Compare commits

...

7 Commits

Author SHA1 Message Date
e87e7010ec improvements
* chance to continue on mismatch url for redirectTab
* support empty field selectors
* add Extractor.results()
* add ExtractResult.walk(), ExtractResult.visit()
* add ! directive to click elements
* code optimize
2021-04-20 14:20:05 +08:00
108ebb835f fix task running state 2021-04-20 12:01:10 +08:00
e0b0a5e986 add timeout for messaging 2021-04-20 12:00:59 +08:00
9cd25e3c1d update url 2021-04-19 15:58:04 +08:00
7827d385bd refactor 2020-06-16 14:45:36 +08:00
ade0670415 update readme 2020-01-17 11:01:13 +08:00
63aec616b1 code optimize 2020-01-17 09:38:40 +08:00
14 changed files with 245 additions and 75 deletions

View File

@ -17,6 +17,8 @@ Extract current page
```js
$('.item', ['a', 'a@href']);
new Extractor().task('.item', ['a', 'a@href']).start();
// fieldSelectors can be empty strings if items have no child to select
new Extractor().task('.item a', ['', '@href']).start();
```
> `$(...args)` is the short form of `new Extractor().task(...args).start();`, which is introduced later.
@ -67,7 +69,9 @@ job = new Extractor().task('.search-list-item', ['a@href'], ["http://sample.com/
job.stop();
```
## Extract Attributes.
> Next time you call `job.start();`, the job will continues from where it stopped.
## Extract Attributes
e.g.: link text and target (use 'selector@attribute')
@ -75,6 +79,14 @@ e.g.: link text and target (use 'selector@attribute')
new Extractor().task('.item', ['a', 'a@href']).start();
```
## Click Selected Elements
The following clicks selected links and extracts link `text` and `href`
```js
new Extractor().task('.item', ['!a', 'a@href']).start();
```
## Advanced Usage
### Use Task Chain.
@ -185,6 +197,43 @@ e.start();
> The uploaded state will be cleaned in 30 seconds, if you don't load it.
## Watch Mode
Watch mode tries to exract data from every page you visit **in current window**.
```js
e = new Extractor();
e.task('.search-list-item', ['a@href'], ["http://sample.com/abc"])
.task('list-item', ["a.title", "p.content"]);
e.watch(1); // start watching for first task
```
To stop watching, you can either `close current window`, or:
```js
e.stop();
```
## Results Operation
To get the results of a task:
```js
let results = job.results(0);
```
Visit URLs (if any) in the results one by one:
```js
results.visit();
```
Walk through all results one by one:
```js
results.walk((row,col,value)=>{console.log(value)});
```
## Developpment
Clone this project and execute:

