Skip to content

Application Configuration

Application Settings

The Record Linker application is configured via a pydantic-settings configuration class. The settings are loaded from environment variables or a .env file, allowing for easy customization without modifying the code. When both an environment variable and a .env file are present, the environment variable takes precedence.

💡 For information on algorithm configuration, see the Algorithm Configuration section.

DB_URI (Required)

The URI for the application database.

Example: postgresql+psycopg2://user:password@localhost/dbname

DB_TABLE_PREFIX (Optional)

Prefix applied to all database tables, useful for namespace separation.

Docker Default: ""

Development Default: ""

TEST_DB_URI (Optional)

The URI for the application database used when running tests.

Docker Default: sqlite:///testdb.sqlite3

Development Default: sqlite:///testdb.sqlite3

CONNECTION_POOL_SIZE (Optional)

Number of connections in the MPI database connection pool.

Docker Default: 5

Development Default: 5

CONNECTION_POOL_MAX_OVERFLOW (Optional)

Maximum number of overflow connections allowed in the connection pool.

Docker Default: 10

Development Default: 10

LOG_CONFIG (Optional)

Path to a logging configuration file, used to configure logging settings on startup. A value of an empty string will cause logging to use default settings.

Docker Default: "assets/production_log_config.json"

Development Default: ""

SPLUNK_URI (Optional)

URI for the Splunk HTTP Event Collector (HEC) endpoint. When set, logs will be sent to the configured Splunk instance for analysis. The format is splunkhec://<token>@<host>:<port>?index=<index>&proto=<protocol>&source=<source>

Docker Default: ""

Development Default: ""

AUTO_MIGRATE (Optional)

Apply all pending database migrations when the application starts. A fake migration will happen on the first application if the database is missing the alembic_version table. In this case, the record linker and migration tables will be auto created and the alembic_version table will be initialized with the latest migration version. On all subsequent application starts, the database will be checked for a pending migration and applied if necessary.

Docker Default: true

Development Default: true

INITIAL_ALGORITHMS (Optional)

Path to a JSON file with initial algorithms to load if the algorithms table is empty. NOTE: This will only be used if AUTO_MIGRATE is true.

Docker Default: assets/initial_algorithms.json

Development Default: assets/initial_algorithms.json

API_ROOT_PATH (Optional)

Root path from which the Record Linker API will be exposed.

Docker Default: /api

Development Default: /api

TUNING_ENABLED (Optional)

Whether to enable the tuning API endpoints.

Docker Default: false

Development Default: false

TUNING_TRUE_MATCH_PAIRS (Optional)

Number of true match pairs to use for tuning.

Docker Default: 10000

Development Default: 10000

TUNING_NON_MATCH_PAIRS (Optional)

Number of non match pairs to use for tuning.

Docker Default: 100000

Development Default: 100000

TUNING_NON_MATCH_SAMPLE (Optional)

Number of patient records to sample from when selecting non match pairs.

Docker Default: 250000

Development Default: 250000

TUNING_JOB_TIMEOUT (Optional)

Number of seconds to wait for a tuning job to finish before canceling it.

Docker Default: 3600

Development Default: 3600

Database Options

The DB_URI and TEST_DB_URI settings can be configured to connect to a compatible SQLAlchemy database. By default, the following database drivers are installed allowing for connections to:

  • sqlite3 (sqlite)

    Example: sqlite:///db.sqlite3

  • psycopg2 (postgresql)

    Example: postgresql+psycopg2://user:password@localhost/dbname

  • pymysql (mysql / mariadb)

    Example: mysql+pymysql://user:password@localhost/dbname

  • pyodbc (sql server)

    Example: mssql+pyodbc://user:password@localhost/dbname?driver=ODBC+Driver+18+for+SQL+Server&TrustServerCertificate=yes

Docker Configuration

In addition to the above application settings that can be configured via environment variables, the Dockerfile can be built with the build arguments to customize the container.

ENVIRONMENT (Optional)

The environment to install python packages for. Options are dev or prod, see the pyproject.toml file for details on the list of packages installed for each environment.

Default: prod

PORT (Optional)

The port the application will listen on.

Default: 8080

USE_MSSQL (Optional)

Whether to install the pyodbc package for connecting to a Microsoft SQL Server database. This is required if connecting to a Microsoft SQL Server database.

Default: true

USE_OTEL (Optional)

Whether to install the opentelemetry-instrumentation package for tracing.

Default: false