Skip to content

Commit

Permalink
CI_for_Specimin (#351)
Browse files Browse the repository at this point in the history
This CI evaluates accuracy of specimin for any push/pull events on main.
It will check file CI_Latest_run_percentage.txt for the latest accuracy
percentage of specimin.

The CI will pass if the current accuracy( accuracy calculated for
current change) matches the existing accuracy of main branch, if not the
CI will result in failure.

In case the calculated accuracy doesn't match, and is higher than the
value present in the file, it needs to be manually updated to the txt
file.
  • Loading branch information
NiharikaJamble authored Aug 9, 2024
1 parent 0c6ebd9 commit 0201a94
Show file tree
Hide file tree
Showing 7 changed files with 531 additions and 0 deletions.
132 changes: 132 additions & 0 deletions .github/workflows/check_compilable_percentage.yml
Original file line number Diff line number Diff line change
@@ -0,0 +1,132 @@
name: check_compilable_percentage_CI

on:
push:
branches:
- main
pull_request:
branches:
- main

jobs:
specimin-evaluation:
runs-on: ubuntu-latest

steps:
- name: Checkout repository
uses: actions/checkout@v2
with:
ref: ${{ github.event.pull_request.head.sha }}
fetch-depth: 0

- name: Set up environment
uses: actions/setup-python@v2
with:
python-version: '3.8'
- uses: actions/setup-java@v2
with:
java-version: '21'
distribution: 'adopt'

- name: Install dependencies
run: |
python -m pip install --upgrade pip
sudo apt-get update
sudo apt-get install -y jq curl bc
- name: Display CSV File Contents
run: |
CSV_PATH="/home/runner/work/specimin/specimin/CI_repository_list.csv"
if [ -f "$CSV_PATH" ]; then
cat "$CSV_PATH"
else
echo "File $CSV_PATH does not exist"
exit 1
fi
- name: Download git-clone-related and dependencies
run: |
curl -L -o git-clone-related https://raw.githubusercontent.com/plume-lib/git-scripts/main/git-clone-related
curl -L -o git-find-fork https://raw.githubusercontent.com/plume-lib/git-scripts/main/git-find-fork
curl -L -o git-find-branch https://raw.githubusercontent.com/plume-lib/git-scripts/main/git-find-branch
chmod +x git-clone-related git-find-fork git-find-branch
- name: Clone ASHE Project using git-clone-related
run: |
./git-clone-related njit-jerse ASHE_Automated-Software-Hardening-for-Entrypoints ASHE
- name: Check and Rename Properties File
run: |
set -ex
CONFIG_PATH="ASHE/src/main/resources/config.properties"
EXAMPLE_PATH="ASHE/src/main/resources/example.properties"
if [ -f "$CONFIG_PATH" ]; then
echo "config.properties already exists"
elif [ -f "$EXAMPLE_PATH" ]; then
echo "example.properties found, renaming to config.properties"
mv "$EXAMPLE_PATH" "$CONFIG_PATH"
if [ -f "$CONFIG_PATH" ]; then
echo "config.properties created successfully"
else
echo "Failed to create config.properties"
exit 1
fi
else
echo "Neither config.properties nor example.properties found"
exit 1
fi
chmod +w "$CONFIG_PATH"
ls -l "$CONFIG_PATH"
- name: Update ASHE Config File to update SPECIMIN path
run: |
set -ex
sed -i 's|^specimin.tool.path=.*|specimin.tool.path='$(pwd)'|' ASHE/src/main/resources/config.properties
- name: Make all scripts under ashe_scripts executable
run: |
set -ex
chmod +x ashe_scripts/*.py
- name: Run the script
run: |
set -ex
python3 ashe_scripts/run_ashe_for_stats.py \
$(pwd)/ASHE \
$(pwd)/CI_repository_list.csv \
$(pwd)/ASHE/CI_REPO_CLONE_SPACE \
$(pwd)/ASHE/src/main/resources/config.properties
- name: Parse accuracy percentage
id: parse_accuracy_percentage
run: |
current_accuracy=$(grep 'Fully successful from minimization to compilation' "$(pwd)/ASHE/logs/specimin_statistics.txt" | awk '{print $NF}' | tr -d '()%')
echo "Current accuracy: $current_accuracy"
echo "::set-output name=current_accuracy::$current_accuracy"
- name: Read latest run percentage from file
id: read_latest_run_percentage
run: |
if [ -f "$(pwd)/CI_Latest_run_percentage.txt" ]; then
latest_run_accuracy=$(cat "$(pwd)/CI_Latest_run_percentage.txt" | tr -d '()%')
echo "Latest run accuracy: $latest_run_accuracy"
echo "::set-output name=latest_run_accuracy::$latest_run_accuracy"
else
echo "File CI_Latest_run_percentage.txt does not exist"
exit 1
fi
- name: Validate accuracy
id: validate_accuracy
run: |
current_accuracy="${{ steps.parse_accuracy_percentage.outputs.current_accuracy }}"
latest_run_accuracy="${{ steps.read_latest_run_percentage.outputs.latest_run_accuracy }}"
if [ "$current_accuracy" != "$latest_run_accuracy" ]; then
echo "Current accuracy ($current_accuracy) does not match latest run accuracy ($latest_run_accuracy)."
exit 1
else
echo "Accuracy validation passed."
fi
1 change: 1 addition & 0 deletions CI_Latest_run_percentage.txt
Original file line number Diff line number Diff line change
@@ -0,0 +1 @@
88.61
2 changes: 2 additions & 0 deletions CI_repository_list.csv
Original file line number Diff line number Diff line change
@@ -0,0 +1,2 @@
Repository,Branch
https://github.com/NiharikaJamble/plume-util,master
10 changes: 10 additions & 0 deletions ashe_scripts/README.ME
Original file line number Diff line number Diff line change
@@ -0,0 +1,10 @@
# Specimin Statistics and Exception Ranking

