sbrl-github/research-rainfallradar
1
0
Fork 0
mirror of https://github.com/sbrl/research-rainfallradar synced 2024-11-22 09:13:01 +00:00
Code Issues Projects Releases Packages Wiki Activity

wrangler bugfix: crashes; logging output

Browse source
This commit is contained in:
Starbeamrainbowlabs 2022-08-10 17:33:10 +01:00
parent b52c7f89a7
commit 231c832888
Signed by: sbrl
GPG key ID: 1BE5172E637709C2
2 changed files with 16 additions and 10 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

12
rainfallwrangler/src/lib/async/debounce.mjs → rainfallwrangler/src/lib/misc/debounce.mjs
View file

@ -11,10 +11,16 @@
* @return {function} A debounced wrapper function around the specified function. * @return {function} A debounced wrapper function around the specified function.
*/ */
export default function(fn, interval) { export default function(fn, interval) {
const time_last = 0; if(typeof interval !== "number")
throw new Error(`Error: Expected interval to be of type number, but got value of type ${typeof interval}`);
let time_last = 0;
return (...args) => { return (...args) => {
const now = new Date(); const now = new Date();
if(now - time_last > interval) // console.log(`DEBUG:debounce time since last`, now - time_last, `interval`, interval)
if(now - time_last > interval) {
time_last = now;
fn(...args); fn(...args);
}
} }
} }

14
rainfallwrangler/src/lib/record/jsonl_to_tf.mjs
View file

@ -6,7 +6,7 @@ import os from 'os';
import p_map from 'p-map'; import p_map from 'p-map';
import pretty_ms from 'pretty-ms'; import pretty_ms from 'pretty-ms';
import debounce from '../async/debounce.mjs'; import debounce from '../misc/debounce.mjs';
import py_jsonl2tfrecord from '../python/py_jsonl2tfrecord.mjs'; import py_jsonl2tfrecord from '../python/py_jsonl2tfrecord.mjs';
import log from '../../lib/io/NamespacedLog.mjs'; const l = log("jsonl2tf"); import log from '../../lib/io/NamespacedLog.mjs'; const l = log("jsonl2tf");
@ -20,12 +20,12 @@ import log from '../../lib/io/NamespacedLog.mjs'; const l = log("jsonl2tf");
export default async function(dirpath_source, dirpath_target) { export default async function(dirpath_source, dirpath_target) {
const files = await fs.promises.readdir(dirpath_source); const files = await fs.promises.readdir(dirpath_source);
let time_start = new Date(), lines_processed = 0, files_complete = 0; let time_start = new Date(), lines_processed = 0, files_processed = 0;
const update_progress_force = () => { const update_progress_force = () => {
process.stdout.write(`${files_complete}/${lines_processed} files/lines complete | ${((new Date() - time_start) / lines_processed).toFixed(3)} lines/sec | ${((files_processed / files.length)*100).toFixed(2)}% complete\r`); process.stdout.write(`${files_processed}/${lines_processed} files/lines complete | ${((new Date() - time_start) / lines_processed).toFixed(3)} lines/sec | ${((files_processed / files.length)*100).toFixed(2)}% complete\r`);
}; };
const update_progress = debounce(update_progress_force); const update_progress = debounce(update_progress_force, 2000);
await p_map(files, async (filename, i) => { await p_map(files, async (filename, i) => {
const filepath_source = path.join(dirpath_source, filename); const filepath_source = path.join(dirpath_source, filename);
@ -35,15 +35,15 @@ export default async function(dirpath_source, dirpath_target) {
let time_start = new Date(), lines_done = 0; let time_start = new Date(), lines_done = 0;
for await (let line_number of py_jsonl2tfrecord(filepath_source, filepath_dest, filepath_meta)) { for await (let line_number of py_jsonl2tfrecord(filepath_source, filepath_dest, filepath_meta)) {
lines_processed++; lines_processed++;
lines_done = line_number; lines_done++;
update_progress(); update_progress();
} }
files_complete++; files_processed++;
l.info(`DEBUG lines_done`, lines_done); l.info(`DEBUG lines_done`, lines_done);
const time_per_line = (new Date() - time_start) / (lines_done === 0 ? 1 : lines_done); const time_per_line = (new Date() - time_start) / (lines_done === 0 ? 1 : lines_done);
l.log(`converted ${filename}: ${lines_done} lines @ ${pretty_ms(time_per_line)}/line`); l.log(`converted ${filename}: ${lines_done} lines @ ${pretty_ms(time_per_line)}/line`);
}, { concurrency: os.cpus().length }); }, { concurrency: os.cpus().length });
update_progress_force(); update_progress_force();
l.log(`complete: ${lines_processed}/${files_complete} lines/files processed in ${pretty_ms(new Date() - time_start)}`); l.log(`complete: ${lines_processed}/${files_processed} lines/files processed in ${pretty_ms(new Date() - time_start)}`);
} }