View File

@ -1,6 +1,6 @@
import { Actions, Request } from "../common";
import { sendMessage, ResponseChecker } from "./messaging";
import { logger } from "./logger";
import { logger } from "../common/logger";
/**
* redirect tab to url.
@ -8,25 +8,27 @@ import { logger } from "./logger";
* @param {string} url target URL
* @returns {Promise<string[]>} a promise of target URL
*/
export function redirectTab(tab: chrome.tabs.Tab, url: string) {
export function redirectTab(tab: chrome.tabs.Tab, url: string, check?: boolean) {
return queryUrl(tab).then(u => {
if (url !== u) {
let req: Request = {
action: Actions.GOTO_URL,
url: url
}
let checker: ResponseChecker<string> = async (r, err, tryCount): Promise<string> => {
let checker: ResponseChecker<string> = !check ? undefined : async (r, err, tryCount): Promise<string> => {
let queryErr: any;
let newURL = await queryUrl(tab).catch(e => queryErr = e);
if (queryErr) {
return Promise.reject(queryErr);
throw queryErr;
}
if (newURL == url) return url;
if (
tryCount % 1 == 0 &&
!confirm('Cannot navigate to target url. \nPress OK to continue, Cancel to stop.')
confirm(`Cannot navigate to target url.
expected: ${url}\n
actual: ${newURL}\n
Press OK to continue, Cancel to retry. Close the tab to stop`)
) {
return Promise.reject("Tasks stopped by user.");
return newURL;
}
return undefined;
}
@ -82,7 +84,7 @@ export async function ping(tab, count = 1) {
let checker: ResponseChecker<string> = (r, e, c) =>
r.result == "pong" ? r.result : undefined;
let pong = await sendMessage<string>(tab, req, 'Check tab availability...', checker, 1000, count).catch(() => { });
let pong = await sendMessage<string>(tab, req, 'Check tab availability...', checker, 1000, 1000, count).catch(() => { });
return pong == "pong";
}
@ -111,7 +113,7 @@ export function scrollToBottom(tab: chrome.tabs.Tab) {
return sendMessage(tab, req, 'Scroll to page bottom...');
}
export async function createTab(url: string, active: boolean) {
export async function createTab(url: string, active: boolean): Promise<chrome.tabs.Tab> {
return new Promise((resolve, reject) => {
findIncognitoWindow().then(
incognitoWindow => {
@ -179,7 +181,7 @@ export async function CreateIncognitoWindow() {
});
}
export async function getActiveTab(currentWindow: boolean) {
export async function getActiveTab(currentWindow: boolean): Promise<chrome.tabs.Tab> {
return new Promise((resolve, reject) => {
chrome.tabs.query({
active: true,
@ -190,7 +192,7 @@ export async function getActiveTab(currentWindow: boolean) {
})
}
export async function getTabByID(id: number) {
export async function getTabByID(id: number): Promise<chrome.tabs.Tab> {
return new Promise((resolve, reject) => {
chrome.tabs.get(id, function (tab) {
chrome.runtime.lastError;

View File

@ -1,4 +1,4 @@
import { logger } from "./logger";
import { logger } from "../common/logger";
import { Actions } from "../common";
import { messageSubscribers } from "./messaging";

View File

@ -1,9 +1,9 @@
import { Task } from "./task";
import { saveFile } from "./tools";
import { createTab, getActiveTab, ping } from "./actions";
import { ExtractResult } from "./result";
import { logger } from "./logger";
import { parseUrls, saveFile } from "./tools";
import { createTab, getActiveTab, ping, redirectTab } from "./actions";
import { logger } from "../common/logger";
import { caches } from "./caches";
import { ExtractResult } from "./result";
export class Extractor {
private _tasks: Task[] = [];
@ -12,6 +12,14 @@ export class Extractor {
constructor(options?) {
if (options) this._options = options;
}
static async ping(count: number = 1) {
let tab = await getActiveTab(true) || await getActiveTab(false);
let succ = await ping(tab, count);
if (!succ) {
logger.error('Cannot contact with active tab.');
return;
}
}
/**
* Save current state, in case we restore it later.
*/
@ -42,6 +50,17 @@ export class Extractor {
this._tasks.push(new Task(this._options, ...args));
return this;
}
/**
* Add a task to Extractor. \n
* One Extractor could has multiple tasks, which orgnized in a task chian.
* If url arguments not given within later tasks, they will use previous task result as input (target url list).
* @param {...any} args itemsSelector, fieldSelectors, and more args to specify target urls.
*/
results(id?: number): ExtractResult {
id = this._checkTaskId(id);
if (id < 0) return;
return this._tasks[id].results;
}
/**
* Clear tasks and task caches.
*/
@ -113,7 +132,7 @@ export class Extractor {
if (i < from) return;
if (i > 0) {
let prevTask = this._tasks[i - 1];
return task.execute(tab, new ExtractResult(prevTask.results));
return task.execute(tab, prevTask.results);
}
return task.execute(tab);
});
@ -135,22 +154,22 @@ export class Extractor {
let id = this._checkTaskId(taskid, this._tasks.length - 1);
if (id < 0) return;
let results = this._tasks[id].results
if (!results.length) {
let count = results.data.length
if (!count) {
logger.info(`No result for task #${id}. Forget to call ".start()"?`);
return;
}
results.unshift(this._tasks[id].fieldSelectors);
let exResults = new ExtractResult(results);
results.header = this._tasks[id].fieldSelectors;
let msg = `
Please confirm to download (${results.length - 1} items)
Please confirm to download (${count} items)
${exResults.toString(50) || "- Empty -"}
${results.toString(50) || "- Empty -"}
`.trim();
if (confirm(msg)) {
saveFile(exResults.toString(), "text/csv");
saveFile(results.toString(), "text/csv");
}
}
_checkTaskId(id: number, defaultId?: number) {
private _checkTaskId(id: number, defaultId?: number) {
if (!this._tasks.length) {
logger.info("No task found.");
return -1;