### specimin_statistics.py
The script to parse the ASHE log files and generate statistical data from Specimin's minimization process.

### specimin_exception_ranking.py
The script to parse the ASHE log files and generate a ranking of the exceptions that occurred during the minimization process.

### run_ashe_for_stats.py
The script that clones ASHE, builds and runs it, and then runs the specimin_statistics.py and specimin_exception_rank.py scripts.
124 changes: 124 additions & 0 deletions ashe_scripts/run_ashe_for_stats.py
Original file line number Diff line number Diff line change
@@ -0,0 +1,124 @@
"""
Script to run Ashe.RepositoryAutomationEngine and Specimin scripts to analyze the log file generated by ASHE in dryrun mode.
https://github.com/jonathan-m-phillips/ASHE_Automated-Software-Hardening-for-Entrypoints
Created by: Jonathan Phillips, https://github.com/jonathan-m-phillips
Date: April 13, 2024
Usage:
python3 run_ashe_for_stats.py <path_to_clone_ashe> <path_to_csv> <path_to_clone_csv_repositories> <path_to_config.properties>
"""
import subprocess
import sys
import threading
import datetime
import time
import os


def run(ashe_path: str, csv_path: str, clone_path: str, props_file_path: str):
"""
Run ASHE and Specimin scripts to analyze the log file.
Args:
ashe_path: absolute path to clone the ASHE repository
csv_path: absolute path to the CSV file containing the repositories ASHE will iterate over
clone_path: absolute path to clone the repositories in the CSV file ASHE will iterate over
props_file_path: absolute path to the directory containing the config.properties files for ASHE
"""

ashe_url: str = "https://github.com/jonathan-m-phillips/ASHE_Automated-Software-Hardening-for-Entrypoints"
# clone or update repository
__git_clone_or_update(ashe_url, ashe_path)

start_time: datetime = datetime.datetime.now()
status_thread: threading.Thread = threading.Thread(target=__print_ashe_runtime, args=(start_time,))
status_thread.daemon = True
status_thread.start()
current_dir = os.path.dirname(os.path.abspath(__file__))
print(f"Current directory path: {current_dir}")
current_dir = current_dir.replace('ASHE/ashe_scripts', 'ashe_scripts')
main_project_dir = os.path.abspath(os.path.join(current_dir, '..', '..'))
stats_script = os.path.join(current_dir, 'specimin_statistics.py')
rank_script = os.path.join(current_dir, 'specimin_exception_rank.py')
print(f"Current directory path after normalising: {current_dir}")
print(f"main project path: {main_project_dir}")
print(f"Statistics script path: {stats_script}")
print(f"Exception rank script path: {rank_script}")

