@gscdump/engine-sqlite

Run typed analytics over SQLite or D1 through an async executor.

createSqliteQuerySource wraps an executor as an AnalysisQuerySource scoped to a Site. The Engine resolver builds the SQL; your executor runs it.

Install

npm install @gscdump/engine-sqlite @gscdump/engine drizzle-orm

Usage

import {
  aggCtr,
  compileSqlite,
  gsc_keywords,
  sql,
} from '@gscdump/engine-sqlite'

const queryExpr = sql`
  SELECT ${gsc_keywords.query} as query,
         SUM(${gsc_keywords.clicks}) as clicks,
         ${aggCtr(gsc_keywords)} as ctr
  FROM ${gsc_keywords}
  WHERE ${gsc_keywords.site_id} = ${siteId}
  GROUP BY ${gsc_keywords.query}
`

const { sql: compiledSql, params } = compileSqlite(queryExpr)
const { rows } = await executor(compiledSql, params, 'all')

Engine source for analyzer dispatch

import { createSqliteQuerySource } from '@gscdump/engine-sqlite'

const source = createSqliteQuerySource({
  executor, // (sql, params, mode) => { rows }
  siteId,
  regex: true, // Enable only if your executor supports REGEXP
})

Import sql from @gscdump/engine-sqlite so it uses the same Drizzle instance as the package. This avoids SQL<unknown> type mismatches between installations.

Exports

• createSqliteQuerySource({ executor, siteId, regex? }) : builds an AnalysisQuerySource.
• createSqliteInsightRunner({ executor }) : sqlite-proxy drizzle adapter.
• compileSqlite(sql) : compile to { sql, params } for any HTTP executor.
• aggClicks / aggImpressions / aggCtr / aggPosition : aggregate helpers.
• scopeFor(table, { siteId, window }) / mergeScope() : tenant scope predicates.
• gsc_pages / gsc_keywords / gsc_page_keywords / gsc_countries / gsc_devices / gsc_query_dim / schema : drizzle schema mirroring @gscdump/engine/schema.
• sqliteResolverAdapter / createSqliteResolverAdapter / probeSqliteRegex : dialect adapter for the resolver kit.
• resolveWindow (re-exported from @gscdump/engine/period).

Storage tables

The gsc_* tables remain exported for SQLite and D1 consumers. This adapter also exports createD1ManifestStore and its r2* manifest tables. Your application owns table creation, migrations, and data ingestion. The CLI uses a Parquet Store; it does not sync into these SQLite tables.

Related

• @gscdump/engine : Storage contracts + dialect-neutral resolver.
• @gscdump/analysis : Analyzer registry + dispatcher.
• @gscdump/engine/node : Node DuckDB counterpart.
• @gscdump/engine-duckdb-wasm : Browser DuckDB-WASM counterpart.

License

MIT