View File

@ -1,6 +1,6 @@
import { Request, Actions, Response } from "../common";
import { getTabByID } from "./actions";
import { logger } from "./logger";
import { logger } from "../common/logger";
export type ResponseCheckerSync<T> = (r: Response<T>, err: chrome.runtime.LastError, count: number) => T;
@ -21,12 +21,14 @@ export type ResponseChecker<T> = ResponseCheckerSync<T> | ResponseCheckerAsync<T
*/
export function sendMessage<T>(
tab: chrome.tabs.Tab,
req,
req: Request,
log?: string,
dataChecker?: ResponseChecker<T>,
timeout?: number,
interval?: number,
limit?: number
) {
timeout = timeout || 10;
interval = interval || 500;
limit = isNaN(limit) ? 0 : limit;
let count = 0;
@ -47,43 +49,23 @@ export function sendMessage<T>(
return;
}
count++;
let timeout = setTimeout(() => { reject(`${Actions[req.action]} requset timeout after ${timeout}s`) }, 10000);
chrome.tabs.sendMessage(tab.id, req, async (r: Response<T>) => {
clearTimeout(timeout);
// check error but do nothing until dataChecker.
let err = chrome.runtime.lastError;
let result: T;
// r could be undefined if the content script is interrupted.
if (r) {
result = r.result;
if (dataChecker) {
let pms: T | Promise<T>;
try {
pms = dataChecker(r, err, count);
} catch (error) {
reject(error);
return;
}
// don't catch if it's not a Promise
if (pms instanceof Promise) {
let checkerError: any;
pms = pms.catch(e => checkerError = e);
result = await pms;
if (checkerError) {
reject(checkerError);
return;
}
} else {
result = pms;
}
}
let [result, error] = await checkResponse(dataChecker, r, err, count);
if (error) {
reject(error);
return;
}
let flag = result !== undefined && result !== null;
let flag = result !== undefined;
if (log) logger.info(log, flag ? '(OK)' : '(failed)');
if (flag) {
resolve(result);
} else {
setTimeout(() => {
logger.debug('Invalid response', r, 'retry...');
loop();
}, interval);
}
@ -92,6 +74,39 @@ export function sendMessage<T>(
});
}
async function checkResponse<T>(
dataChecker: ResponseChecker<T>,
response: Response<T>,
error: chrome.runtime.LastError,
tryCount: number
): Promise<[T, string]> {
// response could be undefined if the content script is interrupted.
// don't check, tell sendMessage to retry.
if (!response) return [undefined, undefined];
if (!dataChecker) {
return [response.result, response.error];
}
let result: T;
let pms: T | Promise<T>;
try {
pms = dataChecker(response, error, tryCount);
} catch (err) {
return [undefined, err];
}
// don't catch if it's not a Promise
if (pms instanceof Promise) {
let checkerError: any;
pms = pms.catch(e => checkerError = e);
result = await pms;
if (checkerError) {
return [undefined, checkerError];
}
} else {
result = pms;
}
return [result, undefined];
}
export type ActionSubscriberSync = (request: Request, sender: chrome.runtime.MessageSender, sendResponse: (response?: any) => void) => void;
export type ActionSubscriberAsync = (request: Request, sender: chrome.runtime.MessageSender, sendResponse: (response?: any) => void) => Promise<void>;
export type ActionSubscriber = ActionSubscriberSync | ActionSubscriberAsync;

