Skillware skills ยท data_engineering/unit_normalizer
Unit Normalizer
Deterministic, dependency-free unit normalization, compound string parsing (5 ft 11 in, 3 1/2 gal, 120 km/h), and dimensional audit trail for batch agent pipelines, tables, and engineering feeds.
Capabilities
- Deterministic, dependency-free unit normalization for agent pipelines
- Single values, record batches (normalize_batch), and table columns (normalize_table)
- Compound string parser (5 ft 11 in, 3 1/2 gal, 120 km/h) and SI canonicalization
- Ambiguity detection (ton, gal, oz, pt, qt) with risk_variance and assume lock
- Per-record structured error codes without aborting batch; compact bounded output
Problems this skill solves
People search for tasks โ not always "skillware":
- deterministic unit conversion ai agent
- normalize mixed units batch python
- parse compound quantity string 5 ft 11 in python
- si unit canonicalization engineering pipeline
- dependency free unit converter python
Install
pip install skillware then SkillLoader.load_skill("data_engineering/unit_normalizer")
FAQ
- Why use a deterministic unit normalizer instead of letting an LLM convert units?
- LLMs frequently hallucinate conversion factors, temperature offsets, or ambiguous units (ton, gal, oz). Unit Normalizer uses one deterministic {factor, offset} table, declares ambiguities with risk_variance, and never aborts batches.