feat(wip): ui
This commit is contained in:
@@ -1,109 +1,91 @@
|
|||||||
import React, { useState, useRef, useEffect } from 'react';
|
import React, { useState, useEffect, useRef } from 'react';
|
||||||
|
import ReactDOM from 'react-dom';
|
||||||
import axios from 'axios';
|
import axios from 'axios';
|
||||||
|
|
||||||
function Scraper() {
|
const Scraper = () => {
|
||||||
const [url, setUrl] = useState('');
|
const [url, setUrl] = useState('');
|
||||||
const [selections, setSelections] = useState([]);
|
const [selectedSelectors, setSelectedSelectors] = useState([]);
|
||||||
const [data, setData] = useState(null);
|
const [scrapedData, setScrapedData] = useState(null);
|
||||||
const iframeRef = useRef(null);
|
const containerRef = useRef(null);
|
||||||
|
|
||||||
|
useEffect(() => {
|
||||||
|
const handleSelectElement = async (selector) => {
|
||||||
|
try {
|
||||||
|
await window.playwright.selectElement(selector);
|
||||||
|
setSelectedSelectors((prevSelectors) => [...prevSelectors, selector]);
|
||||||
|
} catch (error) {
|
||||||
|
console.error('Error selecting element:', error);
|
||||||
|
}
|
||||||
|
};
|
||||||
|
|
||||||
|
window.playwright = {
|
||||||
|
selectElement: handleSelectElement,
|
||||||
|
};
|
||||||
|
|
||||||
|
return () => {
|
||||||
|
window.playwright = null;
|
||||||
|
};
|
||||||
|
}, []);
|
||||||
|
|
||||||
const handleUrlChange = (e) => {
|
const handleUrlChange = (e) => {
|
||||||
setUrl(e.target.value);
|
setUrl(e.target.value);
|
||||||
};
|
};
|
||||||
|
|
||||||
const handleSelectElement = (e) => {
|
const handleOpenSite = async () => {
|
||||||
const selection = {
|
try {
|
||||||
name: `selection_${selections.length + 1}`,
|
const containerId = containerRef.current.id;
|
||||||
selector: generateSelector(e.target),
|
const response = await axios.post('http://localhost:3000/scrape', { url, containerId });
|
||||||
};
|
|
||||||
setSelections((prevSelections) => [...prevSelections, selection]);
|
|
||||||
};
|
|
||||||
|
|
||||||
const generateSelector = (element) => {
|
// Mount the Playwright context into the frontend container
|
||||||
let selector = element.tagName.toLowerCase();
|
if (response.data.success) {
|
||||||
let parent = element.parentNode;
|
const { context } = response.data;
|
||||||
|
await context.setDefaultViewportSize({ width: 1280, height: 720 });
|
||||||
while (parent && parent.tagName !== 'HTML') {
|
await context.exposeBinding('selectElement', async (selector) => {
|
||||||
const id = parent.id;
|
const selection = {
|
||||||
const classes = Array.from(parent.classList).join('.');
|
name: `selection_${selectedSelectors.length + 1}`,
|
||||||
|
selector,
|
||||||
if (id) {
|
};
|
||||||
selector = `#${id} > ${selector}`;
|
setSelectedSelectors((prevSelectors) => [...prevSelectors, selection]);
|
||||||
break;
|
});
|
||||||
} else if (classes) {
|
ReactDOM.render(<div id="playwright-container" />, containerRef.current);
|
||||||
selector = `.${classes} > ${selector}`;
|
await context.mount(document.getElementById('playwright-container'));
|
||||||
break;
|
|
||||||
} else {
|
} else {
|
||||||
selector = `${parent.tagName.toLowerCase()} > ${selector}`;
