Canon — the Damn Filters product database

Canon is the engine behind the Damn Filters catalog. It takes raw supplier exports and turns them into one clean, structured dataset that classifies every cartridge filter, grades it for bidding and pricing, and feeds the purchase-order automation.

The pipeline

flowchart LR A[Matrixify export] --> B[Normalized database] B --> C[Classify into dimensional classes] C --> D[Grade: house × natural × anchor] D --> E[Final export to Shopify] E --> F[PO automation]

At the heart of it is the class — a fingerprint of the eight attributes that make a filter what it is: outside diameter, inside diameter, length, shape, media, top endcap, bottom endcap, and build configuration. Each class is the seed for a Damn Filters house part number.

The documents

README — what Canon is, how to run it, and where everything lives. Start here.
AGENTS — how the project is run: the locked rules, the data files, and the traps we've learned to avoid. The source of truth for anyone touching the data.
CHANGELOG — the append-only log of every pass, dated and attributed.
Commits — the complete git commit history, newest first.
Domain Model — the canonical business rules: the class definition, product-page/OEM-PN convention, and endcap/flange/build rules.
Grade & Pricing — house × natural × anchor grades and the pricing framework.
Media Taxonomy — the #12–#68 media codes (80/20, Nano, Spun Bond, exotics).
Open Questions — the live decision backlog.

How to use this handbook

New to Canon? Start with README for orientation, then AGENTS for the rules. The business documents — Domain Model, Grade & Pricing, and Media Taxonomy — carry the specifics; Open Questions is where to comment on anything that looks wrong or unresolved. CHANGELOG and Commits show what changed and when.