Skip to main content

Pure Python SVG to Numpy 2D Array

Pure functional implimentation.
Should be easy to follow the implementation and extend.

Functionality:

  • Parse SVG file to Python List representing SVG paths
  • Generate list of 2D points from SVG paths
  • Convert SVG relative paths to absolute representation

Hosted on PyPI

pip install mountain-svgpath

Parse SVG paths

import svgpath

python_logo = """<?xml version="1.0" encoding="iso-8859-1"?>
<svg>		
  <path d="M11.298,8.02c1.295-0.587,1.488-5.055,0.724-6.371c-0.998-1.718-5.742-1.373-7.24-0.145
    C4.61,2.114,4.628,3.221,4.636,4.101h4.702v0.412H4.637c0,0.006-2.093,0.013-2.093,0.013c-3.609,0-3.534,7.838,1.228,7.838
    c0,0,0.175-1.736,0.481-2.606C5.198,7.073,9.168,8.986,11.298,8.02z M6.375,3.465c-0.542,0-0.981-0.439-0.981-0.982
    c0-0.542,0.439-0.982,0.981-0.982c0.543,0,0.982,0.44,0.982,0.982C7.358,3.025,6.918,3.465,6.375,3.465z"/>
  <path d="M13.12,4.691c0,0-0.125,1.737-0.431,2.606c-0.945,2.684-4.914,0.772-7.045,1.738
    C4.35,9.622,4.155,14.09,4.92,15.406c0.997,1.719,5.741,1.374,7.24,0.145c0.172-0.609,0.154-1.716,0.146-2.596H7.603v-0.412h4.701
    c0-0.006,2.317-0.013,2.317-0.013C17.947,12.53,18.245,4.691,13.12,4.691z M10.398,13.42c0.542,0,0.982,0.439,0.982,0.982
    c0,0.542-0.44,0.981-0.982,0.981s-0.981-0.439-0.981-0.981C9.417,13.859,9.856,13.42,10.398,13.42z"/>
</svg>"""

tree = svgpath.parse(python_logo)
for path, start, end in tree:
    print(f"Path: {start}-{end}")
    for token in path:
        print("\t", token)
Path: 72-500
         ['M', '11.298', '8.02']
         ['c', '1.295', '-0.587', '1.488', '-5.055', '0.724', '-6.371']
         ['c', '-0.998', '-1.718', '-5.742', '-1.373', '-7.24', '-0.145']
         ['C', '4.61', '2.114', '4.628', '3.221', '4.636', '4.101']
         ['h', '4.702']
         ['v', '0.412']
         ['H', '4.637']
         ['c', '0', '0.006', '-2.093', '0.013', '-2.093', '0.013']
         ['c', '-3.609', '0', '-3.534', '7.838', '1.228', '7.838']
         ['c', '0', '0', '0.175', '-1.736', '0.481', '-2.606']
         ['C', '5.198', '7.073', '9.168', '8.986', '11.298', '8.02']
         ['z']
         ['M', '6.375', '3.465']
         ['c', '-0.542', '0', '-0.981', '-0.439', '-0.981', '-0.982']
         ['c', '0', '-0.542', '0.439', '-0.982', '0.981', '-0.982']
         ['c', '0.543', '0', '0.982', '0.44', '0.982', '0.982']
         ['C', '7.358', '3.025', '6.918', '3.465', '6.375', '3.465']
         ['z']
Path: 515-944
         ['M', '13.12', '4.691']
         ['c', '0', '0', '-0.125', '1.737', '-0.431', '2.606']
         ['c', '-0.945', '2.684', '-4.914', '0.772', '-7.045', '1.738']
         ['C', '4.35', '9.622', '4.155', '14.09', '4.92', '15.406']
         ['c', '0.997', '1.719', '5.741', '1.374', '7.24', '0.145']
         ['c', '0.172', '-0.609', '0.154', '-1.716', '0.146', '-2.596']
         ['H', '7.603']
         ['v', '-0.412']
         ['h', '4.701']
         ['c', '0', '-0.006', '2.317', '-0.013', '2.317', '-0.013']
         ['C', '17.947', '12.53', '18.245', '4.691', '13.12', '4.691']
         ['z']
         ['M', '10.398', '13.42']
         ['c', '0.542', '0', '0.982', '0.439', '0.982', '0.982']
         ['c', '0', '0.542', '-0.44', '0.981', '-0.982', '0.981']
         ['s', '-0.981', '-0.439', '-0.981', '-0.981']
         ['C', '9.417', '13.859', '9.856', '13.42', '10.398', '13.42']
         ['z']

