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
This commit is contained in:
2021-04-20 14:20:05 +08:00
parent 108ebb835f
commit e87e7010ec
7 changed files with 146 additions and 28 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.
@ -202,6 +214,26 @@ To stop watching, you can either `close current window`, or:
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

@ -8,14 +8,14 @@ import { logger } from "../common/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) {
@ -23,10 +23,12 @@ export function redirectTab(tab: chrome.tabs.Tab, url: string) {
}
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`)
) {
throw "Tasks stopped by user.";
return newURL;
}
return undefined;
}
@ -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

@ -1,9 +1,9 @@
import { Task } from "./task";
import { saveFile } from "./tools";
import { createTab, getActiveTab, ping } from "./actions";
import { ExtractResult } from "./result";
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[] = [];
@ -50,6 +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.
*/
@ -121,7 +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);
});
@ -143,22 +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 +1,11 @@
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) {
@ -35,4 +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

@ -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;

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

@ -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;
}