Skip to main content

🌐 HtmlAnalyzerNode Module

The HtmlAnalyzerNode module analyzes HTML code based on the desired information to be extracted and returns a reduced version of the HTML content. The output of this node is primarily intended to enhance the reasoning capabilities of the GenerateCodeNode, enabling it to generate more accurate and context-aware code.

Classes​

HtmlAnalyzerNode​

HtmlAnalyzerNode generates an analysis of the HTML code and returns a reduced version of the HTML, primarily to improve the GenerateCodeNode's reasoning.

Attributes​

  • llm_model: An instance of a language model client, configured for generating answers.
  • verbose (bool): A flag indicating whether to show print statements during execution.
  • additional_info: Additional context to be included in the HTML analysis.
  • reduction (int): The level of reduction to apply to the HTML content.
    • 0: Minification only.
    • 1: Minification and removal of unnecessary tags and attributes
    • 2: Minification, removal of unnecessary tags and attributes, simplification of text content, and removal of the <head> tag.

Methods​

  • __init__(self, input: str, output: List[str], node_config: Optional[dict] = None, node_name: str = "HtmlAnalyzer")

    • Initializes the HtmlAnalyzerNode with necessary configurations.
    • Args:
      • input (str): Boolean expression defining the input keys needed from the state.
      • output (List[str]): List of output keys to be updated in the state.
      • node_config (dict, optional): Additional configuration for the node.
      • node_name (str, optional): The unique identifier name for the node. Defaults to "HtmlAnalyzer".

  • execute(self, state: dict) -> dict

    • Generates an analysis of the provided HTML code based on the wanted information to be extracted and returns a reduced version of the HTML content. The analysis is primarily used to enhance the reasoning of the GenerateCodeNode.
    • The information to be extracted can be either the user's original input or, preferably, the refined prompt from the PromptRefinerNode for better results.
    • Args:
      • state (dict): The current state of the graph. The input keys will be used to fetch the correct data from the state.
    • Returns:
      • dict: The updated state with the following output keys:
        • The generated HTML analysis.
        • The reduced HTML code based on the specified reduction strength.
    • Raises:
      • KeyError: If the input keys are not found in the state, indicating that the necessary information for generating an answer is missing.

Example Usage​

Here is an example of how to use the HtmlAnalyzerNode class:

from html_analyzer_node import HtmlAnalyzerNode

# Define a html analyzer node
html_analyzer_node = HtmlAnalyzerNode(
    input="refined_prompt & original_html",
    output=["html_info", "reduced_html"]
)

# Define the state
state = {"refined_prompt": "What are the attractions in Chioggia?",
        "original_html": "<html>...</html>"}

# Execute the generate answer node
state = html_analyzer_node.execute(state)

# Retrieve the generated answer from the state
html_info = state["html_info"]
reduced_html = state["reduced_html"]

print(f"HTML Analysis: {html_info}")