4.6.7 first pr (#726)

This commit is contained in:
Archer
2024-01-10 23:35:04 +08:00
committed by GitHub
parent 414b693303
commit 006ad17c6a
186 changed files with 2996 additions and 1838 deletions

View File

@@ -0,0 +1,6 @@
import path from 'path';
export const tmpFileDirPath =
process.env.NODE_ENV === 'production' ? '/app/tmp' : path.join(process.cwd(), 'tmp');
export const previewMaxCharCount = 3000;

View File

@@ -1 +0,0 @@
export const imageBaseUrl = '/api/system/img/';

View File

@@ -1,5 +1,5 @@
import { UploadImgProps } from '@fastgpt/global/common/file/api';
import { imageBaseUrl } from './constant';
import { imageBaseUrl } from '@fastgpt/global/common/file/image/constants';
import { MongoImage } from './schema';
export function getMongoImgUrl(id: string) {
@@ -8,10 +8,13 @@ export function getMongoImgUrl(id: string) {
export const maxImgSize = 1024 * 1024 * 12;
export async function uploadMongoImg({
type,
base64Img,
teamId,
expiredTime,
metadata
metadata,
shareId
}: UploadImgProps & {
teamId: string;
}) {
@@ -20,12 +23,16 @@ export async function uploadMongoImg({
}
const base64Data = base64Img.split(',')[1];
const binary = Buffer.from(base64Data, 'base64');
const { _id } = await MongoImage.create({
type,
teamId,
binary: Buffer.from(base64Data, 'base64'),
binary,
expiredTime: expiredTime,
metadata
metadata,
shareId
});
return getMongoImgUrl(String(_id));

View File

@@ -1,5 +1,7 @@
import { TeamCollectionName } from '@fastgpt/global/support/user/team/constant';
import { connectionMongo, type Model } from '../../mongo';
import { MongoImageSchemaType } from '@fastgpt/global/common/file/image/type.d';
import { mongoImageTypeMap } from '@fastgpt/global/common/file/image/constants';
const { Schema, model, models } = connectionMongo;
const ImageSchema = new Schema({
@@ -12,12 +14,18 @@ const ImageSchema = new Schema({
type: Date,
default: () => new Date()
},
binary: {
type: Buffer
},
expiredTime: {
type: Date
},
binary: {
type: Buffer
},
type: {
type: String,
enum: Object.keys(mongoImageTypeMap),
required: true
},
metadata: {
type: Object
}
@@ -25,14 +33,13 @@ const ImageSchema = new Schema({
try {
ImageSchema.index({ expiredTime: 1 }, { expireAfterSeconds: 60 });
ImageSchema.index({ type: 1 });
ImageSchema.index({ teamId: 1 });
} catch (error) {
console.log(error);
}
export const MongoImage: Model<{
teamId: string;
binary: Buffer;
metadata?: { fileId?: string };
}> = models['image'] || model('image', ImageSchema);
export const MongoImage: Model<MongoImageSchemaType> =
models['image'] || model('image', ImageSchema);
MongoImage.syncIndexes();

View File

@@ -0,0 +1,68 @@
import * as pdfjs from 'pdfjs-dist/legacy/build/pdf.mjs';
// @ts-ignore
import('pdfjs-dist/legacy/build/pdf.worker.min.mjs');
import { ReadFileParams } from './type';
type TokenType = {
str: string;
dir: string;
width: number;
height: number;
transform: number[];
fontName: string;
hasEOL: boolean;
};
export const readPdfFile = async ({ path }: ReadFileParams) => {
const readPDFPage = async (doc: any, pageNo: number) => {
const page = await doc.getPage(pageNo);
const tokenizedText = await page.getTextContent();
const viewport = page.getViewport({ scale: 1 });
const pageHeight = viewport.height;
const headerThreshold = pageHeight * 0.95;
const footerThreshold = pageHeight * 0.05;
const pageTexts: TokenType[] = tokenizedText.items.filter((token: TokenType) => {
return (
!token.transform ||
(token.transform[5] < headerThreshold && token.transform[5] > footerThreshold)
);
});
// concat empty string 'hasEOL'
for (let i = 0; i < pageTexts.length; i++) {
const item = pageTexts[i];
if (item.str === '' && pageTexts[i - 1]) {
pageTexts[i - 1].hasEOL = item.hasEOL;
pageTexts.splice(i, 1);
i--;
}
}
page.cleanup();
return pageTexts
.map((token) => {
const paragraphEnd = token.hasEOL && /([。?!.?!\n\r]|(\r\n))$/.test(token.str);
return paragraphEnd ? `${token.str}\n` : token.str;
})
.join('');
};
const loadingTask = pdfjs.getDocument(path);
const doc = await loadingTask.promise;
const pageTextPromises = [];
for (let pageNo = 1; pageNo <= doc.numPages; pageNo++) {
pageTextPromises.push(readPDFPage(doc, pageNo));
}
const pageTexts = await Promise.all(pageTextPromises);
loadingTask.destroy();
return {
rawText: pageTexts.join('')
};
};

