Random Number MCP

Instructions

Implementation Reference

• src/random_number_mcp/server.py:42-67 (handler)

  MCP tool handler for 'random_choices'. Handles optional JSON string weights and delegates core logic to tools.random_choices.

  @app.tool()
  def random_choices(
      population: list[Any],
      k: int = 1,
      weights: list[int | float] | str | None = None,
  ) -> list[Any]:
      """Choose k items from population with replacement, optionally weighted.
  
      Args:
          population: List of items to choose from
          k: Number of items to choose (default 1)
          weights: Optional weights for each item (default None for equal weights)
  
      Returns:
          List of k chosen items
      """
      numeric_weights: list[int | float] | None = None
      if isinstance(weights, str):
          try:
              numeric_weights = json.loads(weights)
          except json.JSONDecodeError as e:
              raise ValueError(f"Invalid JSON string for weights: {weights}") from e
      else:
          numeric_weights = weights
      return tools.random_choices(population, k, numeric_weights)

• src/random_number_mcp/tools.py:57-80 (helper)

  Core helper function implementing random choices logic using Python's random.choices, with input validation.

  def random_choices(
      population: list[Any], k: int = 1, weights: list[int | float] | None = None
  ) -> list[Any]:
      """Choose k items from population with replacement, optionally weighted.
  
      Args:
          population: List of items to choose from
          k: Number of items to choose (default 1)
          weights: Optional weights for each item (default None for equal weights)
  
      Returns:
          List of k chosen items
  
      Raises:
          ValueError: If population is empty, k < 0, or weights length doesn't match
          TypeError: If k is not an integer
      """
      validate_list_not_empty(population, "population")
      validate_positive_int(k, "k")
  
      if weights is not None:
          validate_weights_match_population(population, weights)
  
      return random.choices(population, weights=weights, k=k)