Installing packages with micropip#
micropip can install following types of packages:
Pure Python packages
Pure Python packages are packages that do not have any compiled code.
Most pure Python packages can be installed directly from PyPI with
if they have a pure Python wheel. Check if this is the case by trying
Python packages that contain C-extensions
If a package has C-extensions (or any other compiled codes like Rust), it will not have a pure Python wheel on PyPI.
Trying to install such a package with
micropip.install will result in an error like:
ValueError: Can't find a pure Python 3 wheel for 'tensorflow'. See: https://pyodide.org/en/stable/usage/faq.html#micropip-can-t-find-a-pure-python-wheel You can use `micropip.install(..., keep_going=True)`to get a list of all packages with missing wheels.
To install such a package, you need to first build a Python wheels for WASM/Emscripten for it.
Note that pyodide provides several commonly used packages with pre-built wheels.
Those packages can be installed with
You can find a list of packages with pre-built wheels in the Pyodide documentation. If your package is not in the list, you can build a wheel for it yourself. See the Building packages section of the Pyodide documentation for more information.
import micropip # snoballstemmer is a pure Python package # and has a pure Python wheel on PyPI # so it can be installed directly await micropip.install("snowballstemmer") # numpy is a package that has C-extensions # and does not have a pure Python wheel on PyPI # but it is provided by Pyodide await micropip.install("numpy") # It is also possible to install from # - arbitrary URLs await micropip.install("https://.../package.whl") # - local files inside the Pyodide virtual file system await micropip.install("emfs://.../package.whl")
You can pass multiple packages to
await micropip.install(["pkg1", "pkg2"])
You can specify additional constraints:
await micropip.install("snowballstemmer==2.2.0") await micropip.install("snowballstemmer>=2.2.0") await micropip.install("snowballstemmer[all]")
micropip does dependency resolution by default, but you can disable it, this is useful if you want to install a package that has a dependency which is not a pure Python package, but it is not mandatory for your use case:
await micropip.install("pkg", deps=False)