__build_and_run_ashe(csv_path, clone_path, props_file_path, working_dir=ashe_path)

current_dir = os.path.dirname(os.path.abspath(__file__))
print(f"Current directory path: {current_dir}")
current_dir = current_dir.replace('ASHE/ashe_scripts', 'ashe_scripts')
main_project_dir = os.path.abspath(os.path.join(current_dir, '..', '..'))
stats_script = os.path.join(current_dir, 'specimin_statistics.py')
rank_script = os.path.join(current_dir, 'specimin_exception_rank.py')
print(f"Current directory path after normalising: {current_dir}")
print(f"main project path: {main_project_dir}")
print(f"Statistics script path: {stats_script}")
print(f"Exception rank script path: {rank_script}")
# run Specimin scripts
log_path: str = os.path.join(ashe_path, "logs", "app.log")
print("Running statistics script...")
__run_command(f"python3 {stats_script} {log_path}")

print("Running exception rank script...")
__run_command(f"python3 {rank_script} {log_path}")


def __run_command(command, working_dir=None):
try:
result = subprocess.run(command, cwd=working_dir, shell=True, check=True, stdout=subprocess.PIPE,
stderr=subprocess.PIPE)
print(result.stdout.decode())
except subprocess.CalledProcessError as e:
print("Error executing command:", e.stderr.decode())


def __git_clone_or_update(repo_url, ashe_path):
"""Clone or update the git repository."""
if not os.path.exists(ashe_path):
print("Cloning the repository...")
__run_command(f"git clone {repo_url} {ashe_path}")
else:
print("Repository exists. Checking if it's a Git repository...")
if not os.path.exists(os.path.join(ashe_path, '.git')):
print(f"The directory {ashe_path} is not a Git repository.")
__run_command(f"git clone {repo_url} {ashe_path}")
else:
print("Updating the repository...")
os.chdir(ashe_path)
__run_command("git pull")


def __build_and_run_ashe(csv_path: str, clone_path: str, props_file_path: str, working_dir: str):
"""Build and run the ASHE project using gradle."""
# build ASHE
build_command: str = './gradlew build'
model_type: str = "dryrun"
run_automation_command: str = f"./gradlew runRepositoryAutomation -PrepositoriesCsvPath=\"{csv_path}\" -PcloneDirectory=\"{clone_path}\" -Pllm=\"{model_type}\" -PpropsFilePath=\"{props_file_path}\""

print("Building ASHE...")
__run_command(build_command, working_dir=working_dir)

print("Running ASHE...")
__run_command(run_automation_command, working_dir=working_dir)


def __print_ashe_runtime(start_time):
"""Function to print the elapsed time since ASHE started."""
print("ASHE started.")
print("ASHE runtime: 00:00:00")
while True:
time.sleep(300) # sleep for 5 minute
elapsed_time = datetime.datetime.now() - start_time
# format elapsed time into H:M:S
formatted_time = str(elapsed_time).split('.')[0] # remove microseconds
print(f"ASHE runtime: {formatted_time}")


if __name__ == "__main__":
if len(sys.argv) < 4:
print("Usage: python3 run_ashe_for_stats.py <path_to_clone_ashe> <path_to_csv> <path_to_clone_csv_repositories> <path_to_config.properties>")
sys.exit(1)
run(sys.argv[1], sys.argv[2], sys.argv[3], sys.argv[4])
125 changes: 125 additions & 0 deletions ashe_scripts/specimin_exception_rank.py
Original file line number Diff line number Diff line change
@@ -0,0 +1,125 @@
"""
Script for analyzing log files generated by ASHE in dryrun mode.
https://github.com/jonathan-m-phillips/ASHE_Automated-Software-Hardening-for-Entrypoints
Created by: Jonathan Phillips, https://github.com/jonathan-m-phillips
Date: April 13, 2024
Description:
This script reads a log file and ranks the exceptions by how frequently they occur. If the exceptions
occur more often, they are ranked higher. These exception rankings come from running the
Ashe.RepositoryAutomationEngine in dryrun mode.
Output:
Rankings written to a txt file in the same directory as the provided log file.
Usage:
python3 specimin_exception_rank.py <path_to_log_file.log>
"""

