Pyodide#

Overview#

Pyodide is a client-side Python implementation that runs in a web browser.
The Pyodide build of PyMuPDF is currently experimental.

Building a PyMuPDF wheel for Pyodide#

A PyMuPDF wheel for Pyodide can be built by running scripts/gh_release.py with some environmental variable settings. This is regularly tested on Github by .github/workflows/test_pyodide.yml.

Here is an example of this, a single Linux command (to be run with the current directory set to a PyMuPDF checkout), that builds a Pyodide wheel:

inputs_sdist=0 \
inputs_PYMUPDF_SETUP_MUPDF_BUILD="git:--recursive --depth 1 --shallow-submodules --branch master https://github.com/ArtifexSoftware/mupdf.git" \
inputs_wheels_default=0 \
inputs_wheels_linux_pyodide=1 \
inputs_wheels_implementations=b \
./scripts/gh_release.py build

This does the following (all inside Python venv’s):

Download (git clone and pip install) and customise a Pyodide build environment.
Download (git clone) the latest MuPDF.
Build MuPDF and PyMuPDF in the Pyodide build environment.
Create a wheel in dist/.

For more information, see the comments for functions build_pyodide_wheel() and pyodide_setup() in scripts/gh_release.py.

Using a Pyodide wheel#

Upload the wheel (for example PyMuPDF/dist/PyMuPDF-1.24.2-cp311-cp311-emscripten_3_1_32_wasm32.whl) to a webserver which has been configured to allow Cross-origin resource sharing (https://en.wikipedia.org/wiki/Cross-origin_resource_sharing).
The wheel can be used in a Pyodide console running in a web browser, or a JupyterLite notebook running in a web browser.
- To create a Pyodide console, go to:
  
  https://pyodide.org/en/stable/console.html
- To create a JupyterLite notebook, go to:
  
  https://jupyterlite.readthedocs.io/en/latest/_static/lab/index.html
In both these cases, one can use the following code to download the wheel (replace url with the URL of the uploaded wheel) and import it:
```
import pyodide_js
await pyodide_js.loadPackage(url)
import fitz
```
- Note that micropip.install() does not work, because of PyMuPDF’s use of shared libraries.

Loading a PDF document from a URL into PyMuPDF#

Pyodide browser console does not have generic network access, so for example urllib.request.urlopen(url) fails. But Pyodide has a built-in pyodide.http module that uses javascript internally, which one can use to download into a bytes instance, which can be used to create a PyMuPDF Document instance:
```
import pyodide.http
r = await pyodide.http.pyfetch('https://...')
data = await r.bytes()
doc = fitz.Document(stream=data)
```
It looks like this only works with https://, not http://.

Do you have any feedback on this page?

This software is provided AS-IS with no warranty, either express or implied. This software is distributed under license and may not be copied, modified or distributed except as expressly authorized under the terms of that license. Refer to licensing information at artifex.com or contact Artifex Software Inc., 39 Mesa Street, Suite 108A, San Francisco CA 94129, United States for further information.