Convert to SVG Absolute

Breaking relative dependence between SVG tokens.
Allows for the partial display of an SVG path.

from svgpath import (
    tree_to_paths,
    absolute_paths,
)

# expand generator of generators (needed to compare generator outputs)
paths = [[token for token in path] for path in tree_to_paths(tree)]
apaths = [[token for token in path] for path in absolute_paths(paths)]

for i, (path, apath) in enumerate(zip(paths, apaths)):
    print(f"Path: {i}")
    for token, atoken in zip(path, apath):
        print(f"  {token} {atoken}")
Path: 0
  ['M', '11.298', '8.02'] ['M', '11.298', '8.02']
  ['c', '1.295', '-0.587', '1.488', '-5.055', '0.724', '-6.371'] ['C', '12.593', '7.433', '12.786', '2.965', '12.022', '1.649']
  ['c', '-0.998', '-1.718', '-5.742', '-1.373', '-7.24', '-0.145'] ['C', '11.024', '-0.069', '6.28', '0.276', '4.782', '1.504']
  ['C', '4.61', '2.114', '4.628', '3.221', '4.636', '4.101'] ['C', '4.61', '2.114', '4.628', '3.221', '4.636', '4.101']
  ['h', '4.702'] ['H', '9.338']
  ['v', '0.412'] ['V', '4.513']
  ['H', '4.637'] ['H', '4.637']
  ['c', '0', '0.006', '-2.093', '0.013', '-2.093', '0.013'] ['C', '4.637', '4.519', '2.544', '4.526', '2.544', '4.526']
  ['c', '-3.609', '0', '-3.534', '7.838', '1.228', '7.838'] ['C', '-1.065', '4.526', '-0.99', '12.364', '3.772', '12.364']
  ['c', '0', '0', '0.175', '-1.736', '0.481', '-2.606'] ['C', '3.772', '12.364', '3.947', '10.628', '4.253', '9.758']
  ['C', '5.198', '7.073', '9.168', '8.986', '11.298', '8.02'] ['C', '5.198', '7.073', '9.168', '8.986', '11.298', '8.02']
  ['z'] ['Z']
  ['M', '6.375', '3.465'] ['M', '6.375', '3.465']
  ['c', '-0.542', '0', '-0.981', '-0.439', '-0.981', '-0.982'] ['C', '5.833', '3.465', '5.394', '3.026', '5.394', '2.483']
  ['c', '0', '-0.542', '0.439', '-0.982', '0.981', '-0.982'] ['C', '5.394', '1.941', '5.833', '1.501', '6.375', '1.501']
  ['c', '0.543', '0', '0.982', '0.44', '0.982', '0.982'] ['C', '6.918', '1.501', '7.357', '1.941', '7.357', '2.483']
  ['C', '7.358', '3.025', '6.918', '3.465', '6.375', '3.465'] ['C', '7.358', '3.025', '6.918', '3.465', '6.375', '3.465']
  ['z'] ['Z']
Path: 1
  ['M', '13.12', '4.691'] ['M', '13.12', '4.691']
  ['c', '0', '0', '-0.125', '1.737', '-0.431', '2.606'] ['C', '13.12', '4.691', '12.995', '6.428', '12.689', '7.297']
  ['c', '-0.945', '2.684', '-4.914', '0.772', '-7.045', '1.738'] ['C', '11.744', '9.981', '7.775', '8.069', '5.644', '9.035']
  ['C', '4.35', '9.622', '4.155', '14.09', '4.92', '15.406'] ['C', '4.35', '9.622', '4.155', '14.09', '4.92', '15.406']
  ['c', '0.997', '1.719', '5.741', '1.374', '7.24', '0.145'] ['C', '5.917', '17.125', '10.661', '16.78', '12.16', '15.551']
  ['c', '0.172', '-0.609', '0.154', '-1.716', '0.146', '-2.596'] ['C', '12.332', '14.942', '12.314', '13.835', '12.306', '12.955']
  ['H', '7.603'] ['H', '7.603']
  ['v', '-0.412'] ['V', '12.543']
  ['h', '4.701'] ['H', '12.304']
  ['c', '0', '-0.006', '2.317', '-0.013', '2.317', '-0.013'] ['C', '12.304', '12.537', '14.621', '12.53', '14.621', '12.53']
  ['C', '17.947', '12.53', '18.245', '4.691', '13.12', '4.691'] ['C', '17.947', '12.53', '18.245', '4.691', '13.12', '4.691']
  ['z'] ['Z']
  ['M', '10.398', '13.42'] ['M', '10.398', '13.42']
  ['c', '0.542', '0', '0.982', '0.439', '0.982', '0.982'] ['C', '10.94', '13.42', '11.38', '13.859', '11.38', '14.402']
  ['c', '0', '0.542', '-0.44', '0.981', '-0.982', '0.981'] ['C', '11.38', '14.944', '10.94', '15.383', '10.398', '15.383']
  ['s', '-0.981', '-0.439', '-0.981', '-0.981'] ['S', '9.417', '14.944', '9.417', '14.402']
  ['C', '9.417', '13.859', '9.856', '13.42', '10.398', '13.42'] ['C', '9.417', '13.859', '9.856', '13.42', '10.398', '13.42']
  ['z'] ['Z']