View File

@ -1,6 +1,11 @@
import { logger } from "../common/logger";
import { getActiveTab, ping, redirectTab } from "./actions";
import { parseUrls } from "./tools";
export class ExtractResult {
private _header: string[];
private _data: string[][] = [];
constructor(data) {
constructor(data: string[][]) {
this._data = data || [];
}
row(index: number): string[] {
@ -14,11 +19,17 @@ export class ExtractResult {
squash(): string[] {
return this._data.reduce((p, c) => p.concat(c), []);
}
set header(h: string[]) {
this._header = h
}
get data(): string[][] {
return this._data;
}
toString(rowsCount: number = 0): string {
let data = rowsCount > 0 ? this._data.slice(0, rowsCount) : this._data;
if (this._header && this._header.length) {
data.unshift(this._header);
}
return data.slice().reduce(
(csv, lineCells) => {
if (!lineCells || !lineCells.length) {
@ -26,6 +37,7 @@ export class ExtractResult {
}
let line = lineCells.reduce(
(lineText, cell, idx) => {
cell = cell || "";
cell = '"' + cell.trim().replace(/"/g, '""') + '"';
return lineText + cell + (idx == lineCells.length - 1 ? "" : ",")
}, "");
@ -34,4 +46,40 @@ export class ExtractResult {
""
);
}
async walk(fn: (row: number, col: number, value: string) => void) {
let pms = Promise.resolve(null);
for (let i = 0; i < this._data.length; i++) {
let cells = this._data[i];
for (let j = 0; j < cells.length; j++) {
let row = i;
let col = j;
let value = cells[j];
pms = pms.then(
() => fn(row, col, value)
)
}
}
return pms.catch(err => {
logger.error(err);
});
}
async visit() {
let urls = parseUrls(this);
let tab = await getActiveTab(true) || await getActiveTab(false);
let succ = await ping(tab);
if (!succ) {
logger.error('Cannot contact with active tab.');
return;
}
return urls.reduce(
(pms, url: string, i: number) => {
return pms.then(
async () => {
return redirectTab(tab, url, false);
});
}, Promise.resolve<void>(undefined)
).catch(err => {
logger.error(err);
});
}
}

View File

@ -20,7 +20,7 @@ function(itemsSelector:string, fieldSelectors:string[], urls:string[]);
$(".item", ["a", "a@href"]);
## See Detailed Help:
https://git.jebbs.co/jebbs/data-extracter-extesion
https://git.qjebbs.com/jebbs/data-extracter-extesion
`.trim();
export function testArgs(...args: any) {

View File

@ -4,7 +4,7 @@ import { testArgs, signitures } from "./signiture";
import { ExtractResult } from "./result";
import { messageSubscribers, ActionSubscriber } from "./messaging";
import { Actions } from "../common";
import { logger } from "./logger";
import { logger } from "../common/logger";
export class Task {
private _data: { [key: string]: string[][] } = {};
@ -41,10 +41,11 @@ export class Task {
get urls(): string[] {
return this._urls;
}
get results(): string[][] {
return this._data_keys.reduce((p, c) => {
get results(): ExtractResult {
let rs: string[][] = this._data_keys.reduce((p, c) => {
return p.concat(this._data[c]);
}, []);
return new ExtractResult(rs);
}
get fieldSelectors(): string[] {
return this._fieldSelectors;
@ -87,7 +88,7 @@ export class Task {
logger.info(`Watcher #${taskID} starts.`);
let pm = this.makeOptionalTasks(sender.tab);
return pm.then(
() => extractTabData(sender.tab, this._itemsSelector, this._fieldSelectors, sender.tab.url, false)
() => extractTabData(sender.tab, this._itemsSelector, this._fieldSelectors, sender.tab.url, true)
).then(
results => {
if (results && results.length) {
@ -103,15 +104,26 @@ export class Task {
messageSubscribers.addListener(Actions.REPORT_NEW_PAGE, listener);
}
async execute(tab: chrome.tabs.Tab, upstreamData?: ExtractResult): Promise<void> {
if (!tab) return Promise.reject("No tab to execute the task.");
if (this._running) return Promise.reject("The task is running. Please wait...");
if (!tab) throw "No tab to execute the task.";
if (this._running) throw "The task is running. Please wait...";
this._running = true;
let urls = this._urls
if (!urls.length) {
if (upstreamData) {
urls = parseUrls(upstreamData);
} else {
urls = [await queryUrl(tab)];
let tabURL: string;
await queryUrl(tab)
.then(u => {
tabURL = u;
})
.catch(() => {
e => {
this._running = false;
return Promise.reject(e);
}
});
urls = [tabURL];
}
}
return urls.reduce((p, url, i) => p.then(
@ -134,13 +146,13 @@ export class Task {
if (results && results.length) {
let lastURL = urls[urls.length - 1];
this.saveResult(results, lastURL);
this._running = false;
}
this._running = false;
}
).catch(
e => {
this._running = false;
return Promise.reject(e);
throw e;
}
);
}
@ -152,7 +164,7 @@ export class Task {
return pm;
}
private runningCheck(fn: () => Promise<any>): Promise<any> {
if (!this._running) return Promise.reject("The task is stopped by user.");
if (!this._running) throw "The task is stopped by user.";
return fn();
}
private saveResult(results, key) {

