muller-dataset

name	muller-dataset
description	Create, query, and manage MULLER datasets (multimodal data lake with version control). Use when user wants to work with datasets, create tensors, append data, query samples, or inspect dataset information.
compatibility	Requires Python 3.11+, muller package installed

MULLER Dataset Management

IMPORTANT: How to Use This Skill

DO NOT create new Python files. Always use the existing scripts provided in this skill:

• Use scripts/dataset_manager.py for dataset and tensor management
• Use scripts/data_operations.py for data CRUD operations

Execute these scripts directly with python3 command. Never write new scripts to the project root.

DEMO MODE: Speed Instructions

When executing operations:

• Run commands immediately without pre-explanation
• Do not summarize or explain results after execution
• Only show the JSON output from the script
• Skip all follow-up suggestions unless user asks

When to Use This Skill

Use this skill when the user wants to:

• Create or load MULLER datasets
• Add data to datasets (images, text, labels, vectors, etc.)
• Query or filter dataset samples
• Manage dataset structure (create/delete/rename tensors)
• Inspect dataset information (summary, statistics)
• Import data from files or directories

Available Scripts

scripts/dataset_manager.py

Manages dataset lifecycle and structure.

Operations:

• create - Create a new dataset
• load - Load existing dataset info
• delete - Delete a dataset
• info - Get dataset information
• stats - Get dataset statistics
• create-tensor - Create a new tensor
• delete-tensor - Delete a tensor
• rename-tensor - Rename a tensor

Usage:

# Create dataset
python3 .claude/skills/muller-dataset/scripts/dataset_manager.py create --path ./my_dataset

# Create with tensors
python3 .claude/skills/muller-dataset/scripts/dataset_manager.py create --path ./my_dataset \
  --tensors "images:image:jpg,labels:class_label:uint32"

# Get info
python3 .claude/skills/muller-dataset/scripts/dataset_manager.py info --path ./my_dataset

# Create tensor
python3 .claude/skills/muller-dataset/scripts/dataset_manager.py create-tensor --path ./my_dataset \
  --name embeddings --htype vector --dtype float32

scripts/data_operations.py

Handles data CRUD operations.

Operations:

• append - Add single sample
• extend - Add multiple samples
• update - Update existing sample
• delete - Delete samples
• query - Query and filter samples
• import - Import data from files

Usage:

# Append sample
python3 .claude/skills/muller-dataset/scripts/data_operations.py append --path ./my_dataset \
  --data '{"images": "path/to/img.jpg", "labels": 1}'

# Query samples
python3 .claude/skills/muller-dataset/scripts/data_operations.py query --path ./my_dataset \
  --filter "labels > 5" --limit 10

# Import from file
python3 .claude/skills/muller-dataset/scripts/data_operations.py import --path ./my_dataset \
  --source data.jsonl

Quick Reference

For detailed workflows and examples, check references/ directory.