UI Test-Case Generation Tool

Outline

This tool helps create well-defined and well-structured functional test cases based on the provided HTML content. This tool uses a crew of AI agents to interpret the received HTML data, perceive UI components and features, understand, use, and then curate a list of well-defined test cases.

This tool is built using Flask and communicates with the CrewAI workload to manage cooperation across an agentic crew. Also, it uses a variety of additional libraries, such as BeautifulSoup, to parse and clean provided HTML content.

Features

• Compatibility : This tool is compatible with any of the Langchain Supported Models

• Identification of Functionality : As per the structure of the provided HTML, the tool understands the expected use case and also keeps the context of the web app intact, and even considers the possible user interactions and prospects.

• TestCase Generation : Based upon all the different components of the HTML corpus comes a set of well-defined functional test cases. The AI agents understand crucial points like navigational flows, cases, user auths and access, data manipulation operations, and dynamic content handling.

• Handling for Roles and Permissions : The tool takes into account the user's roles and various levels of approval and access.

• Edge Cases : The boundary conditions and potential edge cases are perceived.

• Response: A JSON is defined as illustrated.

    {
        "message" : "success",
        "data": [
            {
                "test_case_description": "Test case description",
                "dependencies": "List of dependencies",
                "test_steps": "Steps to execute the test case",
                "expected_result": "Expected result of the test case"
            }
            
        ]
    } 

Getting Started

To use the tool follow these steps:

1. Clone The Repository

2. Use pip install -r requirements.txt to install the required dependencies.

3. Create a Virtual Environment

• Navigate to your project directory where you want to create the virtual environment and Open a terminal (or command prompt).

• Run the following command to create the virtual environment:

  python -m venv myenv

  Replace myenv with your preferred name for the virtual environment.

Activate the Virtual Environment

• On Windows Run the following command:

myenv\Scripts\activate

• On macOS and Linux Run the following command:

source myenv/bin/activate

4. This tool leverages the use of LangChain. Using LangChain this tool is compatible with any language model that Langchain supports, for more information regarding LangChain please refer to the following documentation

(https://python.langchain.com/v0.2/docs/introduction/)

• After the dependencies are installed, set the required environment variables as per the langchain documentation.

• We've used OpenAI gpt-4-0125-preview as the default LLM for the tool. If you would like to use the same, then set the OpenAI API key as an environment variable: OPENAI_API_KEY : "Your OpenAI API key"

5. To use any other LangChain model, instantiate a variable as per the Langchain documentation for that model and pass this variable as a parameter to the TestCrew Class Object in the app.py file as shown below

#....Continued code 
from langchain_openai import ChatOpenAI

llm = ChatOpenAI(
    model="gpt-4-0125-preview",
    temperature=0,
    max_tokens=None,
)
@app.route('/htmlagent',methods=['POST'])

    #.....The rest of the code remains the same
    

    for snippet in snippets:
        crew = TestCrew(str(snippet),llm=llm).kickoff_crew() #Pass your llm here
        
        
    # The rest of the code remains the same.......

Here is how you can use ChatAnthropic LLM for the application(https://python.langchain.com/v0.2/docs/integrations/platforms/anthropic/):

• Run pip install -U langchain-anthropic To install dependencies (Please refer to Langchain Docs).

• Set up your necessary environment variables :

  • ANTHROPIC_API_KEY: Your Anthropic API key

• In your app.py file

from langchain_anthropic import ChatAnthropic

llm = ChatAnthropic(model='claude-3-opus-20240229')

#Code Continued...

• Pass llm as a parameter in the TestCrew Class Object :

#....Continued code 
@app.route('/htmlagent',methods=['POST'])

    #.....The rest of the code remains the same
    

    for snippet in snippets:
        crew = TestCrew(str(snippet),llm=llm).kickoff_crew() #Pass your LLM here
        
        
    # The rest of the code remains the same.......

6. Run the Flask Application Server

7. Hit /htmlagent endpoint using a POST method.

8. The HTML data must be sent in the following JSON format with the HTML stringified.

{
   "html": "Your Stringified HTML content"
}

Contributing To the Project

Being an open-source project, all your contributions are highly anticipated! If you take an interest in this and encounter any issues or have further suggestions for improvements, please open an issue or submit a pull request. Do make sure to follow the Project's coding conventions.

To contribute to this project,

1. Fork the repositories
2. Make a distinct branch for the feature or issue fixes.
3. Make your modifications and commit them.
4. Publish your modifications to your forked repository.
5. Submit a new pull request to the original repository.