View File

@ -2,7 +2,7 @@ import { ExtractResult } from "./result";
const URL_REG = /^\s*(https?):\/\//im;
export function parseUrls(...args) {
export function parseUrls(...args): string[] {
if (!args.length) return [];
let arg = args.shift();
if (arg instanceof Array) {

View File

@ -13,7 +13,7 @@ export class Logger {
constructor(logLevel, notifyLevel) {
if (logLevel) this._log_level = logLevel;
if (notifyLevel) this._notify_level = notifyLevel;
chrome.notifications.onClosed.addListener((id, byUser) => { this._notify_level = undefined });
if (chrome.notifications) chrome.notifications.onClosed.addListener((id, byUser) => { this._notify_level = undefined });
}
get logLevel() {
return this._log_level;

View File

@ -1,3 +1,5 @@
import { logger } from "../common/logger";
export function extract(itemsSelector: string, fieldSelectors: string[], expectedURL: string): string[][] {
if (expectedURL && location.href != expectedURL) {
throw 'Target tab URL changed, aborting...';
@ -14,19 +16,42 @@ export function extract(itemsSelector: string, fieldSelectors: string[], expecte
item => {
return fieldSelectors.map(
selector => {
let doClick = false;
if (selector.startsWith("!")) {
doClick = true;
selector = selector.substring(1);
}
let [cls, attr] = selector.split('@').slice(0, 2);
let fieldVals = Array.from(item.querySelectorAll(cls));
if (!fieldVals.length) {
let fieldElements: Element[];
cls = cls.trim()
if (cls != "") {
fieldElements = Array.from(item.querySelectorAll(cls));
} else {
fieldElements = [item];
}
if (!fieldElements.length) {
return;
}
fieldFound[selector] = true;
return fieldVals.map(find => attr ? find[attr] : find.textContent.trim()).join('\n')
return fieldElements.map(find => {
if (doClick) {
let e = document.createEvent("MouseEvents");
e.initEvent("click", true, true);
find.dispatchEvent(e);
}
return attr ? find[attr] : find.textContent.trim();
}).join('\n')
}
)
}
);
// TODO: configurable wait logic
// if it exists a field, which is not found in any row, the sender should retry.
let shouldWait = fieldSelectors.reduce((p, c) => p || !fieldFound[c], false);
let notFoundFields = fieldSelectors.filter(f => !fieldFound[f]);
let shouldWait = notFoundFields.length > 0;
if (shouldWait) {
logger.debug('should wait for:', fieldSelectors.filter(f => !fieldFound[f]).join(','));
}
return shouldWait ? [] : results;
}

View File

@ -10,7 +10,7 @@ window.onload = function () {
document.querySelector('#link-document')
.addEventListener('click', () => {
chrome.tabs.create({
'url': `https://git.jebbs.co/jebbs/data-extracter-extesion`
'url': `https://git.qjebbs.com/jebbs/data-extracter-extesion`
});
})
document.querySelector('#state-input')

View File

@ -52,7 +52,7 @@
<p>
<b>Full document at:</b>
<br>
<a href="#" id="link-document">https://git.jebbs.co/jebbs/data-extracter-extesion</a>
<a href="#" id="link-document">https://git.qjebbs.com/jebbs/data-extracter-extesion</a>
</p>
</div>
</div>