def test_argparse_no_patch(self):
parser = argparse.ArgumentParser()
method = parser.parse_args
slogging.add_logging_args(parser, False)
self.assertEqual(method, parser.parse_args)
parser = argparse.ArgumentParser()
method = parser.parse_args
slogging.add_logging_args(parser)
self.assertNotEqual(method, parser.parse_args)
def run_slogging_main():
parser = argparse.ArgumentParser()
slogging.add_logging_args(parser)
def my_setup(level: Union[str, int], structured: bool, config_path: str):
print(level)
print(structured)
print(config_path)
slogging.setup = my_setup
parser.parse_args()
def test_argparse_erase_args(self):
log_args = {"log_level", "log_structured", "log_config"}
parser = argparse.ArgumentParser()
slogging.add_logging_args(parser, erase_args=True)
with patch.object(sys, "argv"):
with patch("modelforge.slogging.setup"):
def my_setup(*args):
pass
slogging.setup = my_setup
args = parser.parse_args()
self.assertEqual(len(log_args.intersection(vars(args))), 0)
def create_parser() -> ArgumentParser:
"""
Create a parser for the lookout.style.format utility.
:return: an ArgumentParser with an handler defined in the handler attribute.
"""
# Deferred imports to speed up loading __init__
from lookout.style.format.benchmarks.compare_quality_reports import \
compare_quality_reports_entry
from lookout.style.format.benchmarks.evaluate_smoke import evaluate_smoke_entry
from lookout.style.format.benchmarks.generate_smoke import generate_smoke_entry
from lookout.style.format.benchmarks.quality_report import generate_quality_report
from lookout.style.format.benchmarks.general_report import print_reports
from lookout.style.format.benchmarks.quality_report_noisy import quality_report_noisy
from lookout.style.format.benchmarks.expected_vnodes_number import \
calc_expected_vnodes_number_entry
parser = ArgumentParser(
formatter_class=ArgumentDefaultsHelpFormatterNoNone)
# General options
slogging.add_logging_args(parser)
subparsers = parser.add_subparsers(help="Commands")
def add_parser(name, help):
return subparsers.add_parser(
name,
help=help,
formatter_class=ArgumentDefaultsHelpFormatterNoNone)
# Evaluation
eval_parser = add_parser("eval",
"Evaluate trained model on given dataset.")
eval_parser.set_defaults(handler=print_reports)
add_input_pattern_arg(eval_parser)
add_bblfsh_arg(eval_parser)
add_model_args(eval_parser)
eval_parser.add_argument(
"-n",
"--n-files",
default=0,
type=int,
help="How many files with most mispredictions to show. "
"If n <= 0 show all.")
# Generate quality report for the given data
quality_report_parser = add_parser(
"quality-report", "Generate quality report on a given data.")
quality_report_parser.set_defaults(handler=generate_quality_report)
quality_report_parser.add_argument(
"-i",
"--input",
required=True,
help=
"csv file with repositories to make report. Should contain url, to and from columns."
)
quality_report_parser.add_argument("-o",
"--output",
required=True,
help="Directory where to save results.")
quality_report_parser.add_argument(
"-f",
"--force",
default=False,
action="store_true",
help="Force to overwrite results stored in output directory if True. \
Stored results will be used if False.")
quality_report_parser.add_argument("-b",
"--bblfsh",
help="Bblfsh address to use.")
quality_report_parser.add_argument(
"--config",
type=json.loads,
default="{}",
help="Config for analyzer in json format.")
quality_report_parser.add_argument(
"--database",
default=None,
help="sqlite3 database path to store the models."
"Temporary file is used if not set.")
quality_report_parser.add_argument(
"--fs",
default=None,
help="Model repository file system root. "
"Temporary directory is used if not set.")
# Generate the quality report based on the artificial noisy dataset
quality_report_noisy_parser = add_parser(
"quality-report-noisy", "Quality report on the "
"artificial noisy dataset")
quality_report_noisy_parser.set_defaults(handler=quality_report_noisy)
add_bblfsh_arg(quality_report_noisy_parser)
add_rules_thresholds(quality_report_noisy_parser)
quality_report_noisy_parser.add_argument(
"-l",
"--language",
default="javascript",
help="Programming language to use.")
quality_report_noisy_parser.add_argument(
"--repos",
type=str,
help=
"list of urls or paths to the repositories to analyze. Should be strings separated "
"by newlines.")
quality_report_noisy_parser.add_argument(
"--precision-threshold",
type=float,
default=0.95,
help="Precision threshold tolerated for the model.")
quality_report_noisy_parser.add_argument(
"-o",
"--dir-output",
required=True,
type=str,
help=
"Path to the output directory where to store the quality report and the "
"precision-recall curve.")
