import matplotlib
matplotlib.use('Agg')
import matplotlib.pyplot as plt
import io
import numpy as np
from .models import *
# Make output deterministic
matplotlib.rcParams['svg.hashsalt'] = 'plot-mcp'
def _setup_figure(width: float, height: float, title: str = None, x_label: str = None, y_label: str = None, is_polar=False):
"""Setup figure with simplified parameters"""
# dpi=72 is standard for screen
dpi = 72
fig = plt.figure(figsize=(width / dpi, height / dpi), dpi=dpi)
# Transparent background
fig.patch.set_alpha(0.0)
# Default margins
margin_left = 50 / width
margin_right = 1.0 - (20 / width)
margin_bottom = 40 / height
margin_top = 1.0 - (40 / height)
fig.subplots_adjust(left=margin_left, bottom=margin_bottom, right=margin_right, top=margin_top)
if is_polar:
ax = fig.add_subplot(111, projection='polar')
else:
ax = fig.add_subplot(111)
if title:
ax.set_title(title)
# Axis labels
if not is_polar:
if x_label:
ax.set_xlabel(x_label)
if y_label:
ax.set_ylabel(y_label)
return fig, ax
def _finalize(fig, width: float, height: float) -> PlotOutput:
"""Finalize and return plot output"""
f = io.StringIO()
fig.savefig(f, format='svg', transparent=True)
plt.close(fig)
full_svg = f.getvalue()
# Extract <svg ... > ... </svg>
start_idx = full_svg.find("<svg")
if start_idx == -1:
svg_content = full_svg
else:
svg_content = full_svg[start_idx:]
return PlotOutput(
svg=svg_content,
width=width,
height=height,
viewBox=f"0 0 {width} {height}"
)
def render_line(params: LineParams) -> PlotOutput:
"""Render line plot with flat parameters"""
fig, ax = _setup_figure(params.width, params.height, params.title, params.x_label, params.y_label)
for series in params.series:
# Map style names to matplotlib
mpl_ls = "-"
if params.line_style == "dashed":
mpl_ls = "--"
elif params.line_style == "dotted":
mpl_ls = ":"
ax.plot(
series.x,
series.y,
label=series.name,
linestyle=mpl_ls,
linewidth=params.stroke_width,
marker='o' if params.show_markers else None
)
if len(params.series) > 1:
ax.legend()
return _finalize(fig, params.width, params.height)
def render_scatter(params: ScatterParams) -> PlotOutput:
"""Render scatter plot with flat parameters"""
fig, ax = _setup_figure(params.width, params.height, params.title, params.x_label, params.y_label)
ax.scatter(
params.x, params.y,
s=params.point_radius ** 2, # s is area
c=params.color,
alpha=params.opacity
)
return _finalize(fig, params.width, params.height)
def render_bar(params: BarParams) -> PlotOutput:
"""Render bar chart with flat parameters"""
fig, ax = _setup_figure(params.width, params.height, params.title, params.x_label, params.y_label)
if params.orientation == "vertical":
ax.bar(params.categories, params.values, width=params.bar_width, color=params.color)
else:
ax.barh(params.categories, params.values, height=params.bar_width, color=params.color)
return _finalize(fig, params.width, params.height)
def render_area(params: AreaParams) -> PlotOutput:
"""Render area plot with flat parameters"""
fig, ax = _setup_figure(params.width, params.height, params.title, params.x_label, params.y_label)
ax.fill_between(params.x, params.y, color=params.fill_color, alpha=params.opacity)
return _finalize(fig, params.width, params.height)
def render_histogram(params: HistogramParams) -> PlotOutput:
"""Render histogram with flat parameters"""
fig, ax = _setup_figure(params.width, params.height, params.title, params.x_label, params.y_label)
ax.hist(
params.values,
bins=params.bins,
density=params.density,
color=params.color,
edgecolor='black'
)
return _finalize(fig, params.width, params.height)
def render_box(params: BoxParams) -> PlotOutput:
"""Render box plot with flat parameters"""
fig, ax = _setup_figure(params.width, params.height, params.title, params.x_label, params.y_label)
# Prepare data for boxplot
values = [g.values for g in params.groups]
labels = [g.name for g in params.groups]
bp = ax.boxplot(
values,
labels=labels,
widths=params.box_width,
patch_artist=True
)
for patch in bp['boxes']:
patch.set_facecolor(params.color)
return _finalize(fig, params.width, params.height)
def render_heatmap(params: HeatmapParams) -> PlotOutput:
"""Render heatmap with flat parameters"""
fig, ax = _setup_figure(params.width, params.height, params.title)
matrix = np.array(params.matrix)
im = ax.imshow(
matrix,
cmap=params.color_scale,
aspect='auto'
)
# Tick labels
ax.set_xticks(np.arange(len(params.x_labels)), labels=params.x_labels)
ax.set_yticks(np.arange(len(params.y_labels)), labels=params.y_labels)
# Rotate the tick labels
plt.setp(ax.get_xticklabels(), rotation=45, ha="right", rotation_mode="anchor")
if params.show_values:
for i in range(len(params.y_labels)):
for j in range(len(params.x_labels)):
ax.text(j, i, f"{matrix[i, j]:.2f}",
ha="center", va="center", color="w")
return _finalize(fig, params.width, params.height)
def render_contour(params: ContourParams) -> PlotOutput:
"""Render contour plot with flat parameters"""
fig, ax = _setup_figure(params.width, params.height, params.title, params.x_label, params.y_label)
X, Y = np.meshgrid(params.x, params.y)
Z = np.array(params.z)
cs = ax.contour(
X, Y, Z,
levels=params.levels,
linewidths=params.stroke_width
)
ax.clabel(cs, inline=True, fontsize=10)
return _finalize(fig, params.width, params.height)
def render_pie(params: PieParams) -> PlotOutput:
"""Render pie chart with flat parameters"""
fig, ax = _setup_figure(params.width, params.height, params.title, is_polar=False)
wedgeprops = {}
if params.inner_radius_ratio > 0:
wedgeprops['width'] = 1.0 - params.inner_radius_ratio
ax.pie(
params.values,
labels=params.labels,
startangle=params.start_angle,
wedgeprops=wedgeprops
)
return _finalize(fig, params.width, params.height)
