How to install numpy and scipy for Ironpython27? Old method doesn't work

[COMMENT BY ENTHOUGHT SUPPORT: The link in this answer is not valid. See answer below by Jonathan March on August 17, 2018]

For those struggling to get numpy/scipy install for ironpythopn, enthought have moved the download link to https://store.enthought.com/repo/.iron/ . The link would only allow you in if you are registered.

Therefore first up you'd have to register yourself for free, then open the above link, then follow the steps below

1. Download the IronPython-2.7.msi and install it.

2. Download ironpkg-1.0.0.py from the above link.

3. Using command line navigate to the directory where you placed ironpkg-1.0.0.py and run ipy ironpkg-1.0.0.py --install Check whether the install worked using ironpkg -h

4. The last step is lightly different to the one suggested by enthoughts. Running ironpkg scipy won't work as it looks at the old web address for download. Instead download all the eggs and index-depend.txt from the above link. For installation to work, you would have to modify the download location in the config file to point to the local drive instead of website. The config file can be found at user directory eg.C:\Users\Nilster\.ironpkg . Open it in the textpad and change the location to directory where you downloaded the eggs Eg, mine looks like

   IndexedRepos = ['file://C:\Work\Python\Enthought_Eggs',]

5. Then run the following to install numpy/scipy ironpkg scipy

6. Check whether the install worked using ipy -X:Frames -c "import scipy"

Enthought support here.

The Iron Python numpy and scipy packages can be downloaded here: http://code.enthought.com/.iron/README.txt http://code.enthought.com/.iron/eggs/index.html

FYI, Microsoft stopped work on the IronPython project in 2012 in favor of supporting standard CPython. Those archived versions of numpy and scipy were built in 2011 (so contain no newer features or fixes), and are 32-bit-only. We do not plan to update them in any way.

FYI, we typically recommend that those who wish to use Python in a .net context consider using the actively developed pythonnet package to interface with the living CPython ecosystem.