# Compare two quality reports summaries
compare_quality_parser = add_parser(
"compare-quality",
"Creates a file with the differences in quality metrics between two reports."
)
compare_quality_parser.set_defaults(handler=compare_quality_reports_entry)
compare_quality_parser.add_argument(
"--base",
type=str,
required=True,
help=
"Baseline report. Usually the latest report from ./report/ directory.")
compare_quality_parser.add_argument(
"--new",
type=str,
required=True,
help=
"New report. Usually It is a report generated for master or any local \
change you did and want to validate.")
compare_quality_parser.add_argument(
"-o",
"--output",
type=str,
required=True,
help="Path to the file to save result or - to print to stdout.")
# Generate dataset of different styles in code for smoke testing.
gen_smoke_parser = add_parser(
"gen-smoke-dataset", "Generate dataset with different styles. "
"Helps to check the basic system functionality. "
"Only JavaScript code is supported now.")
gen_smoke_parser.set_defaults(handler=generate_smoke_entry)
gen_smoke_parser.add_argument(
"inputpath",
type=str,
help="Path to the tar.xz archive containing initial repositories.")
gen_smoke_parser.add_argument(
"outputpath",
type=str,
help=
"Path to the directory where the generated dataset should be stored.")
gen_smoke_parser.add_argument("--force",
default=False,
action="store_true",
help="Override output directory if exists.")
# Evaluate on different styles dataset
eval_smoke_parser = add_parser(
"eval-smoke-dataset", "Evaluate on the dataset with different styles.")
eval_smoke_parser.set_defaults(handler=evaluate_smoke_entry)
eval_smoke_parser.add_argument(
"inputpath",
type=str,
help="Path to the directory where the generated dataset is stored. "
"To generate a dataset run gen-smoke-dataset command.")
eval_smoke_parser.add_argument(
"reportdir",
type=str,
help="Path for report performance output directory.")
eval_smoke_parser.add_argument("--bblfsh",
help="Babelfish server's address.")
eval_smoke_parser.add_argument("--config",
type=json.loads,
default="{}",
help="JSON config for FormatAnalyzer.")
eval_smoke_parser.add_argument(
"--database",
type=str,
default=None,
help="Path to the sqlite3 database with trained models metadata. "
"Enables reusing previously trained models.")
rule_parser = add_parser("rule", "Print rule description by its hash.")
rule_parser.set_defaults(handler=dump_rule_entry)
rule_parser.add_argument("model", help="Path to the model file.")
rule_parser.add_argument("hash", help="Hash of the rule (8 chars).")
# FIXME(zurk): remove when https://github.com/src-d/style-analyzer/issues/557 is resolved
calc_expected_vnodes = add_parser(
"calc-expected-vnodes-number",
"Write the CSV file with expected numbers of virtual nodes "
"extracted from repositories. Required for quality report "
"generation. It is a workaround for "
"https://github.com/src-d/style-analyzer/issues/557. "
"Docker service is required to be running.")
calc_expected_vnodes.set_defaults(
handler=calc_expected_vnodes_number_entry)
calc_expected_vnodes.add_argument(
"-i",
"--input",
required=True,
help="CSV file with repositories for quality report."
"Should contain url, to and from columns.")
calc_expected_vnodes.add_argument("-o",
"--output",
required=True,
help="Path to a output csv file.")
calc_expected_vnodes.add_argument(
"-r",
"--runs",
default=3,
help="Number of repeats to ensure the result correctness.")
return parser
def parse_args() -> argparse.Namespace:
"""
Create the cmdline argument parser.
"""
parser = argparse.ArgumentParser(
formatter_class=ArgumentDefaultsHelpFormatterNoNone)
slogging.add_logging_args(parser, patch=True, erase_args=False)
# Create and construct subparsers
subparsers = parser.add_subparsers(help="Commands", dest="command")
def add_parser(name, help_message):
return subparsers.add_parser(
name,
help=help_message,
formatter_class=ArgumentDefaultsHelpFormatterNoNone)
# --------------------------------------------------------------------------------------------
clickhouse2deps_parser = add_parser(
"clickhouse2deps", "Extract dependencies from a ClickHouse DB.")
clickhouse2deps_parser.set_defaults(handler=clickhouse2deps)
clickhouse2deps_parser.add_argument(
"-o",
"--output-path",
type=Path,
help=
"Output path to the resulting ASDF model with the extracted dependencies.",
)
clickhouse2deps_parser.add_argument(
"-f",
"--force",
action="store_true",
help=
"Boolean indicating whether to overwrite the existing ASDF model specified by "
"-o/--output-path.",
)
clickhouse2deps_parser.add_argument("--user",
default="default",
help="Username for the DB.")