import sys
import os
import re
from collections import defaultdict


def analyze_log(file_path: str):
directory = os.path.dirname(file_path)
output_file_path = os.path.join(directory, 'specimin_exception_rank.txt')

with open(file_path, 'r') as file:
content = file.readlines()

exceptions = __extract_exceptions(content)
ranked_exceptions = __rank_exceptions(exceptions)

__write_ranked_exceptions(ranked_exceptions, output_file_path)
print("Write successful")


def __extract_exceptions(log_lines):
"""
Extract exceptions from the log lines. An exception is defined as a line that starts with "Exception in thread"
Args:
log_lines: A list of log lines
Returns: A list of tuples (name, message, example_line)
"""
# Enhanced to capture an example line following the exception message
exception_pattern = re.compile(r'^Exception in thread ".*?" (\w+.*?):(.*?)(?=\n\S|\Z)', re.DOTALL)
context_pattern = re.compile(r'^\s+at (.+)$', re.MULTILINE)
exceptions = []
for i, line in enumerate(log_lines):
match = exception_pattern.search(line)
if match:
exception_name, message = match.groups()
# find the next line that starts with whitespace followed by "at" to capture the context
context_match = context_pattern.search(log_lines[i + 1] if i + 1 < len(log_lines) else "")
example_line = context_match.group(1).strip() if context_match else "No code context available"
exceptions.append([exception_name.strip(), message.strip(), example_line])
return exceptions


def __rank_exceptions(exceptions):
"""
Rank the exceptions by how frequently they occur. If the exceptions occur more often, they are ranked higher.
Args:
exceptions: A list of tuples (name, message, example_line)
Returns: A sorted list of tuples (count, examples, name, message)
"""
grouped_exceptions = defaultdict(list)
for name, message, example in exceptions:
simplified_message = simplify_message(message)
grouped_exceptions[(name, simplified_message)].append(example)

# convert grouped data into a sorted list of tuples (count, examples, name, message)
sorted_exceptions = sorted(((len(v), v, k[0], k[1]) for k, v in grouped_exceptions.items()), reverse=True,
key=lambda x: x[0])
return sorted_exceptions


def simplify_message(message):
"""
Simplify the exception message by removing certain patterns that are not helpful for distinguishing exceptions.
Args:
message: The exception message for Specimin developers to analyze
Returns: A simplified version of the message
"""
message = re.sub(r'\bat [\w\.$<>]+\(.*?\)', '', message)
message = re.sub(r'\bLine \d+\b', '', message)
message = re.sub(r'\bmemory address 0x[\da-f]+\b', '', message, flags=re.I)
return message.strip()


def __write_ranked_exceptions(ranked_exceptions, output_file_path):
current_rank = 1
last_count = None
rank_increment = 0 # keeps track of how many ranks we should jump after ties

with open(output_file_path, 'w') as output_file:
for count, examples, name, message in ranked_exceptions:
if last_count != count:
current_rank += rank_increment
rank_increment = 1 # reset for next potential tie group
else:
rank_increment += 1 # increment to account for a tie when next different count comes

last_count = count
output_line = f"""
Rank: {current_rank},
Count: {count},
Exception: {name},
Message: {message},
Example: {examples[0]}
"""
output_file.write(output_line)


if __name__ == '__main__':
if len(sys.argv) < 2:
print("Usage: python3 specimin_exception_rank.py <path_to_log_file.log>")
sys.exit(1)
analyze_log(sys.argv[1])
Loading

0 comments on commit 0201a94

Please sign in to comment.