Prompt for Generating Performance Analysis Reports

[apierr]

Objective

You are tasked with producing a data-driven performance analysis report comparing three cryptographic signing schemes used in Hyperledger Fabric:

• ECDSA (classical)
• PQC-only (e.g., Dilithium)
• Hybrid (ECDSA + PQC)

The goal is to highlight performance differences under varying load conditions using reproducible, statistically sound visualizations derived from existing CSV benchmark data.

---

Role and Methodology

Act as an expert in data analysis and performance benchmarking.

You must:

• Solve the problem step by step
• Explain your reasoning at each step
• Pause after each step and wait for confirmation before proceeding
• Avoid jumping to final conclusions prematurely

---

Input Data Context

You are provided with multiple CSV files generated via Monte Carlo simulations. Each file represents a single run under a specific cryptographic mode and load profile.

Key CSV Fields

• Identification: crypto_mode, load_profile, run_id
• Performance: tx_rate, latency_avg, latency_p95
• System Resources: cpu_util, mem_util
• Blockchain Metrics: block_size, block_commit_time
• Cryptographic Overhead: sig_gen_time, sig_verify_time

Load profiles include LOWLOAD and HIGHLOAD, with multiple runs per configuration.

---

Project Structure Constraints

All reporting scripts must integrate cleanly into the following project layout:

tools/
├── reporting/
│   ├── generate_performance_curve.py
│   ├── generate_overhead_breakdown.py
│   ├── generate_latex_tables.py
│   └── utils/

Engineering Constraints

• Scripts must follow SOLID and DRY principles
• Each file must contain no more than 50 lines of code
• Logic must be decomposed into reusable external modules
• Scripts must be executable via python -m

Example:

python -m tools.reporting.generate_performance_curve \
  --csv data/fixtures/monte_carlo/workshop/ECDSA_LOWLOAD_RUN1.csv \
  --output /tmp/performance_curve.png

---

Step-by-Step Tasks

Step 1 — Visualization Strategy

Propose and justify the types of charts that can be generated from the available CSV data.

Examples (non-exhaustive):

• Box plots
• Histograms
• Scatter plots
• Stacked bar charts
• Heatmaps

For each proposed visualization:

• Specify which metrics it applies to
• Explain what analytical insight it provides
• Explain why it is suitable for comparing ECDSA, PQC-only, and Hybrid schemes

Stop after this step and wait for confirmation before proceeding.

---

Step 2 — Script Design and File Structure

Once the visualization type is confirmed:

• Propose the script name
• Define the file/module structure
• Describe responsibilities of each module
• Explain how SOLID and DRY principles are enforced

Do not implement full code unless explicitly requested.

Stop after this step and wait for confirmation before proceeding.

---

Expected Output

The final outcome (after all steps are approved) will be:

• One or more production-ready reporting scripts
• Reproducible plots suitable for academic papers or technical reports
• Clear separation between data loading, transformation, and visualization logic

---

Tone and Quality Bar

• Use precise technical language
• Favor clarity, reproducibility, and statistical rigor
• Assume the audience consists of researchers and performance engineers

Do not simplify concepts unnecessarily. Focus on correctness and analytical value.