FastEHR.database.collector¶

Classes¶

SQLiteDataCollector

A class to interface with an SQLite database to collect and collate

Module Contents¶

class FastEHR.database.collector.SQLiteDataCollector(db_path: str)¶

A class to interface with an SQLite database to collect and collate patient records.

This class provides functionality for extracting structured patient data, aggregating medical events, and computing metadata for pre-processing from an SQLite database.

Inherits from:

Static - Handles static patient data, such as birth
year and ethnicity.
Diagnoses - Handles diagnosis-related records.
Measurements - Handles event-based measurements, which may
optionally include an associated value.

Attributes¶

db_pathstr: Path to the SQLite database file.
connectionsqlite3.Connection.: SQLite connection object, initialized when connect() is called.
cursorsqlite3.Cursor.: Cursor for executing SQL queries.

Methods¶

connect()

Establish the SQLite database connection.

disconnect()

Close the SQLite database connection.

_extract_distinct()

Extracts distinct values of a given column across multiple tables.

_extract_AGG()

Performs grouped aggregations over tables.

_t_digest_values()

Uses the t-digest algorithm to approximate percentiles of a given measurement. <https://github.com/tdunning/t-digest>

_generate_lazy_by_distinct()

Generates Polars LazyFrames for distinct patient or practice: identifiers.

_collate_lazy_tables()

Merges static and dynamic patient records into a single LazyFrame.

get_meta_information()

Collects metadata from the SQLite database, including distributions: of diagnoses and measurements.

Initializes the SQLiteDataCollector.

Parameters¶

db_pathstr: Path to the SQLite database file.

db_path¶

connection = None¶

cursor = None¶

connect()¶

Establishes a connection to the SQLite database.

If the connection is already established, this method does nothing.

Raises¶

sqlite3.Error: If an error occurs while connecting to the database.

disconnect()¶

Closes the SQLite database connection.

This method ensures that both the connection and cursor are properly: closed.

get_meta_information(practice_ids: list | None = None, static: bool = True, diagnoses: bool = True, measurement: bool = True) → dict¶

Collects metadata from the SQLite database, such as distributions of diagnoses and measurements.

Parameters¶

practice_idslist, optional: List of practice IDs to filter metadata collection (default is None).
staticbool, optional: Whether to collect static patient information (default is True).
diagnosesbool, optional: Whether to collect diagnosis-related metadata (default is True).
measurementbool, optional: Whether to collect measurement-related metadata (default is True).

Returns¶

dict: A dictionary containing metadata tables.