Spaces:
Running
Running
File size: 11,480 Bytes
4c3fe29 10dbcf2 35ae779 3819331 1bb1ac8 31d5b37 35ae779 15fdb32 35ae779 31d5b37 d74e8cc 80fef4e d164b37 80fef4e 006e72f 1bb1ac8 d164b37 7f2bf6a d164b37 31d5b37 7f2bf6a d164b37 3819331 006e72f 7f2bf6a d164b37 7f2bf6a 10dbcf2 7f2bf6a f2b00a1 10dbcf2 f2b00a1 7f2bf6a d164b37 31d5b37 10dbcf2 1bb1ac8 10dbcf2 1bb1ac8 10dbcf2 1bb1ac8 10dbcf2 1bb1ac8 80fef4e 006e72f 7f2bf6a 80fef4e 7f2bf6a 10dbcf2 1bb1ac8 80fef4e 7f2bf6a d164b37 7f2bf6a 3819331 1bb1ac8 10dbcf2 1bb1ac8 10dbcf2 1bb1ac8 80fef4e 7f2bf6a d164b37 80fef4e 10dbcf2 35ae779 15fdb32 7f2bf6a 80fef4e 7f2bf6a 35ae779 80fef4e 7f2bf6a 80fef4e 35ae779 10dbcf2 7f2bf6a 10dbcf2 f2b00a1 31d5b37 7f2bf6a 10dbcf2 3819331 1bb1ac8 10dbcf2 1bb1ac8 10dbcf2 3819331 31d5b37 006e72f 10dbcf2 1bb1ac8 10dbcf2 31d5b37 d74e8cc 1bb1ac8 80fef4e 1bb1ac8 10dbcf2 d74e8cc 10dbcf2 d74e8cc 10dbcf2 d74e8cc 1bb1ac8 10dbcf2 3819331 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 |
import os
os.system("playwright install")
# app.py (Final, Working Async Version with All Bugs Fixed)
import gradio as gr
from playwright.async_api import async_playwright
from bs4 import BeautifulSoup, NavigableString
import urllib.parse
import os
from itertools import cycle
import uuid
# --- 1. GLOBAL RESOURCES & STATE ---
P = None
BROWSER = None
REVOLVER = None
LIVE_CONTEXTS = {}
APP_STARTED = False
SEARCH_ENGINES = {
"DuckDuckGo": "https://duckduckgo.com/html/?q={query}",
"Google": "https://www.google.com/search?q={query}",
"Bing": "https://www.bing.com/search?q={query}",
"Brave": "https://search.brave.com/search?q={query}",
"Ecosia": "https://www.ecosia.org/search?q={query}"
}
# --- 2. PLAIN DATA STATE CLASSES ---
class TabState:
def __init__(self, tab_id, proxy_used="Direct Connection"):
self.id, self.url, self.title = tab_id, "about:blank", "New Tab"
self.parsed_text, self.links, self.proxy_used = "Welcome!", [], proxy_used
class BrowserState:
def __init__(self): self.tabs, self.active_tab_id = [], None
def get_active_tab(self): return next((t for t in self.tabs if t.id == self.active_tab_id), None)
class CredentialRevolver:
def __init__(self, proxy_string: str):
self.proxies = self._parse_proxies(proxy_string)
self.proxy_cycler = cycle(self.proxies) if self.proxies else None
def _parse_proxies(self, proxy_string: str):
proxies = [];
for line in proxy_string.strip().splitlines():
try:
parsed = urllib.parse.urlparse(f"//{line.strip()}"); server = f"{parsed.scheme or 'http'}://{parsed.hostname}:{parsed.port}"
proxies.append({"server": server, "username": parsed.username, "password": parsed.password})
except: pass
return proxies
def get_next(self): return next(self.proxy_cycler) if self.proxy_cycler else None
def count(self): return len(self.proxies) if self.proxies else 0
# --- 3. CORE ASYNC LOGIC & FORMATTING ---
def _process_element_to_markdown(element):
if isinstance(element, NavigableString): return element.strip()
if element.name is None: return ''
inner_text = ''.join(_process_element_to_markdown(child) for child in element.children)
if element.name in ['p', 'div', 'article', 'section']: return f"\n{inner_text.strip()}\n"
if element.name == 'h1': return f"\n# {inner_text.strip()}\n"
if element.name == 'h2': return f"\n## {inner_text.strip()}\n"
if element.name == 'h3': return f"\n### {inner_text.strip()}\n"
if element.name in ['h4', 'h5', 'h6']: return f"\n#### {inner_text.strip()}\n"
if element.name == 'li': return f"* {inner_text.strip()}\n"
if element.name in ['ul', 'ol']: return f"\n{inner_text}\n"
if element.name in ['strong', 'b']: return f"**{inner_text.strip()}**"
if element.name in ['em', 'i']: return f"*{inner_text.strip()}*"
if element.name in ['pre', 'code']: return f"\n```\n{inner_text.strip()}\n```\n"
if element.name == 'a': return f"[{inner_text.strip()}]({element.get('href', '')})"
if element.name == 'hr': return "\n---\n"
return inner_text
def _format_html_to_markdown(soup):
content_node = soup.find('main') or soup.find('body')
if not content_node: return "Could not find main body content."
