[Project Name]

This is a template for my personal research projects.

Overview

Brief description of what this project investigates.

Structure

• proposal/ - Project proposal and planning documents
• knowledge/ - Papers and notes (literature review, domain knowledge)
• experiments/ - Numbered experiment directories
• results/ - Aggregated results across experiments
• notebooks/ - Jupyter notebooks for exploration
• src/project_utils/ - Reusable code specific to this project
• tests/ - Tests for project utilities
• meetings/ - Meeting transcripts, summaries, and prep documents
• scripts/ - Helper scripts

Setup

1. Clone and Install

# Clone this project
git clone <repo-url>
cd <project-name>

# Install library dependencies (adjust path to your library)
pip install -e ../../libraries/<library-name>

# Install project dependencies
pip install -r requirements.txt

2. Configure for Your Project

1. Edit CLAUDE.md and update:

   • Project name and description
   • Library path
   • Any project-specific conventions

2. Edit proposal/proposal.tex with your actual proposal

3. Add papers to knowledge/papers/ as markdown

3. Start First Experiment

# Create a new experiment
./scripts/new-experiment.sh "initial-exploration"

# Work in it
cd experiments/NNN-initial-exploration   # use the dir name the script printed
# Edit README.md, main.py, etc.

Workflow

Running Experiments

cd experiments/NNN-name   # e.g. 003-my-experiment
python main.py

Running on HPC (SLURM)

Sync the repo to the cluster (clone or rsync), install dependencies in a venv/conda env, then:

# Submit an experiment (edit scripts/jobs/experiment.sbatch for partition, time, resources)
./scripts/submit-experiment.sh 3        # by number, or use NNN-name

For interactive runs on a node, cd experiments/NNN-name, set export MPLBACKEND=Agg, then python main.py. See CLAUDE.md “Running on HPC” for details.

Meeting Preparation

# Generate meeting prep document (before meeting)
./scripts/prep-meeting.sh

# Process transcript (after meeting)
claude /meeting-processor meetings/transcripts/YYYY-MM-DD-meeting.txt

Working with Library

When you need to add features to the library:

1. Create issue in THIS repo: gh issue create --title "[LIBRARY] Feature name"
2. Switch to library repo: cd ../../libraries/<library-name>
3. Create branch: git checkout -b feature/feature-name
4. Develop with TDD
5. Return to project to validate
6. Merge library PR after validation

Project Status

• Phase: [Exploration/Implementation/Analysis/Writing]
• Current Experiment: [See latest in experiments/]

Citation

If you use this work, please cite:

@article{,
  title={},
  author={},
  journal={},
  year={},
}