Skip to main content

Storage: SQLite / MySQL / PostgreSQL

ByJG\TextClassifier\Storage\Rdbms persists the word list to a relational database using byjg/micro-orm. It supports SQLite, MySQL, and PostgreSQL via a URI connection string.

Setup​

SQLite​

use ByJG\TextClassifier\Degenerator\ConfigDegenerator;
use ByJG\TextClassifier\Degenerator\StandardDegenerator;
use ByJG\TextClassifier\Storage\Rdbms;
use ByJG\Util\Uri;

$storage = new Rdbms(
    new Uri('sqlite:///var/data/spam.db'),
    new StandardDegenerator(new ConfigDegenerator())
);

// Run once to create the schema
$storage->createDatabase();

MySQL​

$storage = new Rdbms(
    new Uri('mysql://user:password@localhost/mydb'),
    new StandardDegenerator(new ConfigDegenerator())
);
$storage->createDatabase();

PostgreSQL​

$storage = new Rdbms(
    new Uri('pgsql://user:password@localhost/mydb'),
    new StandardDegenerator(new ConfigDegenerator())
);
$storage->createDatabase();

createDatabase()​

createDatabase() uses byjg/migration to apply the schema migrations located in the db/ directory of this package. It calls reset() internally, which means:

  • Drops and recreates the migration version table
  • Runs all up migration scripts from scratch

Call this once when setting up a new database. Do not call it on a database that already contains training data — it will wipe the existing data.

URI format​

scheme://[user[:password]@]host[:port]/database
SchemeDriver
sqlite://SQLite (path after // is the file path)
mysql://MySQL / MariaDB
pgsql://PostgreSQL

Database table​

The tc_wordlist table stores one row per unique token:

CREATE TABLE tc_wordlist (
    token      VARCHAR(255) NOT NULL,
    count_ham  INTEGER DEFAULT NULL,
    count_spam INTEGER DEFAULT NULL,
    PRIMARY KEY (token)
);

Two internal rows are seeded automatically:

tokenpurpose
tc*dbversionSchema version (value: 3)
tc*textsTotal ham/spam text count

When to use this backend​

  • Production deployments requiring durability
  • Shared access across multiple processes or servers
  • Large training sets where memory footprint matters
  • When you already have a relational database in your infrastructure