clickhouse2deps_parser.add_argument("--password",
default="",
help="Password for the DB.")
clickhouse2deps_parser.add_argument("--host",
default="0.0.0.0",
help="Host for the DB.")
clickhouse2deps_parser.add_argument("--port",
default=9000,
type=int,
help="Port for the DB.")
clickhouse2deps_parser.add_argument("--database",
default="default",
help="Database name for the DB.")
clickhouse2deps_parser.add_argument("--table",
default="uasts",
help="Table name for the DB.")
clickhouse2deps_parser.add_argument(
"--langs",
nargs="+",
default=CLICKHOUSE_LANGS,
choices=CLICKHOUSE_LANGS,
help="Languages to consider while extracting dependencies.",
)
# --------------------------------------------------------------------------------------------
collect_stdlibs_parser = add_parser(
"collect-stdlibs",
"Collect the lists of standard libraries for each language Babelfish can parse.",
)
collect_stdlibs_parser.set_defaults(handler=collect_stdlibs)
collect_stdlibs_parser.add_argument(
"-o",
"--output-path",
type=Path,
help=
"Output path to the resulting ASDF model with the extracted standard libraries.",
)
collect_stdlibs_parser.add_argument(
"-f",
"--force",
action="store_true",
help=
"Boolean indicating whether to overwrite the existing ASDF model specified by "
"-o/--output-path.",
)
args = parser.parse_args()
if not hasattr(args, "handler"):
args.handler = lambda _: parser.print_usage() # noqa: E731
return args
def create_parser() -> ArgumentParser:
"""
Create a parser for the lookout.style.typos utility.
:return: an ArgumentParser with an handler defined in the handler attribute.
"""
parser = ArgumentParser(
formatter_class=ArgumentDefaultsHelpFormatterNoNone)
# General options
slogging.add_logging_args(parser)
subparsers = parser.add_subparsers(help="Commands")
def add_parser(name, help):
return subparsers.add_parser(
name,
help=help,
formatter_class=ArgumentDefaultsHelpFormatterNoNone)
# Prepare raw data for corrector
prepare_parser = add_parser("prepare-data",
"Prepare raw dataset for corrector training.")
prepare_parser.set_defaults(handler=prepare_data)
add_config_arg(prepare_parser)
# Train new fasttext model on gien data
fasttext_parser = add_parser(
"train-fasttext", "Train fasttext model on the given dataset"
"of code identifiers.")
fasttext_parser.set_defaults(handler=cli_train_fasttext)
add_data_path_arg(fasttext_parser)
add_config_arg(fasttext_parser)
# Create train and test datasets with artificial typos
datasets_parser = add_parser(
"get-datasets", "Create the train and the test datasets of typos.")
datasets_parser.set_defaults(handler=cli_get_datasets)
add_data_path_arg(datasets_parser)
add_config_arg(datasets_parser)
# Create, train and evaluate new corrector model
train_parser = add_parser(
"train", "Create and train TyposCorrector model on the given data.")
train_parser.set_defaults(handler=cli_train_corrector)
train_parser.add_argument(
"--train",
required=False,
type=str,
default=DEFAULT_CORRECTOR_CONFIG["datasets"]["train_path"],
help=
".csv dump of a Dataframe with columns Columns.Split and Columns.Frequency.",
)
train_parser.add_argument(
"--test",
required=False,
type=str,
default=DEFAULT_CORRECTOR_CONFIG["datasets"]["test_path"],
help=
".csv dump of a Dataframe with columns Columns.Split and Columns.Frequency.",
)
train_parser.add_argument(
"-v",
"--vocabulary-path",
required=False,
type=str,
default=os.path.join(
DEFAULT_CORRECTOR_CONFIG["preparation"]["data_dir"],
DEFAULT_CORRECTOR_CONFIG["preparation"]["vocabulary_filename"]),
help="Path to a .csv file with vocabulary.",
)
train_parser.add_argument(
"-f",
"--frequencies-path",
required=False,
type=str,
default=os.path.join(
DEFAULT_CORRECTOR_CONFIG["preparation"]["data_dir"],
DEFAULT_CORRECTOR_CONFIG["preparation"]["frequencies_filename"]),
help="Path to a .csv file with tokens' frequencies.",
)
train_parser.add_argument(
"-e",
"--fasttext-path",
required=False,
type=str,
default=DEFAULT_CORRECTOR_CONFIG["fasttext"]["path"],
help="Path to a FastText model's dump (.bin).",
)
add_config_arg(train_parser)
add_corrector_path_arg(train_parser)
########################################
# One command to rule them all
########################################
train_from_scratch_parser = add_parser(
"train-from-scratch",
"Create and train TyposCorrector model on the given data.")
