Brain Path Query Planner
Trace a brain:// address from raw bytes to its canonical representation
A brain:// address resolves through an explicit nine-stage query planner. This tool makes the derivation concrete and shows that it is not hand-written per modality: enter any canonical address (or pick an example) and the planner parses it, then derives the plan by searching a declarative registry of transforms.
Each transform declares its precondition (the representation it consumes), its effect (what it produces), and a cost. The planner selects the transforms whose effects reach the requested :space/:dtype/:qualifiers and orders them by their preconditions, so step order is derived, not scripted - the same capability-directed assembly a SQL or federated-GraphQL planner runs over a typed catalog.
Where a matching derivative already exists in the cache (e.g. an fMRIPrep output in MNI152), the planner reuses it and skips the upstream steps - a cache hit; toggle reuse cached derivatives to see it. The view below shows the core values the planner computes - the parsed segments, the input files it gathers, the derived operator sequence with each step’s needs/gives/cost, and the registry it was drawn from. See the full Brain Path Specification for the address grammar and the staged planner.
Enter a brain:// address, or pick one of the examples: