chore: import upstream snapshot with attribution
Deploy local.promptfoo.app / Deploy to Cloudflare Pages (push) Waiting to run
Test and Publish Multi-arch Docker Image / test (push) Waiting to run
Test and Publish Multi-arch Docker Image / build-docker-and-push-digests (map[digest-suffix:linux-amd64 platform:linux/amd64 runner:ubuntu-latest]) (push) Blocked by required conditions
Test and Publish Multi-arch Docker Image / build-docker-and-push-digests (map[digest-suffix:linux-arm64 platform:linux/arm64 runner:ubuntu-24.04-arm]) (push) Blocked by required conditions
Test and Publish Multi-arch Docker Image / merge-docker-digests (push) Blocked by required conditions
Test and Publish Multi-arch Docker Image / Attest Multi-arch Image (push) Blocked by required conditions
Validate Renovate Config / Validate Renovate Configuration (push) Waiting to run
CI / Shell Format Check (push) Has been cancelled
CI / Check Ruby (3.4) (push) Has been cancelled
CI / CI Config (push) Has been cancelled
CI / Test on Node ${{ matrix.node }} and ${{ matrix.os }}${{ matrix.shard && format(' (shard {0}/3)', matrix.shard) || '' }} (push) Has been cancelled
CI / Build on Node ${{ matrix.node }} (push) Has been cancelled
CI / Style Check (push) Has been cancelled
CI / Generate Assets (push) Has been cancelled
CI / Check Python (3.14) (push) Has been cancelled
CI / Check Python (3.9) (push) Has been cancelled
CI / Build Docs (push) Has been cancelled
CI / Code Scan Action (push) Has been cancelled
CI / Site tests (push) Has been cancelled
CI / webui tests (push) Has been cancelled
CI / Run Integration Tests (push) Has been cancelled
CI / Run Smoke Tests (push) Has been cancelled
CI / Go Tests (push) Has been cancelled
CI / Share Test (push) Has been cancelled
CI / Redteam (Production API) (push) Has been cancelled
CI / Redteam (Staging API) (push) Has been cancelled
CI / GitHub Actions Lint (push) Has been cancelled
CI / Check Ruby (3.0) (push) Has been cancelled
release-please / release-please (push) Has been cancelled
release-please / build (push) Has been cancelled
release-please / publish-npm (push) Has been cancelled
release-please / publish-npm-backfill (push) Has been cancelled
release-please / docker (push) Has been cancelled
release-please / publish-code-scan-action (push) Has been cancelled
release-please / attest-code-scan-action (push) Has been cancelled

This commit is contained in:
wehub-resource-sync
2026-07-13 13:24:08 +08:00
commit 0d3cb498a3
5438 changed files with 1316560 additions and 0 deletions
+8
View File
@@ -0,0 +1,8 @@
# integration-browser (Browser)
Examples for using promptfoo with browser-based testing.
## Examples
- [existing-session](./existing-session/) - Connect to an existing browser session
- [headless](./headless/) - Headless browser automation
@@ -0,0 +1,79 @@
# integration-browser/existing-session (Browser Existing Session)
Test OAuth-authenticated applications by connecting to existing Chrome browser sessions.
You can run this example with:
```bash
npx promptfoo@latest init --example integration-browser/existing-session
cd integration-browser/existing-session
```
## Overview
This example demonstrates two browser provider modes:
1. **Connecting to existing Chrome sessions** - For OAuth/SSO authenticated testing
2. **Launching new browsers** - Normal browser automation
## Quick Start
### 1. Start the test server
```bash
cd examples/integration-browser/existing-session
node server.js
```
### 2. For OAuth testing (existing session)
```bash
# Start Chrome with debugging
chrome --remote-debugging-port=9222 --user-data-dir=/tmp/chrome-test
# Manually log into your application in that Chrome instance
# Run tests using the existing session
npx promptfoo eval --filter-providers existing-session
```
### 3. For normal browser testing
```bash
# Just run the tests (launches new browser)
npx promptfoo eval --filter-providers new-browser
```
## Configuration
The `promptfooconfig.yaml` includes both modes:
```yaml
providers:
# Connect to existing session
- id: browser
label: existing-session
config:
connectOptions:
debuggingPort: 9222
# ... test steps
# Launch new browser
- id: browser
label: new-browser
config:
headless: true
# ... test steps
```
## Prerequisites
```bash
npm install playwright @playwright/browser-chromium playwright-extra puppeteer-extra-plugin-stealth
```
## Files
- `promptfooconfig.yaml` - Example configuration showing both modes
- `server.js` - Test web server
- `test-page.html` - Mock authenticated chat interface
@@ -0,0 +1,90 @@
# yaml-language-server: $schema=https://promptfoo.dev/config-schema.json
description: Browser provider - OAuth authentication testing example
# This example demonstrates two modes:
# 1. Connecting to an existing Chrome browser session (for OAuth/SSO testing)
# 2. Launching a new browser (normal operation)
providers:
# Example 1: Connect to existing Chrome session with active authentication
- id: browser
label: existing-session
config:
connectOptions:
debuggingPort: 9222 # Chrome DevTools port
steps:
- action: navigate
args:
url: 'http://localhost:8080'
- action: type
args:
selector: '#chat-input'
text: '{{prompt}}'
- action: click
args:
selector: '#send-button'
- action: wait
args:
ms: 1000
- action: extract
args:
selector: '.chat-message:last-child .message-content'
name: response
transformResponse: 'extracted.response'
# Example 2: Launch new browser (normal operation)
- id: browser
label: new-browser
config:
headless: true # Set to false to see the browser
steps:
- action: navigate
args:
url: 'http://localhost:8080'
- action: type
args:
selector: '#chat-input'
text: '{{prompt}}'
- action: click
args:
selector: '#send-button'
- action: wait
args:
ms: 1000
- action: extract
args:
selector: '.chat-message:last-child .message-content'
name: response
transformResponse: 'extracted.response'
prompts:
- 'Who am I logged in as?'
tests:
- description: Test authenticated session access
vars:
prompt: 'Who am I logged in as?'
assert:
- type: contains
value: 'John Doe'
# Instructions:
#
# For OAuth/SSO testing (existing session):
# 1. Start Chrome with debugging: chrome --remote-debugging-port=9222
# 2. Log into your application manually
# 3. Run: npx promptfoo eval --filter-providers existing-session
#
# For normal browser testing:
# Run: npx promptfoo eval --filter-providers new-browser
@@ -0,0 +1,30 @@
const http = require('http');
const fs = require('fs');
const path = require('path');
const PORT = 8080;
const server = http.createServer((req, res) => {
console.log(`Request: ${req.method} ${req.url}`);
if (req.url === '/' || req.url === '/chat') {
const filePath = path.join(__dirname, 'test-page.html');
fs.readFile(filePath, 'utf8', (err, content) => {
if (err) {
res.writeHead(500);
res.end('Error loading page');
return;
}
res.writeHead(200, { 'Content-Type': 'text/html' });
res.end(content);
});
} else {
res.writeHead(404);
res.end('Not found');
}
});
server.listen(PORT, () => {
console.log(`Test server running at http://localhost:${PORT}`);
console.log('Open this URL in Chrome to test the authenticated session');
});
@@ -0,0 +1,148 @@
<!doctype html>
<html>
<head>
<title>Test Chat Interface</title>
<style>
body {
font-family: Arial, sans-serif;
max-width: 800px;
margin: 0 auto;
padding: 20px;
}
#user-info {
background: #f0f0f0;
padding: 10px;
margin-bottom: 20px;
border-radius: 5px;
}
#chat-container {
border: 1px solid #ccc;
padding: 20px;
min-height: 300px;
margin-bottom: 20px;
}
.chat-message {
margin: 10px 0;
padding: 10px;
background: #e9e9e9;
border-radius: 5px;
}
.message-content {
margin-top: 5px;
}
#chat-input {
width: 70%;
padding: 10px;
font-size: 16px;
}
#send-button {
padding: 10px 20px;
font-size: 16px;
background: #007bff;
color: white;
border: none;
border-radius: 5px;
cursor: pointer;
}
[data-testid='user-avatar'] {
display: inline-block;
width: 40px;
height: 40px;
background: #007bff;
color: white;
border-radius: 50%;
text-align: center;
line-height: 40px;
margin-right: 10px;
}
</style>
</head>
<body>
<div id="user-info">
<span data-testid="user-avatar">JD</span>
<strong>Logged in as:</strong> John Doe (john.doe@example.com)
</div>
<h1>Test Chat Interface</h1>
<div id="chat-container">
<div class="chat-message">
<strong>System:</strong>
<div class="message-content">Welcome! You are authenticated and can start chatting.</div>
</div>
</div>
<div>
<input type="text" id="chat-input" placeholder="Type your message..." />
<button id="send-button">Send</button>
</div>
<script>
// Simple chat simulation
document.getElementById('send-button').addEventListener('click', function () {
const input = document.getElementById('chat-input');
const message = input.value.trim();
if (message) {
// Add user message
const userMsg = document.createElement('div');
userMsg.className = 'chat-message';
const userLabel = document.createElement('strong');
userLabel.textContent = 'You:';
userMsg.appendChild(userLabel);
const userContent = document.createElement('div');
userContent.className = 'message-content';
userContent.textContent = message;
userMsg.appendChild(userContent);
document.getElementById('chat-container').appendChild(userMsg);
// Simulate bot response
setTimeout(() => {
const botMsg = document.createElement('div');
botMsg.className = 'chat-message';
let response = '';
if (message.toLowerCase().includes('who am i')) {
response =
'You are logged in as John Doe (john.doe@example.com). You have authenticated access to this chat system.';
} else if (message.toLowerCase().includes('status')) {
response =
'Your account status is: Active. Premium subscription valid until 2025-12-31.';
} else if (message.toLowerCase().includes('activity')) {
response = 'Your recent activity: Last login 2 hours ago. 5 conversations today.';
} else if (message.toLowerCase().includes('permission')) {
response = 'You have the following permissions: chat.read, chat.write, profile.edit';
} else if (message.toLowerCase().includes('other user')) {
response = "Error: You cannot access other users' private data. Permission denied.";
} else {
response = 'I received your message: "' + message + '". How can I help you today?';
}
const botLabel = document.createElement('strong');
botLabel.textContent = 'Assistant:';
botMsg.appendChild(botLabel);
const botContent = document.createElement('div');
botContent.className = 'message-content';
botContent.textContent = response;
botMsg.appendChild(botContent);
document.getElementById('chat-container').appendChild(botMsg);
}, 500);
input.value = '';
}
});
// Allow Enter key to send
document.getElementById('chat-input').addEventListener('keypress', function (e) {
if (e.key === 'Enter') {
document.getElementById('send-button').click();
}
});
</script>
</body>
</html>
@@ -0,0 +1,182 @@
# integration-browser/headless (Headless Browser Automation)
A browser automation example demonstrating how to test web applications using Playwright.
You can run this example with:
```bash
npx promptfoo@latest init --example integration-browser/headless
cd integration-browser/headless
```
## Overview
This example demonstrates how to:
- Test a local Gradio application using browser automation
- Handle dynamic JavaScript-rendered content
- Extract data from web interfaces
- Work with complex UI interactions (forms, tabs, buttons)
## Prerequisites
Ensure you have Python 3 and Node.js installed on your system.
1. **Install Node.js dependencies**:
```bash
npm install playwright @playwright/browser-chromium playwright-extra puppeteer-extra-plugin-stealth
```
2. **Install Python dependencies** (for the demo application):
```bash
pip install -r requirements.txt
```
That's it! No additional setup scripts or configuration needed.
## Running the Example
1. **Start the Gradio demo application**:
```bash
python gradio_demo.py
```
This starts a local server at http://localhost:7860
2. **Run the browser automation tests**:
```bash
npx promptfoo@latest eval -c promptfooconfig.yaml
```
3. **View the results**:
```bash
npx promptfoo@latest view
```
## Test Results
### Chatbot Example
The main configuration (`promptfooconfig.yaml`) tests a chatbot interface with a 100% pass rate:
```text
┌─────────────────────────────────────────────────┬─────────────────────────────────────────────────┐
│ topic │ [browser-provider] Tell me about {{topic}} │
├─────────────────────────────────────────────────┼─────────────────────────────────────────────────┤
│ testing browser automation │ [PASS] Test successful! The browser automation │
│ │ is working correctly. │
├─────────────────────────────────────────────────┼─────────────────────────────────────────────────┤
│ how the system works │ [PASS] I received your message: 'Tell me about │
│ │ how the system works'. This is a simple demo │
│ │ response! │
├─────────────────────────────────────────────────┼─────────────────────────────────────────────────┤
│ a simple greeting │ [PASS] I received your message: 'Tell me about │
│ │ a simple greeting'. This is a simple demo │
│ │ response! │
└─────────────────────────────────────────────────┴─────────────────────────────────────────────────┘
```
### Calculator Example
The `calculator-example.yaml` demonstrates form interactions with a 100% pass rate:
```text
┌───────────────────┬───────────────────┬───────────────────┬───────────────────┬───────────────────┐
│ num1 │ num2 │ operation │ operationSelector │ [browser-provider]│
├───────────────────┼───────────────────┼───────────────────┼───────────────────┼───────────────────┤
│ 10 │ 5 │ Add │ #operation │ [PASS] Calculator │
│ │ │ │ label:nth-child(1)│ interaction │
│ │ │ │ │ successful │
├───────────────────┼───────────────────┼───────────────────┼───────────────────┼───────────────────┤
│ 20 │ 4 │ Multiply │ #operation │ [PASS] Calculator │
│ │ │ │ label:nth-child(3)│ interaction │
│ │ │ │ │ successful │
└───────────────────┴───────────────────┴───────────────────┴───────────────────┴───────────────────┘
```
This example demonstrates:
- Navigate between tabs in a web application
- Fill multiple input fields
- Select radio button options
- Click buttons and wait for results
- Extract and verify content from the page
## Configuration Details
The example configurations demonstrate key concepts:
- **Appropriate delays**: 2-3 seconds between actions for reliability
- **Local testing**: Tests run against localhost:7860
- **Error handling**: Uses `transformResponse` for data extraction
- **Clear assertions**: Validates expected outputs
## Selectors Used
The Gradio application provides consistent selectors:
- `textarea[data-testid="textbox"]` - Message input field
- `button#submit-button` - Submit button
- `div[data-testid="bot"]:last-of-type .prose` - Latest bot response
- `button[value="calculator"]` - Calculator tab button
- `input[type="radio"]` - Operation selection
## Adapting This Example
### Testing Your Own Application
1. Update the `url` in the navigation step
2. Modify selectors to match your UI elements
3. Adjust wait times based on your application's response time
4. Add appropriate assertions for your use case
### Handling Dynamic Content
For single-page applications or AJAX content:
```yaml
- action: waitForNewChildren
args:
parentSelector: '#results-container'
timeout: 10000
```
### Complex Interactions
Chain multiple actions for sophisticated workflows:
```yaml
steps:
- action: navigate
args:
url: 'http://localhost:3000'
- action: click
args:
selector: '#menu-button'
- action: wait
args:
ms: 1000
- action: click
args:
selector: '#dropdown-option-2'
```
## Debugging Tips
| Issue | Solution |
| ----------------------- | ----------------------------------------------- |
| Elements not found | Use browser DevTools to verify selectors |
| Timing issues | Increase wait times or use `waitForNewChildren` |
| Want to see the browser | Set `headless: false` in the configuration |
| Need detailed logs | Run with `npx promptfoo@latest eval --verbose` |
## Additional Resources
- [Browser Provider Documentation](/docs/providers/browser)
- [Playwright Selectors Guide](https://playwright.dev/docs/selectors)
- [Gradio Documentation](https://www.gradio.app/docs)
@@ -0,0 +1,85 @@
# yaml-language-server: $schema=https://promptfoo.dev/config-schema.json
description: 'Calculator demo for browser automation'
prompts:
- 'Calculate {{num1}} {{operation}} {{num2}}'
providers:
- id: browser
config:
headless: false # See the browser in action
steps:
# Navigate to the demo
- action: navigate
args:
url: 'http://localhost:7860'
# Wait for initial page load
- action: wait
args:
ms: 3000
# Navigate to calculator tab
- action: click
args:
selector: 'button[role="tab"]:nth-child(2)'
# Wait for tab to load
- action: wait
args:
ms: 2000
# Enter numbers and perform calculation
- action: type
args:
selector: '#num1 input'
text: '{{num1}}'
- action: type
args:
selector: '#num2 input'
text: '{{num2}}'
# Select operation
- action: click
args:
selector: '{{operationSelector}}'
# Perform calculation
- action: click
args:
selector: '#calculate'
# Wait for calculation
- action: wait
args:
ms: 2000
# Extract entire result area for debugging
- action: extract
args:
selector: '#result'
name: resultArea
# For demo purposes, we'll just check that we got something
transformResponse: 'extracted.resultArea && extracted.resultArea.includes("Result") ? "Calculator interaction successful" : "No result found"'
# Simple demo tests
tests:
- vars:
num1: '10'
num2: '5'
operation: Add
operationSelector: '#operation label:nth-child(1)'
assert:
- type: contains
value: 'successful'
- vars:
num1: '20'
num2: '4'
operation: Multiply
operationSelector: '#operation label:nth-child(3)'
assert:
- type: contains
value: 'successful'
@@ -0,0 +1,141 @@
import time
import gradio as gr
# Simple chatbot function
def chatbot_response(message, history):
"""Return deterministic keyword replies for browser automation tests."""
time.sleep(1) # Simulate processing time
responses = {
"hello": "Hello! How can I help you today?",
"how are you": "I'm doing great, thank you for asking! How are you?",
"test": "Test successful! The browser automation is working correctly.",
"demo": "This is a demo Gradio application for testing browser automation.",
"help": "I can respond to simple greetings and questions. Try saying 'hello' or asking 'how are you'!",
}
# Simple keyword matching
message_lower = message.lower()
for keyword, response in responses.items():
if keyword in message_lower:
return response
# Default response
return f"I received your message: '{message}'. This is a simple demo response!"
