External dataset (#1497)

* perf: read rawText and chunk code

* perf: read raw text

* perf: read rawtext

* perf: token count

* log
This commit is contained in:
Archer
2024-05-16 11:47:53 +08:00
committed by GitHub
parent d5073f98ab
commit c6d9b15897
36 changed files with 531 additions and 267 deletions

View File

@@ -77,9 +77,8 @@ export const urlsFetch = async ({
$,
selector
});
console.log('html====', html);
const md = await htmlToMarkdown(html);
console.log('html====', md);
return {
url,

View File

@@ -12,27 +12,34 @@ import { getNanoid } from '@fastgpt/global/common/string/tools';
import { addLog } from '../../system/log';
export const getTiktokenWorker = () => {
if (global.tiktokenWorker) {
return global.tiktokenWorker;
const maxWorkers = global.systemEnv?.tokenWorkers || 20;
if (!global.tiktokenWorkers) {
global.tiktokenWorkers = [];
}
if (global.tiktokenWorkers.length >= maxWorkers) {
return global.tiktokenWorkers[Math.floor(Math.random() * global.tiktokenWorkers.length)];
}
const worker = getWorker(WorkerNameEnum.countGptMessagesTokens);
const i = global.tiktokenWorkers.push({
index: global.tiktokenWorkers.length,
worker,
callbackMap: {}
});
worker.on('message', ({ id, data }: { id: string; data: number }) => {
const callback = global.tiktokenWorker?.callbackMap?.[id];
const callback = global.tiktokenWorkers[i - 1]?.callbackMap?.[id];
if (callback) {
callback?.(data);
delete global.tiktokenWorker.callbackMap[id];
delete global.tiktokenWorkers[i - 1].callbackMap[id];
}
});
global.tiktokenWorker = {
worker,
callbackMap: {}
};
return global.tiktokenWorker;
return global.tiktokenWorkers[i - 1];
};
export const countGptMessagesTokens = (
@@ -44,20 +51,29 @@ export const countGptMessagesTokens = (
const start = Date.now();
const { worker, callbackMap } = getTiktokenWorker();
const id = getNanoid();
const timer = setTimeout(() => {
resolve(0);
console.log('Count token Time out');
resolve(
messages.reduce((sum, item) => {
if (item.content) {
return sum + item.content.length * 0.5;
}
return sum;
}, 0)
);
delete callbackMap[id];
}, 300);
}, 60000);
callbackMap[id] = (data) => {
// 检测是否有内存泄漏
addLog.info(`Count token time: ${Date.now() - start}, token: ${data}`);
// console.log(process.memoryUsage());
resolve(data);
clearTimeout(timer);
// 检测是否有内存泄漏
// addLog.info(`Count token time: ${Date.now() - start}, token: ${data}`);
// console.log(process.memoryUsage());
};
worker.postMessage({