benchmark

name	benchmark
description	Run csv-nose benchmark accuracy tests against POLLOCK, W3C-CSVW, or CSV Wrangling datasets

Benchmark Skill

Run benchmark accuracy tests. Accept an optional dataset argument: pollock, w3c, wrangling, codec, messy, or all (default).

Commands

Dataset	Command
pollock	cargo run --release -- --benchmark tests/data/pollock --annotations tests/data/annotations/pollock.txt
w3c	cargo run --release -- --benchmark tests/data/w3c-csvw --annotations tests/data/annotations/w3c-csvw.txt
wrangling	cargo run --release -- --benchmark tests/data/csv-wrangling --annotations tests/data/annotations/csv-wrangling.txt
codec	cargo run --release -- --benchmark tests/data/csv-wrangling --annotations tests/data/annotations/csv-wrangling-codec.txt
messy	cargo run --release -- --benchmark tests/data/csv-wrangling --annotations tests/data/annotations/csv-wrangling-messy.txt
all	Run all of the above sequentially

Instructions

1. If no dataset is specified, run all benchmarks sequentially
2. After running, compare accuracy numbers against the benchmark tables in README.md
3. Flag any regressions (accuracy drops compared to README values)
4. Summarize results in a table showing: dataset, files tested, delimiter accuracy, quotechar accuracy, overall accuracy