# Create the Gradio interface
def create_demo():
"""Create the chatbot demo used by browser automation examples."""
with gr.Blocks(title="Browser Testing Demo") as demo:
gr.Markdown(
"""
# 🤖 Browser Automation Test Demo
This is a simple Gradio application designed for testing browser automation with promptfoo.
**Features:**
- Simple chat interface
- Predictable responses for testing
- Clear element identifiers for automation
"""
)
chatbot = gr.Chatbot(
label="Chat History", elem_id="chat-history", type="messages"
)
msg = gr.Textbox(
label="Your Message",
placeholder="Type a message and press Enter...",
elem_id="user-input",
)
with gr.Row():
submit = gr.Button("Submit", elem_id="submit-button")
clear = gr.Button("Clear", elem_id="clear-button")
# Example messages for testing
gr.Examples(
examples=[
"Hello",
"How are you?",
"This is a test",
"Show me a demo",
"Help",
],
inputs=msg,
label="Example Messages",
)
# Handle message submission
def respond(message, chat_history):
"""Append the user message and deterministic bot response."""
bot_message = chatbot_response(message, chat_history)
# Use the new messages format with role and content
chat_history.append({"role": "user", "content": message})
chat_history.append({"role": "assistant", "content": bot_message})
return "", chat_history
# Wire up the interface
msg.submit(respond, [msg, chatbot], [msg, chatbot])
submit.click(respond, [msg, chatbot], [msg, chatbot])
clear.click(lambda: None, None, chatbot, queue=False)
return demo
# Simple calculator for testing form inputs
def create_calculator_demo():
"""Create the calculator demo used for form input testing."""
with gr.Blocks(title="Calculator Demo") as calc_demo:
gr.Markdown("## Simple Calculator for Testing")
with gr.Row():
num1 = gr.Number(label="First Number", elem_id="num1")
num2 = gr.Number(label="Second Number", elem_id="num2")
operation = gr.Radio(
["Add", "Subtract", "Multiply", "Divide"],
label="Operation",
elem_id="operation",
)
calculate_btn = gr.Button("Calculate", elem_id="calculate")
result = gr.Textbox(label="Result", elem_id="result")
def calculate(n1, n2, op):
"""Calculate the selected arithmetic operation."""
if op == "Add":
return str(n1 + n2)
elif op == "Subtract":
return str(n1 - n2)
elif op == "Multiply":
return str(n1 * n2)
elif op == "Divide":
return str(n1 / n2) if n2 != 0 else "Error: Division by zero"
return "Error: Unknown operation"
calculate_btn.click(calculate, [num1, num2, operation], result)
return calc_demo
if __name__ == "__main__":
# Create a tabbed interface with multiple demos
demo1 = create_demo()
demo2 = create_calculator_demo()
tabbed_demo = gr.TabbedInterface(
[demo1, demo2],
["Chatbot Demo", "Calculator Demo"],
title="Browser Automation Test Suite",
)
print("Starting Gradio demo server...")
print("Access the demo at http://localhost:7860")
print("Use Ctrl+C to stop the server")
tabbed_demo.launch(server_name="0.0.0.0", server_port=7860, share=False)
@@ -0,0 +1,68 @@
# yaml-language-server: $schema=https://promptfoo.dev/config-schema.json
description: 'Testing browser automation with local Gradio app'
prompts:
- 'Tell me about {{topic}}'
providers:
- id: browser
config:
headless: false # set to false to see the browser in action
steps:
# Always start with a respectful delay
- action: wait
args:
ms: 2000
# Navigate to local Gradio app
- action: navigate
args:
url: 'http://localhost:7860'
# Wait for the page to load
- action: wait
args:
ms: 2000
# Type the message in the Gradio textbox
- action: type
args:
selector: 'textarea[data-testid="textbox"]'
text: '{{prompt}}'
# Submit the message
- action: click
args:
selector: 'button#submit-button'
# Wait for the response with reasonable timeout
- action: wait
args:
ms: 3000
# Extract the bot's response from the chatbot - get the last message
- action: extract
args:
selector: 'div[data-testid="bot"]:last-of-type .prose'
name: response
transformResponse: 'extracted.response'
tests:
- vars:
topic: testing browser automation
assert:
- type: contains
value: 'Test successful'
- vars:
topic: how the system works
assert:
- type: contains
value: 'simple demo response'
- vars:
topic: a simple greeting
assert:
- type: javascript
value: output.length > 0
@@ -0,0 +1,2 @@
gradio>=5.49.1
playwright>=1.55.0