Plotting

ezdwg can render DWG entities using matplotlib. Install the plotting extra:

pip install "ezdwg[plot]"

Basic Usage

From the Command Line

ezdwg plot drawing.dwg
ezdwg plot drawing.dwg -o drawing.png --dpi 200
ezdwg plot drawing.dwg -o drawing.svg --types "LINE ARC"

Without -o, the command opens an interactive matplotlib window, or an SVG preview in your browser if no interactive backend is available. With -o, it saves the drawing without opening a viewer. See the CLI reference.

From a Document

import ezdwg

doc = ezdwg.read("drawing.dwg")
doc.plot()

From a Layout

msp = doc.modelspace()
msp.plot()

From a File Path

import ezdwg

ezdwg.plot("drawing.dwg")

Parameters

The plot() function accepts the following parameters:

Parameter	Type	Default	Description
types	str \| None	None	Entity type filter (e.g. "LINE ARC")
ax	matplotlib.axes.Axes \| None	None	Existing axes to draw on
show	bool	True	Call plt.show() after drawing
equal	bool	True	Use equal aspect ratio
title	str \| None	None	Plot title
line_width	float	0.5	Line width for geometry, in points
arc_segments	int	64	Segments for arc approximation
auto_fit	bool	True	Auto-fit view to content
fit_margin	float	0.04	Margin around content (fraction)
dimension_color	Any \| None	"black"	Color for dimension entities

Saving to a File

Set show=False and use matplotlib to save:

import matplotlib.pyplot as plt
import ezdwg

doc = ezdwg.read("drawing.dwg")
ax = doc.plot(show=False)
plt.savefig("output.png", dpi=150, bbox_inches="tight")

Plotting Specific Types

# Only lines and arcs
doc.plot(types="LINE ARC")

# Only text entities
doc.plot(types="TEXT MTEXT")

Custom Axes

import matplotlib.pyplot as plt

fig, (ax1, ax2) = plt.subplots(1, 2, figsize=(16, 8))

doc = ezdwg.read("drawing.dwg")
doc.plot(types="LINE", ax=ax1, show=False, title="Lines")
doc.plot(types="ARC CIRCLE", ax=ax2, show=False, title="Arcs & Circles")

plt.tight_layout()
plt.show()

Color Handling

ezdwg resolves entity colors from:

1. True color (24-bit RGB) — if present on the entity
2. ACI color index — AutoCAD Color Index (1–255)
3. Layer color — inherited from the entity's layer

Colors are applied automatically when plotting. ACI index 7 (white/black) is rendered as black for visibility on matplotlib's default light background. On near-white backgrounds, very bright colors are darkened while retaining their hue so yellow, cyan, and green geometry remains visible. Entity color values in the document are unchanged.

Drawing Presentation

Geometry uses thin 0.5-point strokes by default. Dimension strokes are thinner than geometry; their ticks and extension overshoot follow text height rather than the measured length. Dimension text uses its saved attachment point. These are preview styles, not a reproduction of DWG lineweight or plot styles.

Automatic fitting keeps the drawing's rectangular bounds while maintaining equal X/Y scale. The CLI uses a larger canvas with coordinate axes hidden and gray dimensions. Python callers can continue to customize the returned Axes.

Text Size and Dimensions

Drawing text is rendered as vector outlines in drawing coordinates. Capital height follows the DWG text height, so text scales together with geometry when zooming, resizing, or saving at a different DPI. Text outlines and backgrounds are included in automatic view bounds, including for text-only drawings. These outlines are matplotlib patches rather than entries in ax.texts.

Dimension text uses the saved height from its referenced anonymous block when the block's decoded text has one consistent, positive height. The high-level entity exposes this as char_height, with char_height_source="anonymous_block". If no height can be resolved, the plot uses a fallback of one drawing unit; it does not infer text height from the dimension length. This does not regenerate dimension styles or fully reproduce CAD text layout and fonts.

Block references (INSERT / MINSERT) are currently shown as small, faint crosses at their insertion points, rather than filled circles. Their block geometry is not expanded by the plotting API. Actual POINT entities use tiny dots; actual CIRCLE entities retain their decoded radius.