for el in content_node.select('nav, footer, header, aside, form, script, style'): el.decompose()
return _process_element_to_markdown(content_node)
async def _fetch_and_update_tab_state(tab_state: TabState, url: str):
log = f"▶️ Navigating to {url}..."; live_page = LIVE_CONTEXTS[tab_state.id]["page"]
try:
await live_page.goto(url, wait_until='domcontentloaded', timeout=30000)
tab_state.url = live_page.url; tab_state.title = await live_page.title() or "No Title"
log += f"\n✅ Arrived at: {tab_state.url}"
html_content = await live_page.content(); soup = BeautifulSoup(html_content, 'lxml')
tab_state.parsed_text = _format_html_to_markdown(soup)
tab_state.links = [{'text': link.get_text(strip=True) or "[No Link Text]", 'url': urllib.parse.urljoin(tab_state.url, link['href'])} for link in soup.find_all('a', href=True) if link.get('href', '').startswith('http')]
log += f"\n🔗 Found {len(tab_state.links)} links."
except Exception as e:
error_message = str(e).splitlines()[0]; tab_state.title = "Error"; tab_state.url = url
tab_state.parsed_text = f"❌ Failed to load page.\n\nError: {error_message}"
tab_state.links = []; log += f"\n❌ {error_message}"
return log
async def handle_action(browser_state: BrowserState, search_engine: str, action: str, value=None):
active_tab_state = browser_state.get_active_tab()
if action == "go" and active_tab_state and value:
is_url = urllib.parse.urlparse(value).scheme and urllib.parse.urlparse(value).netloc
url = value if is_url else SEARCH_ENGINES.get(search_engine, SEARCH_ENGINES["DuckDuckGo"]).format(query=urllib.parse.quote_plus(value))
log = await _fetch_and_update_tab_state(active_tab_state, url)
elif action == "new_tab":
tab_id, proxy_config = str(uuid.uuid4()), REVOLVER.get_next()
context = await BROWSER.new_context(proxy=proxy_config)
page = await context.new_page()
LIVE_CONTEXTS[tab_id] = {"context": context, "page": page}
new_tab = TabState(tab_id, proxy_used=proxy_config['server'].split('@')[-1] if proxy_config else "Direct")
browser_state.tabs.append(new_tab); browser_state.active_tab_id = tab_id
log = await _fetch_and_update_tab_state(new_tab, "https://www.startpage.com/")
elif action == "click" and active_tab_state and value is not None:
try:
link_index = int(value)
if 0 <= link_index < len(active_tab_state.links): log = await _fetch_and_update_tab_state(active_tab_state, active_tab_state.links[link_index]['url'])
else: log = "Invalid link number."
except: log = "Please enter a valid number to click."
elif action == "close_tab" and active_tab_state:
if len(browser_state.tabs) > 1:
tab_to_close_id = browser_state.active_tab_id; tab_index = browser_state.tabs.index(active_tab_state)
browser_state.tabs.pop(tab_index); new_index = tab_index - 1 if tab_index > 0 else 0
browser_state.active_tab_id = browser_state.tabs[new_index].id
resources = LIVE_CONTEXTS.pop(tab_to_close_id); await resources['context'].close(); log = f"💣 Tab closed."
else: log = "Cannot close the last tab."
elif action == "switch_tab" and value is not None:
browser_state.active_tab_id = value; log = f"Switched to tab."
else: log = "No action taken."
