime.now()
self._querier = GitQuerier(self._git_repo_path, self._logger)
self._dao = GitDao(self._config, self._logger)
project_id = self._dao.select_project_id(self._project_name)
repo_id = self._dao.select_repo_id(self._repo_name)
self._update_repo(repo_id, self._get_import_type(repo_id))
self._dao.restart_connection()
self._dao.fix_commit_parent_table(repo_id)
end_time = datetime.now()
minutes_and_seconds = self._logging_util.calculate_execution_time(end_time, start_time)
self._logger.info("Git2DbUpdate finished after " + str(minutes_and_seconds[0])
+ " minutes and " + str(round(minutes_and_seconds[1], 1)) + " secs")
self._logging_util.remove_file_handler_logger(self._logger, self._fileHandler)
except:
self._logger.error("Git2DbUpdate failed", exc_info=True)
finally:
if self._dao:
self._dao.close_connection()