Show HN: 30k IKEA items in flat text (huggingface.co)
OP here.
I took the unofficial IKEA US dataset (originally scraped by jeffreyszhou) and converted all 30,511 products into a flat, markdown-like protocol called CommerceTXT.
The goal: See if a flatter structure is more efficient for LLM context windows.
The results: - Size: 30k products across 632 categories. - Efficiency: The text version uses ~24% fewer tokens (3.6M saved total) compared to the equivalent minified JSON. - Structure: Files are organized in folders (e.g. /products/category/), which helps with testing hierarchical retrieval routers.
The link goes to the dataset on Hugging Face which has the full benchmarks.
Parser code is here: https://github.com/commercetxt/commercetxt
Happy to answer questions about the conversion logic!
11 comments
[ 3.7 ms ] story [ 39.1 ms ] threadOr just a handy open data set you could use to prove out the concept?
For example, Google’s indexers already use this to surface pricing data. https://developers.google.com/search/docs/appearance/structu...
These things should be put under /.well-known [1], not in the root.
[1] https://en.wikipedia.org/wiki/Well-known_URI
It's funny because it makes zero sense in the body of an initial post!
In comments replying to people downthread - maybe. But opening a top-level post with "Original Poster here" is just silly and shows a lack of respect for community etiquette.
https://hn.algolia.com/?dateRange=pastYear&page=0&prefix=tru...