64 lines
4.5 KiB
JavaScript
64 lines
4.5 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 } 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_);
|
|
/** Path to the parquet database (local copy from pygadm) */
|
|
const PARQUET_PATH = resolve(__dirname_, '../data/gadm_database.parquet');
|
|
// ---------- lazy cache ----------
|
|
let _rows = null;
|
|
function readFile() {
|
|
const buffer = readFileSync(PARQUET_PATH);
|
|
return buffer.buffer.slice(buffer.byteOffset, buffer.byteOffset + buffer.byteLength);
|
|
}
|
|
/** 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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
|