Fermat MCP

Instructions

Implementation Reference

• fmcp/mpl_mcp/core/plot_chart.py:8-110 (handler)

  The handler function `plot_chart` that implements the core logic for plotting line, scatter, or bar charts using matplotlib. It processes input data, creates the plot, customizes it, and returns a PNG image as FastMCP Image object.
  def plot_chart( x_data: List[Union[float, int]], y_data: Union[List[Union[float, int]], List[List[Union[float, int]]]], plot_type: str = "line", # "line", "scatter", or "bar" labels: Optional[Union[str, List[str]]] = None, title: str = "", xlabel: str = "", ylabel: str = "", color: Union[str, List[str]] = "skyblue", save: bool = False, dpi: int = 200, figsize: Optional[List[Union[int, float]]] = None, grid: bool = True, legend: bool = False, ) -> Image: """ Create a customizable plot with support for multiple plot types. Args: x_data: X-axis data points (1D array-like) y_data: Y-axis data points (1D or 2D array-like for multiple series) plot_type: Type of plot to create ("line", "scatter", or "bar") labels: Label or list of labels for the data series title: Plot title xlabel: Label for the x-axis ylabel: Label for the y-axis color: Color or list of colors for the plot elements save: If True, save the figure to a buffer dpi: Output image resolution (dots per inch, default: 200) figsize: List of width and height in inches grid: Whether to show grid lines legend: Whether to show legend Returns: FastMCP Image object with the plotted chart """ # Convert inputs to numpy arrays for processing x = np.asarray(x_data, dtype=float) y = np.asarray(y_data, dtype=float) # Handle 1D y_data case by adding an extra dimension if y.ndim == 1: y = y.reshape(-1, 1) # Ensure x matches the number of data points in y if x.ndim == 1 and len(x) != y.shape[0]: x = np.tile(x, (y.shape[1], 1)).T # Handle labels — normalize to a list of strings if labels is None: labels_list: List[str] = [""] * y.shape[1] elif isinstance(labels, str): labels_list = [labels] else: labels_list = list(labels) # Handle colors if isinstance(color, str): color = [color] * y.shape[1] # Create figure with specified size using OO interface # Normalize figsize to a tuple of floats (matplotlib accepts floats or ints) if figsize and len(figsize) >= 2: figsize_vals = (float(figsize[0]), float(figsize[1])) else: figsize_vals = (6.0, 4.0) fig, ax = plt.subplots(figsize=figsize_vals, dpi=dpi) # Create the appropriate plot type for i in range(y.shape[1]): current_label = labels_list[i] if i < len(labels_list) else f"Series {i+1}" current_color = color[i % len(color)] if plot_type == "line": ax.plot(x, y[:, i], label=current_label, color=current_color) elif plot_type == "scatter": ax.scatter(x, y[:, i], label=current_label, color=current_color) elif plot_type == "bar": ax.bar(x, y[:, i], label=current_label, color=current_color) else: plt.close(fig) raise ValueError(f"Unsupported plot type: {plot_type}") # Customize the plot ax.set_title(title) ax.set_xlabel(xlabel) ax.set_ylabel(ylabel) if grid: ax.grid(True, linestyle="--", alpha=0.7) if legend and any(labels_list): ax.legend() # Save the plot to a buffer and close the figure buf = io.BytesIO() fig.savefig(buf, format="png", dpi=dpi, bbox_inches="tight") plt.close(fig) buf.seek(0) return Image(data=buf.read(), format="png")
• fmcp/mpl_mcp/server.py:21-21 (registration)

  Registers the `plot_chart` function as an MCP tool in the FastMCP server named 'mpl_mcp', likely naming the tool 'mpl_mcp_plot_chart' based on server prefix and function name.
  mpl_mcp.tool(plot_chart, description="Plots line/scatter/bar chart of given datavalues")
• fmcp/mpl_mcp/server.py:4-4 (registration)

  Imports the `plot_chart` handler function required for tool registration.
  from .core.plot_chart import plot_chart