return browser_state, log
# ** CRITICAL BUG FIX: `NameError` is fixed by defining this function before it is called **
def update_ui_components(browser_state: BrowserState):
active_tab = browser_state.get_active_tab()
if not active_tab: return {page_content: gr.Markdown("No active tabs."), url_textbox: "", links_display: "", tab_selector: gr.Radio(choices=[])}
tab_choices = [(f"Tab {i}: {t.title[:25]}... ({t.proxy_used})", t.id) for i, t in enumerate(browser_state.tabs)]
links_md = "### 🔗 Links on Page\n" + ('\n'.join(f"{i}. [{link['text'][:80]}]({link['url']})" for i, link in enumerate(active_tab.links[:25])) if active_tab.links else "_No links found._")
page_md = f"# {active_tab.title}\n**URL:** {active_tab.url}\n\n---\n\n{active_tab.parsed_text}"
return {
page_content: gr.Markdown(page_md),
url_textbox: gr.Textbox(value=active_tab.url), links_display: gr.Markdown(links_md),
tab_selector: gr.Radio(choices=tab_choices, value=active_tab.id, label="Active Tabs"),
}
# --- 4. GRADIO UI AND EVENT HANDLING ---
with gr.Blocks(theme=gr.themes.Soft(), title="Real Browser Demo") as demo:
browser_state = gr.State(BrowserState())
gr.Markdown("# 🛰️ Real Browser Demo v2.1")
with gr.Row():
with gr.Column(scale=4):
with gr.Row(): url_textbox = gr.Textbox(label="Enter URL or Search Query", interactive=True, scale=4); go_btn = gr.Button("Go", variant="primary", scale=1)
with gr.Accordion("Page Content (Formatted)", open=True): page_content = gr.Markdown("Loading...")
with gr.Column(scale=2):
search_engine_selector = gr.Radio(list(SEARCH_ENGINES.keys()), value="DuckDuckGo", label="Search Engine")
with gr.Row(): new_tab_btn = gr.Button("➕ New Tab"); close_tab_btn = gr.Button("❌ Close Tab")
tab_selector = gr.Radio(choices=[], label="Active Tabs", interactive=True)
with gr.Accordion("Clickable Links", open=True):
links_display = gr.Markdown("...");
with gr.Row(): click_num_box = gr.Number(label="Link #", scale=1, minimum=0, step=1); click_btn = gr.Button("Click Link", scale=2)
log_display = gr.Textbox(label="Status Log", interactive=False)
all_outputs = [page_content, url_textbox, links_display, tab_selector, log_display]
async def master_handler(current_state, search_engine, action, value=None):
global APP_STARTED, P, BROWSER, REVOLVER
if not APP_STARTED:
print("🚀 First request received, starting up Playwright...");
P = await async_playwright().start(); BROWSER = await P.firefox.launch(headless=True)
proxy_list_str = os.getenv("PROXY_LIST", ""); REVOLVER = CredentialRevolver(proxy_list_str)
print(f"✅ Playwright started. {REVOLVER.count()} proxies loaded."); APP_STARTED = True
new_state, log = await handle_action(current_state, search_engine, action, value)
ui_updates = update_ui_components(new_state)
return (new_state, ui_updates[page_content], ui_updates[url_textbox], ui_updates[links_display], ui_updates[tab_selector], log)
async def on_load(state, search_engine): return await master_handler(state, search_engine, "new_tab", None)
async def on_go_click(state, search_engine, value): return await master_handler(state, search_engine, "go", value)
async def on_click_link(state, search_engine, value): return await master_handler(state, search_engine, "click", value)
async def on_new_tab(state, search_engine): return await master_handler(state, search_engine, "new_tab", None)
async def on_close_tab(state, search_engine): return await master_handler(state, search_engine, "close_tab", None)
async def on_switch_tab(state, search_engine, value): return await master_handler(state, search_engine, "switch_tab", value)
outputs = [browser_state, *all_outputs]
demo.load(on_load, [browser_state, search_engine_selector], outputs)
go_btn.click(on_go_click, [browser_state, search_engine_selector, url_textbox], outputs, show_progress="full")
url_textbox.submit(on_go_click, [browser_state, search_engine_selector, url_textbox], outputs, show_progress="full")
click_btn.click(on_click_link, [browser_state, search_engine_selector, click_num_box], outputs, show_progress="full")
new_tab_btn.click(on_new_tab, [browser_state, search_engine_selector], outputs, show_progress="full")
close_tab_btn.click(on_close_tab, [browser_state, search_engine_selector], outputs)
tab_selector.input(on_switch_tab, [browser_state, search_engine_selector, tab_selector], outputs)
demo.launch() |