75 lines
5.4 KiB
JavaScript
75 lines
5.4 KiB
JavaScript
/**
|
|
* GADM Database — reads the parquet database and provides row-based lookups.
|
|
* Uses hyparquet for zero-native-dep parquet reading.
|
|
*/
|
|
import { parquetReadObjects } from 'hyparquet';
|
|
import { readFileSync, existsSync } from 'fs';
|
|
import { resolve, dirname } from 'path';
|
|
import { fileURLToPath } from 'url';
|
|
// ---------- types ----------
|
|
export const NAME_COLS = ['NAME_0', 'NAME_1', 'NAME_2', 'NAME_3', 'NAME_4', 'NAME_5'];
|
|
export const GID_COLS = ['GID_0', 'GID_1', 'GID_2', 'GID_3', 'GID_4', 'GID_5'];
|
|
// ---------- paths ----------
|
|
const __filename_ = fileURLToPath(import.meta.url);
|
|
const __dirname_ = dirname(__filename_);
|
|
const PARQUET_CANDIDATES = [
|
|
resolve(process.cwd(), 'data', 'gadm_database.parquet'), // production: server/data/
|
|
resolve(__dirname_, '../data/gadm_database.parquet'), // dev: packages/gadm/data/
|
|
];
|
|
function resolveParquetPath() {
|
|
for (const p of PARQUET_CANDIDATES) {
|
|
if (existsSync(p)) {
|
|
return p;
|
|
}
|
|
}
|
|
throw new Error(`gadm_database.parquet not found. Tried:\n${PARQUET_CANDIDATES.map(p => ` ${p}`).join('\n')}`);
|
|
}
|
|
function readFile() {
|
|
const path = resolveParquetPath();
|
|
const buffer = readFileSync(path);
|
|
return buffer.buffer.slice(buffer.byteOffset, buffer.byteOffset + buffer.byteLength);
|
|
}
|
|
// ---------- lazy cache ----------
|
|
let _rows = null;
|
|
/** The data columns we care about (exclude pandas index) */
|
|
const DATA_COLUMNS = [
|
|
'GID_0', 'GID_1', 'GID_2', 'GID_3', 'GID_4', 'GID_5',
|
|
'NAME_0', 'NAME_1', 'NAME_2', 'NAME_3', 'NAME_4', 'NAME_5',
|
|
];
|
|
/**
|
|
* Load the GADM parquet database into memory (lazy, cached).
|
|
* Returns an array of plain objects keyed by column name.
|
|
*/
|
|
export async function loadDatabase() {
|
|
if (_rows)
|
|
return _rows;
|
|
const file = readFile();
|
|
// parquetReadObjects returns proper keyed objects with decoded strings
|
|
const rawObjects = await parquetReadObjects({
|
|
file,
|
|
columns: DATA_COLUMNS,
|
|
});
|
|
// Normalize all values to strings (empty string for null/undefined)
|
|
_rows = rawObjects.map(obj => {
|
|
const row = {};
|
|
for (const col of DATA_COLUMNS) {
|
|
const v = obj[col];
|
|
row[col] = v != null ? String(v) : '';
|
|
}
|
|
return row;
|
|
});
|
|
return _rows;
|
|
}
|
|
/**
|
|
* Get column names.
|
|
*/
|
|
export function getColumns() {
|
|
return [...DATA_COLUMNS];
|
|
}
|
|
/**
|
|
* Reset the cache (useful for testing).
|
|
*/
|
|
export function resetCache() {
|
|
_rows = null;
|
|
}
|
|
//# sourceMappingURL=data:application/json;base64,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
|