Plot SVG Paths

# python script
from matplotlib import pyplot as plt
from svgpath import paths_to_points
paths = list(tree_to_paths(tree))

_, ax = plt.subplots()
ax.invert_yaxis() # origin at top left
for path in paths_to_points(paths, resolution=100):
    for trace in path:
        ax.plot(trace[:, 0], trace[:, 1])
plt.show()

The library includes a plotting script.
It requires matplotlib and pyqt6 be installed "pip install mountain-svgpath[display]".

svg-plot -i python_logo.svg

Python Logo

Metadata

Release files for mountain-svgpath 1.1.1

For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.

Source distribution (sdist)

Source distribution for mountain-svgpath 1.1.1
File Size Uploaded
mountain_svgpath-1.1.1.tar.gz 9.8 kB Details

Built distribution (wheel)

Table of built distributions (wheels) for mountain-svgpath 1.1.1
File Interpreter ABI Platform
mountain_svgpath-1.1.1-py3-none-any.whl Python 3 none any Details

Total release size: 21.1 kB

Release files / mountain_svgpath-1.1.1.tar.gz

Download URL mountain_svgpath-1.1.1.tar.gz
Size 9.8 kB
Tags Source
SHA-256 checksum
How to use checksums
a004f02114306dfea2d18833821bc0a754a21d82cbc466b4d67e89e4c2a32197
BLAKE2b-256 checksum
How to use checksums
114c74504bd03c1f11a099d3664dc0b1734a738aa7ce89d662eaf846c435268c
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
Yes
Uploaded via twine/6.1.0 CPython/3.13.12

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

PyPI Publish Attestation

PyPI verified that this artifact, at this checksum, originated from the publisher listed below.

Signed by GitHub Actions, verified by PyPI on Jun 7, 2026.

Transparency log

Release files / mountain_svgpath-1.1.1-py3-none-any.whl

Download URL mountain_svgpath-1.1.1-py3-none-any.whl
Size 11.3 kB
Tags Python 3
SHA-256 checksum
How to use checksums
ba258e2d688debbcb635e185a56c9b2eeae8ff92328e54092dd56188a79c1e68
BLAKE2b-256 checksum
How to use checksums
e1a1b9e840ab33eb27bf24bc9ea3a0d86139d92ccac703c21480146c7718e9ef
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
Yes
Uploaded via twine/6.1.0 CPython/3.13.12

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

PyPI Publish Attestation

PyPI verified that this artifact, at this checksum, originated from the publisher listed below.

Signed by GitHub Actions, verified by PyPI on Jun 7, 2026.

Transparency log

Release history Release notifications | RSS feed

This release

1.1.1 This release

2 release files

1.1.0

2 release files

1.0.1

2 release files

1.0.0

2 release files

Anthropic, PBC Visionary sponsor Bloomberg Visionary sponsor Hudson River Trading Visionary sponsor Meta Visionary sponsor NVIDIA Visionary sponsor Microsoft Sustainability sponsor Depot Continuous Integration AWS Cloud computing and Security Sponsor Datadog Monitoring Fastly CDN Google Download Analytics Sentry Error logging StatusPage Status page