|
console.error('Error opening site:', response.data.error);
|
||||||
}
|
}
|
||||||
|
} catch (error) {
|
||||||
parent = parent.parentNode;
|
console.error('Error opening site:', error);
|
||||||
}
|
}
|
||||||
|
|
||||||
return selector;
|
|
||||||
};
|
};
|
||||||
|
|
||||||
const handleScrape = async () => {
|
const handleScrape = async () => {
|
||||||
try {
|
try {
|
||||||
const response = await axios.post('http://localhost:3000/scrape', {
|
const response = await axios.post('http://localhost:3000/scrape', { selectedSelectors });
|
||||||
url,
|
setScrapedData(response.data);
|
||||||
selections,
|
|
||||||
});
|
|
||||||
setData(response.data);
|
|
||||||
console.log('Scraped Data', response.data);
|
|
||||||
} catch (error) {
|
} catch (error) {
|
||||||
console.error('Error scraping:', error);
|
console.error('Error scraping:', error);
|
||||||
}
|
}
|
||||||
};
|
};
|
||||||
|
|
||||||
// useEffect(() => {
|
|
||||||
// if (iframeRef.current && iframeRef.current.contentWindow) {
|
|
||||||
// const contentWindow = iframeRef.current.contentWindow;
|
|
||||||
// const contentDocument = contentWindow.document;
|
|
||||||
|
|
||||||
// contentDocument.body.onclick = handleSelectElement;
|
|
||||||
// }
|
|
||||||
// }, [url]);
|
|
||||||
|
|
||||||
const handleIframeLoad = () => {
|
|
||||||
if (iframeRef.current && iframeRef.current.contentDocument) {
|
|
||||||
const contentDocument = iframeRef.current.contentDocument;
|
|
||||||
contentDocument.body.onclick = handleSelectElement;
|
|
||||||
}
|
|
||||||
};
|
|
||||||
|
|
||||||
return (
|
return (
|
||||||
<div>
|
<div>
|
||||||
<input
|
<input type="text" value={url} onChange={handleUrlChange} placeholder="Enter URL" />
|
||||||
type="text"
|
<button onClick={handleOpenSite}>Open Site</button>
|
||||||
placeholder="Enter URL to scrape"
|
<div ref={containerRef} style={{ width: '100%', height: '500px', border: '1px solid black' }} />
|
||||||
value={url}
|
<button onClick={handleScrape}>Scrape Data</button>
|
||||||
onChange={handleUrlChange}
|
|
||||||
/>
|
|
||||||
<button onClick={handleScrape}>Scrape</button>
|
|
||||||
<p>Click on the elements you want to scrape:</p>
|
|
||||||
<iframe
|
|
||||||
ref={iframeRef}
|
|
||||||
src={url}
|
|
||||||
style={{ width: '100%', height: '500px' }}
|
|
||||||
onLoad={handleIframeLoad}
|
|
||||||
/>
|
|
||||||
<p>Selected Elements:</p>
|
<p>Selected Elements:</p>
|
||||||
<ul>
|
<ul>
|
||||||
{selections.map((selection, index) => (
|
{selectedSelectors.map((selector, index) => (
|
||||||
<li key={index}>
|
<li key={index}>{selector}</li>
|
||||||
{selection.name}: {selection.selector}
|
|
||||||
</li>
|
|
||||||
))}
|
))}
|
||||||
</ul>
|
</ul>
|
||||||
{data && (
|
{scrapedData && (
|
||||||
<div>
|
<div>
|
||||||
<h2>Scraped Data:</h2>
|
<h2>Scraped Data:</h2>
|
||||||
<pre>{JSON.stringify(data, null, 2)}</pre>
|
<pre>{JSON.stringify(scrapedData, null, 2)}</pre>
|
||||||
</div>
|
</div>
|
||||||
)}
|
)}
|
||||||
</div>
|
</div>
|
||||||
);
|
);
|
||||||
}
|
};
|
||||||
|
|
||||||
export default Scraper;
|
export default Scraper;
|
||||||
Reference in New Issue
Block a user