Metadata completeness profile v0.1
Every loaded project now produces a typed metadata report for three separate questions:
- Analysis: can the declared computation be traced to identified sessions, fingerprinted inputs, and an explicit event factor?
- NWB: is the metadata required by the current NWB writer present?
- Publication/reuse: does the project satisfy FiberPhotometry's minimal v0.1 descriptive profile?
The last status is not certification against a journal, repository, FAIR rubric, or ontology. It is an actionable minimum intended to prevent obviously anonymous or uninterpretable exports.
Configuration
Metadata are deliberately open-ended:
[metadata]
experimenters = ["A. Scientist"]
institution = "Example University"
lab = "Example Lab"
experiment_description = "Cue-evoked fluorescence in a within-animal design"
protocol = "https://doi.org/example"
species = "Mus musculus"
indicator = "GCaMP8m"
brain_regions = ["dorsomedial striatum"]
acquisition_system = "TDT Synapse"
data_license = "CC-BY-4.0"
# Lab-specific fields are retained and reported as unrecognized, not rejected.
implant_batch = "2026-07-A"
The recognized publication fields are experimenters, institution, lab, experiment description, protocol, species, indicator, brain regions, acquisition system, and data license. Session start times remain session-level fields because they can vary and NWB requires a timezone-aware value.
Output
fiberphotometry inspect project.toml includes metadata_completeness in its
preflight JSON. A run additionally writes metadata.json, fingerprints it in the
manifest, and embeds the same assessment in each exported NWB file.
Each check has a stable code, target(s), status, scope, and remediation. Each
target reports ready or incomplete plus the exact missing check codes. Missing
publication metadata does not silently block an otherwise declared exploratory
analysis; downstream release policies can choose to enforce the readiness status.
The openness and versioning policy is recorded in SDR-0008.