restgym-TrackB / src /check.py
minhhungg's picture
feat: initial upload for restgym Track B datasets and logs
e675fa3
Raw
History Blame Contribute Delete
7.58 kB
import common
import sqlite3
import os
import sys
import shutil
import re
import time
from rich import print
from rich.progress import Progress
# Collect paths of runs
def collect_runs():
runs = set()
api_dirs = os.scandir(f'{common.RESTGYM_BASE_DIR}/results')
for api_dir in api_dirs:
if api_dir.is_dir():
tool_dirs = os.scandir(api_dir)
for tool_dir in tool_dirs:
if tool_dir.is_dir():
run_dirs = os.scandir(tool_dir)
for run_dir in run_dirs:
if run_dir.name != '.DS_Store':
runs.add(run_dir.path)
return runs
# Read time budget from file
def parse_time_budget(file_path):
with open(file_path, 'r') as f:
content = f.read()
# Match "Time budget: X minutes." where X is integer
match = re.search(r'Time budget:\s*(\d+)', content)
if match:
minutes = int(match.group(1))
if minutes > 0:
return minutes
return 60
# Count code coverage samples
def count_coverage_samples(run):
files = os.listdir(f"{run}{common.CODE_COVERAGE_PATH}")
exec_count = 0
csv_count = 0
for file in files:
if file.endswith('.exec'):
exec_count += 1
elif file.endswith('.csv'):
csv_count += 1
return exec_count, csv_count
# Verify SQLite database integrity
def verify_database_integrity(run):
conn = sqlite3.connect(f'{run}/{common.DB_FILENAME}')
cursor = conn.cursor()
cursor.execute("PRAGMA integrity_check;")
result = cursor.fetchone()
conn.close()
if result[0] == "ok":
return True
else:
return False
# Perform integrity analysis
def analyze():
runs = collect_runs()
print(f"Analyzing {len(runs)} runs.")
analyzed = {}
# Progress bar
with Progress() as progress:
verify_task = progress.add_task("Verifying run data...", total=len(runs)+1)
progress.update(verify_task, advance=1)
for run in runs:
warnings = []
errors = []
# Get time budget for the run
time_budget = parse_time_budget(f'{run}/time-budget.txt')
# Verify started.txt exists
if not os.path.exists(f'{run}/started.txt'):
errors.append("The run did not start.")
# Verify completed.txt exists
if not os.path.exists(f'{run}/completed.txt'):
errors.append("The run did not complete.")
# Verify database exists
if not os.path.exists(f'{run}/{common.DB_FILENAME}'):
errors.append("The result database does not exist.")
else:
# Verify database integrity
if not verify_database_integrity(run):
warnings.append("The database integrity check failed.")
# Connect to the database
conn = sqlite3.connect(f"{run}/{common.DB_FILENAME}")
cursor = conn.cursor()
# Verify the interaction table is the database
if int(cursor.execute("SELECT COUNT(1) FROM sqlite_master WHERE type='table' AND name = 'interactions'").fetchone()[0]) == 0:
errors.append("The interaction table does not exist in the database.")
else:
# Count interactions
interaction_count = int(cursor.execute('SELECT COUNT(1) FROM interactions').fetchone()[0])
# Check if any request was recoded
if interaction_count == 0:
errors.append("No requests recorded in the database.")
else:
# Verify at least 100 requests per minute (in average) have been sent
if interaction_count < 100 * time_budget:
warnings.append(f"Less than 100 requests per minute have been recorded: {interaction_count} requests in {time_budget} minutes.")
# Verify requests time span is at least 80% of the time budget
interaction_time_span = int(int(cursor.execute('SELECT MAX(request_timestamp) - MIN(request_timestamp) FROM interactions').fetchone()[0]) / 60)
if interaction_time_span < int(0.8 * time_budget):
warnings.append(f"Requests time span is of {interaction_time_span}/{time_budget} minutes.")
# Verify code coverage dir exists
if not os.path.exists(f"{run}{common.CODE_COVERAGE_PATH}"):
errors.append("Code coverage directory does not exist.")
else:
# Verify all code coverage samples exist
exec_count, csv_count = count_coverage_samples(run)
if exec_count < 12 * time_budget:
errors.append("Missing code coverage EXEC samples.")
if csv_count < 12 * time_budget:
errors.append("Missing code coverage CSV samples.")
# Add result to returned dictionary
analyzed[run] = {
'errors': errors,
'warnings': warnings
}
# Print if something is wrong
if len(errors) + len(warnings) > 0:
message = f" => Run {'/'.join(os.path.normpath(run).split(os.sep)[-3:])} => "
if len(errors) > 0:
message += f"[red]ERRORS: {errors}[/red] "
if len(warnings) > 0:
message += f"[yellow]WARNINGS: {warnings}[/yellow]"
print(message)
# If no errors, create verified.txt file no error have been encountered
if len(errors) == 0:
content = f"This run was verified at {time.ctime()}.\n"
if len(warnings) > 0:
content += f"The verification passed with WARNINGS.\n\nWarnings: {warnings}\n"
else:
content += "The verification PASSED!\n"
with open(f'{run}/verified.txt', 'a') as f:
f.write(content)
progress.update(verify_task, advance=1)
return analyzed
# Removes runs with problems
def clean(runs):
count = len(runs)
for run in runs:
shutil.rmtree(run, ignore_errors=True)
print(f"Removed {count} runs.")
if __name__ == '__main__':
common.welcome()
print("This is the verify data module. It will check the integrity of experimental data of run.")
input("Press ENTER to start or CTRL+C to cancel...")
analyzed = analyze()
with_errors = []
with_warnings = []
for run in analyzed:
if len(analyzed[run]['errors']) > 0:
with_errors.append(run)
elif len(analyzed[run]['warnings']) > 0:
with_warnings.append(run)
if len(with_errors) + len(with_warnings) == 0:
print(f"All {len(analyzed)} runs passed the verification.")
else:
to_delete = []
print(f"Analyzed: {len(analyzed)}. [red]With errors: {len(with_errors)}.[/red] [yellow]With warnings: {len(with_warnings)}.[/yellow]")
if input(f"Do you want to delete runs with ERRORS? (yes/no): ").strip().lower() == "yes":
to_delete += with_errors
if input(f"Do you want to delete runs with WARNINGS? (yes/no): ").strip().lower() == "yes":
to_delete += with_warnings
if len(to_delete) == 0:
print("Nothing to delete.")
else:
clean(to_delete)