View File

@@ -0,0 +1,18 @@
export type ReadFileParams = {
preview: boolean;
teamId: string;
path: string;
metadata?: Record<string, any>;
};
export type ReadFileResponse = {
rawText: string;
};
export type ReadFileBufferItemType = ReadFileParams & {
rawText: string;
};
declare global {
var readFileBuffers: ReadFileBufferItemType[];
}

View File

@@ -0,0 +1,50 @@
import { readPdfFile } from './pdf';
import { readDocFle } from './word';
import { ReadFileBufferItemType, ReadFileParams } from './type';
global.readFileBuffers = global.readFileBuffers || [];
const bufferMaxSize = 200;
export const pushFileReadBuffer = (params: ReadFileBufferItemType) => {
global.readFileBuffers.push(params);
if (global.readFileBuffers.length > bufferMaxSize) {
global.readFileBuffers.shift();
}
};
export const getReadFileBuffer = ({ path, teamId }: ReadFileParams) =>
global.readFileBuffers.find((item) => item.path === path && item.teamId === teamId);
export const readFileContent = async (params: ReadFileParams) => {
const { path } = params;
const buffer = getReadFileBuffer(params);
if (buffer) {
return buffer;
}
const extension = path?.split('.')?.pop()?.toLowerCase() || '';
const { rawText } = await (async () => {
switch (extension) {
case 'pdf':
return readPdfFile(params);
case 'docx':
return readDocFle(params);
default:
return Promise.reject('Only support .pdf, .docx');
}
})();
pushFileReadBuffer({
...params,
rawText
});
return {
...params,
rawText
};
};

View File

@@ -0,0 +1,22 @@
import mammoth from 'mammoth';
import { htmlToMarkdown } from '../../string/markdown';
import { ReadFileParams } from './type';
/**
* read docx to markdown
*/
export const readDocFle = async ({ path, metadata = {} }: ReadFileParams) => {
try {
const { value: html } = await mammoth.convertToHtml({
path
});
const md = await htmlToMarkdown(html);
return {
rawText: md
};
} catch (error) {
console.log('error doc read:', error);
return Promise.reject('Can not read doc file, please convert to PDF');
}
};

View File

@@ -1,11 +1,9 @@
import type { NextApiRequest, NextApiResponse } from 'next';
import { customAlphabet } from 'nanoid';
import multer from 'multer';
import path from 'path';
import { BucketNameEnum, bucketNameMap } from '@fastgpt/global/common/file/constants';
import fs from 'fs';
const nanoid = customAlphabet('1234567890abcdef', 12);
import { getNanoid } from '@fastgpt/global/common/string/tools';
import { tmpFileDirPath } from './constants';
type FileType = {
fieldname: string;
@@ -17,7 +15,9 @@ type FileType = {
size: number;
};
export function getUploadModel({ maxSize = 500 }: { maxSize?: number }) {
const expiredTime = 30 * 60 * 1000;
export const getUploadModel = ({ maxSize = 500 }: { maxSize?: number }) => {
maxSize *= 1024 * 1024;
class UploadModel {
uploader = multer({
@@ -26,9 +26,12 @@ export function getUploadModel({ maxSize = 500 }: { maxSize?: number }) {
},
preservePath: true,
storage: multer.diskStorage({
filename: (_req, file, cb) => {
// destination: (_req, _file, cb) => {
// cb(null, tmpFileDirPath);
// },
filename: async (req, file, cb) => {
const { ext } = path.parse(decodeURIComponent(file.originalname));
cb(null, nanoid() + ext);
cb(null, `${Date.now() + expiredTime}-${getNanoid(32)}${ext}`);
}
})
}).any();
@@ -75,14 +78,4 @@ export function getUploadModel({ maxSize = 500 }: { maxSize?: number }) {
}
return new UploadModel();
}
export const removeFilesByPaths = (paths: string[]) => {
paths.forEach((path) => {
fs.unlink(path, (err) => {
if (err) {
console.error(err);
}
});
});
};

