Connectors - EncodeBox

A connector is a YAML file that maps your source tables to FDA Sentinel CDM v8.2.2 domains (diagnosis, procedure, dispensing, enrollment, …).

Where to start

Template: connectors/template_sentinel/connector.yml
Guide: connectors/template_sentinel/README.md
Sample: connectors/postgres_local_sentinel/connector.yml

What you configure

Database: type (postgres / redshift / snowflake / bigquery / mysql) and schema
Domain blocks (top-level keys):
- diagnosis, procedure, dispensing, enrollment, demographic, …
- Table name, patient/date/code columns, code-type mappings, optional position mapping

name: my_sentinel_connector
type: claims
database:
  type: postgres
  schema: public

diagnosis:
  table: diagnosis_table
  patient_id_column: patid
  adate_column: adate
  dx_column: dx
  dx_code_type:
    type_column: dx_codetype
    type_mapping:
      "10": ["10", "ICD10"]
      "09": ["09", "9"]
    default_type: "10"

Register and query

import encodebox as eb

eb.register_connector("connectors/my_sentinel_connector/connector.yml")

db = eb.connect(
    "postgres",
    host="localhost",
    port=5433,
    database="mydb",
    username="user",
    password="pass",
    connector="my_sentinel_connector",
)

events = db.diagnosis.matching("E11", code_type="10", match="starts_with")

Registry cache

Registrations are cached locally. Inspect with eb.list_connectors().

Env / API	Effect
`ENCODEBOX_CONNECTOR_REGISTRY_PATH`	Override cache path
`ENCODEBOX_DISABLE_CONNECTOR_CACHE=1`	Disable caching
`eb.unregister_connector("name")`	Remove one
`eb.clear_connectors()`	Clear all

See Which API? for EventFrame / Cohort usage.

⌘I