Log Insight - AI-Powered Log Analysis

An intelligent log analysis tool using LLMs for context-aware insights, benchmarking, and publication-ready statistical evaluation.

Quick Start

python -m venv .venv && source .venv/bin/activate
pip install -r requirements.txt
cp .env.example .env  # set secrets

Core Workflows

Analyze a file and create a dashboard:

python -m src.cli analyze logs/app.log --chunk-size 10 --format html --output $(echo ${RESULTS_ROOT:-results})/reports/analysis_results.html

Run a benchmarking comparison and append stats to the manuscript:

python -m src.cli benchmark path/to/logs.txt path/to/ground_truth.json \
  --out-prefix $(echo ${RESULTS_ROOT:-results})/benchmarks/bench \
  --model o3-mini --manuscript docs/results/manuscript.md

Provision EC2, stream CloudWatch logs, analyze, and save a live-updated dashboard:

python -m src.cli experiment --region us-east-1 --key-name YOUR_KEYPAIR \
  --duration 300 --chunk-size 5 --out-root $(echo ${RESULTS_ROOT:-results})/experiments

Publish figures and stats to the manuscript:

python -m src.cli publish results/experiments/run_YYYYMMDD_HHMMSS --manuscript docs/results/manuscript.md
python -m src.cli publish_stats --results publication_ready_results.json --manuscript docs/results/manuscript.md

One-click statistical analysis (synthetic, for paper numbers):

python quick_statistical_analysis.py

Repository Structure

See docs/REPO_STRUCTURE.md for details. Generated artifacts go under results/ (configurable via RESULTS_ROOT).

Environment Variables

• OPENAI_API_KEY (required for LLM)
• ELASTIC_API_KEY, SPLUNK_USERNAME, SPLUNK_PASSWORD (optional connectors)
• RESULTS_ROOT (default: results)
• AWS_REGION (default: us-east-1)

Makefile Shortcuts

make install
make experiment KEY_NAME=your-keypair AWS_REGION=us-east-1
make benchmark LOG_FILE=path/to/logs.txt GROUND_TRUTH=path/to/gt.json
make publish EXP_DIR=results/experiments/run_YYYYMMDD_HHMMSS
make publish-stats

Notes

• All outputs are parameterized to land under results/*.
• Manuscript lives at docs/results/manuscript.md. Figures and stats can be appended via CLI.