View File

@@ -0,0 +1,33 @@
import fs from 'fs';
import { tmpFileDirPath } from './constants';
export const removeFilesByPaths = (paths: string[]) => {
paths.forEach((path) => {
fs.unlink(path, (err) => {
if (err) {
console.error(err);
}
});
});
};
/* cron job. check expired tmp files */
export const checkExpiredTmpFiles = () => {
// get all file name
const files = fs.readdirSync(tmpFileDirPath).map((name) => {
const timestampStr = name.split('-')[0];
const expiredTimestamp = timestampStr ? Number(timestampStr) : 0;
return {
filename: name,
expiredTimestamp,
path: `${tmpFileDirPath}/${name}`
};
});
// count expiredFiles
const expiredFiles = files.filter((item) => item.expiredTimestamp < Date.now());
// remove expiredFiles
removeFilesByPaths(expiredFiles.map((item) => item.path));
};

View File

@@ -50,8 +50,11 @@ export const cheerioToHtml = ({
.get()
.join('\n');
const title = $('head title').text() || $('h1:first').text() || fetchUrl;
return {
html,
title,
usedSelector
};
};
@@ -70,7 +73,7 @@ export const urlsFetch = async ({
});
const $ = cheerio.load(fetchRes.data);
const { html, usedSelector } = cheerioToHtml({
const { title, html, usedSelector } = cheerioToHtml({
fetchUrl: url,
$,
selector
@@ -79,6 +82,7 @@ export const urlsFetch = async ({
return {
url,
title,
content: md,
selector: usedSelector
};
@@ -87,6 +91,7 @@ export const urlsFetch = async ({
return {
url,
title: '',
content: '',
selector: ''
};

View File

@@ -15,7 +15,9 @@ export const htmlToMarkdown = (html?: string | null) =>
worker.on('message', (md: string) => {
worker.terminate();
resolve(simpleMarkdownText(md));
let rawText = simpleMarkdownText(md);
resolve(rawText);
});
worker.on('error', (err) => {
worker.terminate();

View File

@@ -0,0 +1,6 @@
import nodeCron from 'node-cron';
export const setCron = (time: string, cb: () => void) => {
// second minute hour day month week
return nodeCron.schedule(time, cb);
};

View File

@@ -49,6 +49,7 @@ export const addLog = {
},
error(msg: string, error?: any) {
this.log('error', msg, {
message: error?.message,
stack: error?.stack,
...(error?.config && {
config: {

View File

@@ -2,6 +2,8 @@ export type DeleteDatasetVectorProps = {
id?: string;
datasetIds?: string[];
collectionIds?: string[];
collectionId?: string;
dataIds?: string[];
};

View File

@@ -101,14 +101,19 @@ export const deleteDatasetDataVector = async (
retry?: number;
}
): Promise<any> => {
const { id, datasetIds, collectionIds, dataIds, retry = 2 } = props;
const { id, datasetIds, collectionIds, collectionId, dataIds, retry = 2 } = props;
const where = await (() => {
if (id) return `id=${id}`;
if (datasetIds) return `dataset_id IN (${datasetIds.map((id) => `'${String(id)}'`).join(',')})`;
if (collectionIds)
if (collectionIds) {
return `collection_id IN (${collectionIds.map((id) => `'${String(id)}'`).join(',')})`;
if (dataIds) return `data_id IN (${dataIds.map((id) => `'${String(id)}'`).join(',')})`;
}
if (collectionId && dataIds) {
return `collection_id='${String(collectionId)}' and data_id IN (${dataIds
.map((id) => `'${String(id)}'`)
.join(',')})`;
}
return Promise.reject('deleteDatasetData: no where');
})();