train_from_scratch_parser.set_defaults(handler=train_from_scratch)
add_config_arg(train_from_scratch_parser)
# Report for Typo Commits Dataset
typo_commits_report_parser = add_parser(
"typo-commits-report", "Generate report for Typo Commits Dataset.")
typo_commits_report_parser.set_defaults(
handler=generate_typos_report_entry)
add_config_arg(typo_commits_report_parser)
typo_commits_report_parser.add_argument(
"-i",
"--dataset",
required=True,
help=
"csv file with commits with typos. Must contain wrong_id, correct_id, file, line, "
"commit_fix, repo, commit_typo columns. It is possible to specify the xz compressed "
"file")
typo_commits_report_parser.add_argument(
"-o",
"--output",
required=True,
help="Directory where to save results.")
typo_commits_report_parser.add_argument("-b",
"--bblfsh",
help="Bblfsh address to use.")
typo_commits_report_parser.add_argument(
"--database",
default=None,
help="sqlite3 database path to store the models."
"Temporary file is used if not set.")
typo_commits_report_parser.add_argument(
"--fs",
default=None,
help="Model repository file system root. "
"Temporary directory is used if not set.")
typo_commits_report_parser.add_argument(
"--repos-cache",
default=None,
required=False,
help=
"Directory where to download repositories from the dataset. It is strongly \
recommended to set this parameter if there are more than 20 repositories \
in the dataset. Temporary directory is used if not set.")
return parser
def create_parser() -> ArgumentParser:
"""
Create a parser for the lookout.style.typos utility.
:return: an ArgumentParser with an handler defined in the handler attribute.
"""
parser = ArgumentParser(
formatter_class=ArgumentDefaultsHelpFormatterNoNone)
# General options
slogging.add_logging_args(parser)
subparsers = parser.add_subparsers(help="Commands")
def add_parser(name, help):
return subparsers.add_parser(
name,
help=help,
formatter_class=ArgumentDefaultsHelpFormatterNoNone)
# Prepare raw data for corrector
prepare_parser = add_parser("prepare-data",
"Prepare raw dataset for corrector training.")
prepare_parser.set_defaults(handler=prepare_data)
add_config_arg(prepare_parser)
# Train new fasttext model on gien data
fasttext_parser = add_parser(
"train-fasttext", "Train fasttext model on the given dataset"
"of code identifiers.")
fasttext_parser.set_defaults(handler=cli_train_fasttext)
add_data_path_arg(fasttext_parser)
add_config_arg(fasttext_parser)
# Create train and test datasets with artificial typos
datasets_parser = add_parser(
"get-datasets", "Create the train and the test datasets of typos.")
datasets_parser.set_defaults(handler=cli_get_datasets)
add_data_path_arg(datasets_parser)
add_config_arg(datasets_parser)
# Create, train and evaluate new corrector model
train_parser = add_parser(
"train", "Create and train TyposCorrector model on the given data.")
train_parser.set_defaults(handler=cli_train_corrector)
train_parser.add_argument(
"--train",
required=False,
type=str,
default=DEFAULT_CONFIG["datasets"]["train_path"],
help=
".csv dump of a Dataframe with columns Columns.Split and Columns.Frequency.",
)
train_parser.add_argument(
"--test",
required=False,
type=str,
default=DEFAULT_CONFIG["datasets"]["test_path"],
help=
".csv dump of a Dataframe with columns Columns.Split and Columns.Frequency.",
)
train_parser.add_argument(
"-v",
"--vocabulary-path",
required=False,
type=str,
default=os.path.join(
DEFAULT_CONFIG["preparation"]["data_dir"],
DEFAULT_CONFIG["preparation"]["vocabulary_filename"]),
help="Path to a .csv file with vocabulary.",
)
train_parser.add_argument(
"-f",
"--frequencies-path",
required=False,
type=str,
default=os.path.join(
DEFAULT_CONFIG["preparation"]["data_dir"],
DEFAULT_CONFIG["preparation"]["frequencies_filename"]),
help="Path to a .csv file with tokens' frequencies.",
)
train_parser.add_argument(
"-e",
"--fasttext-path",
required=False,
type=str,
default=DEFAULT_CONFIG["fasttext"]["path"],
help="Path to a FastText model's dump (.bin).",
)
add_corrector_path_arg(train_parser)
########################################
# One command to rule them all
########################################
train_from_scratch_parser = add_parser(
"train-from-scratch",
"Create and train TyposCorrector model on the given data.")
train_from_scratch_parser.set_defaults(handler=train_from_scratch)
add_config_arg(train_from_scratch_parser)
return parser
