# JavaMLBugDetective Configuration File

# =====================================

# This file contains all configuration settings for the bug detection analysis

# Repository Configuration

# ------------------------

# The Git repository URL to analyze

repository.url=https://github.com/google/gson.git

# Local directory where the repository will be cloned/analyzed

repository.local.path=./repositories/gson

# Name of the project (used for database naming)

project.name=gson

# Database Configuration

# ----------------------

# Note: Database name is automatically generated from project.name (e.g., "gson.db")

# Database connection timeout in seconds

database.timeout=30

# Analysis Configuration

# ----------------------

# Regular expression pattern to identify bug-fixing commits

bug.fix.pattern=(?i)(fix|bug|defect|issue|patch)

# File extensions to analyze (comma-separated)

file.extensions=.java

# Maximum number of commits to analyze (0 = all commits)

max.commits=0

# SZZ Algorithm Configuration

# ---------------------------

# Keywords to identify bug-fixing commits (comma-separated)

# More specific keywords to reduce noise and focus on severe bugs

szz.bug_fix_keywords=crash,exception,error,fault,fail,npe,nullpointer,incorrect result

# Confidence score for SZZ algorithm labeling (0.0 to 1.0)

szz.confidence_score=0.8

# Whether to ignore comments in blame analysis

szz.ignore_comments=true

# Whether to ignore blank lines in blame analysis

szz.ignore_blank_lines=true

# Machine Learning Configuration

# ------------------------------

# Train/test split ratio (e.g., 0.8 means 80% training, 20% testing)

ml.train.ratio=0.8

# ML algorithm to use:

# - RandomForest: Random Forest ensemble

# - J48: C4.5 decision tree

# - NaiveBayes: Naive Bayes classifier

# - SMO: Support Vector Machine

# - all: Run all algorithms for comparison

ml.algorithm=all

# Number of cross-validation folds

ml.cv.folds=10

# Enable class balancing for imbalanced datasets

ml.balance.classes=true

# Cost-Sensitive Classification Settings

# The cost of misclassifying a 'buggy' file as 'clean' (False Negative).

# Higher values make the model more aggressive at finding bugs.

ml.cost.fn=10.0

# The cost of misclassifying a 'clean' file as 'buggy' (False Positive).

ml.cost.fp=1.0

# Report Configuration

# --------------------

# Output directory for reports

report.output.dir=./reports

# Report format: markdown, html

report.format=markdown

# Include detailed metrics in reports

report.include.details=true

# GitHub Configuration

# --------------------

# GitHub credentials for private repositories (use a Personal Access Token)

github.username=

github.token=

# Performance Configuration

# -------------------------

# Enable verbose logging

logging.verbose=false

# Memory allocation for JVM operations (in MB)

jvm.memory.max=2048