DPARS service
BabelShark
Metadata intelligence for heterogeneous whole-slide imaging
Purpose
Designed as an independent, observable module.
BabelShark watches scanner outputs, establishes a stable WSI identity, extracts label information, and routes slides through fast, deep and fallback metadata paths.
Capabilities
WSI identity & canonical naming
DataMatrix and OCR-assisted metadata extraction
Configurable format-driven routing
Research-slide protection and special-case routing
This public overview intentionally excludes patient data, internal addresses, credentials and sensitive deployment details.
Service workflow
Inside the pipeline
Execution flow
01
WatchDiscover scanner output
02
IdentifyEstablish WSI identity
03
ResolveFilename → DataMatrix → OCR
04
RouteClinical / research destination
05
PersistMetadata + routing result
Platform integration
D
DagsterCoordinates the service as an observable pipeline stage.
P
PostgreSQLPersists lineage, execution state and audit evidence.
↺
Failure-awareExplicit state transitions make retries and investigation traceable.