VisEval is a benchmark designed to evaluate visualization generation methods. In this repository, we provide both the toolkit to support the benchmarking, as well as the data used for benchmarks.

VisEval evaluates generated visualizations from three dimensions:

1. Whether the generated code can produce the visualization.
2. Whether the generated visualization meets the query.
3. Whether the generated visualization is easy to read.

pip install --upgrade vis-evaluator
# or `git clone https://github.com/microsoft/VisEval.git && cd VisEval && pip install --upgrade -e .`

To access the dataset, please follow these steps:

1. Download the dataset from this link.
2. Once the download is complete, unzip the file to extract the dataset contents.

For additional information about the dataset, please refer to the dataset documentation.

After installation, you can use VisEval by referring to examples/evaluate.py or a follow:

1. Create your generation method by inheriting from the Agent Class. You can find three examples in the examples/agent directory.

from viseval.agent import Agent, ChartExecutionResult

class YourAgent(Agent):
    def __init__(self, llm):
        self.llm = llm
    
    def generate(
        self, nl_query: str, tables: list[str], config: dict
    ) -> Tuple[str, dict]:
        """Generate code for the given natural language query."""
        pass

    def execute(
        self, code: str, context: dict, log_name: str = None
    ) -> ChartExecutionResult:
        """Execute the given code with context and return the result"""
        pass

2. Configure evaluator.

    evaluator = Evaluator(webdriver_path, vision_model)

(You can configure the Evaluator without a webdriver and vision model, in which case the evaluation of the readability of the generated visualizations will be skipped.)

• Install webdriver.

  # download
  wget https://dl.google.com/linux/direct/google-chrome-stable_current_amd64.deb
  # install
  apt install google-chrome-stable_current_amd64.deb
  # verify
  google-chrome --version

• Load vision model (e.g., GPT4-v).

  from langchain_openai import AzureChatOpenAI
  
  import dotenv
  # Copy .env.example to .env and put your API keys in the file.
  dotenv.load_dotenv()
  
  vision_model = AzureChatOpenAI(
      model_name="gpt-4-turbo-v",
      max_retries=999,
      temperature=0.0,
      request_timeout=20,
      max_tokens=4096,
  )

3. Evaluate

from viseval import Dataset

# Configure dataset with the benchmark dataset folder path ( folder), 
# specify the number of tables required to generate visualizations (table_type`: all, single, or multiple),
# and indicate whether to include irrelevant tables (`with_irrelevant_tables`).
dataset = Dataset(folder, table_type, with_irrelevant_tables)

config = {"library": args.library}
result = evaluator.evaluate(agent, dataset, config)
score = result.score()
print(f"Score: {score}")

This project welcomes contributions and suggestions. Most contributions require you to agree to a Contributor License Agreement (CLA) declaring that you have the right to, and actually do, grant us the rights to use your contribution. For details, visit https://cla.opensource.microsoft.com.

When you submit a pull request, a CLA bot will automatically determine whether you need to provide a CLA and decorate the PR appropriately (e.g., status check, comment). Simply follow the instructions provided by the bot. You will only need to do this once across all repos using our CLA.

This project has adopted the Microsoft Open Source Code of Conduct.For more information see the Code of Conduct FAQ or contact [email protected] with any additional questions or comments.

This project may contain trademarks or logos for projects, products, or services. Authorized use of Microsoft trademarks or logos is subject to and must follow Microsoft's Trademark & Brand Guidelines. Use of Microsoft trademarks or logos in modified versions of this project must not cause confusion or imply Microsoft sponsorship. Any use of third-party trademarks or logos are subject to those third-party's policies.

This project has adopted the Microsoft Privacy Statement.