Install py2 to global conda env

[pdmack]

Fixes #906

PTAL
/cc @jlewi
/cc @ankushagarwal

Are we good with just dropping the --user py2 install and letting it get installed to /opt/conda/envs? Are there some other tests I should try?

Also, should we install the Beam pkg @jlewi was referring to in #906 while we're here?

---

This change is 

[ankushagarwal]

Installing beam SGTM

[ankushagarwal]

My bad. tensorflow-model-analysis seems to install apache-beam already : https://github.com/tensorflow/model-analysis/blob/master/setup.py#L189

[pdmack]

Yep, I misinterpreted @jlewi's comment. Doing some more testing.

[pdmack]

@ankushagarwal Do you/we have a particular build environment for the NB images when TFMA is enabled? I consistently get a SIGILL from your post-refactor build in my env:

/bin/bash: line 1:   283 Illegal instruction     (core dumped) jupyter nbextension install --py --symlink tensorflow_model_analysis

This happens even in the NB container itself if I try to manually install it.

[pdmack]

OK figured out that my build CPU env wasn't up to snuff. If there's no AVX support then that extension install fails because it tries to do a TF import. Will do more py2 testing on another env tomorrow.

[jlewi]

Thank you!

[pdmack]

I think this is OK now. @jlewi can you try this out and see if it's satisfactory?

docker pull pdmack/tf-notebook:1.7.0

[jlewi]

Confirmed imports work in a Python 2 environment

import apache_beam
import tensorflow_model_analysis as tfma

/lgtm
/approve

[k8s-ci-robot]

[APPROVALNOTIFIER] This PR is APPROVED

This pull-request has been approved by: jlewi

The full list of commands accepted by this bot can be found here.

The pull request process is described here

Needs approval from an approver in each of these files:

• OWNERS [jlewi]

Approvers can indicate their approval by writing /approve in a comment
Approvers can cancel approval by writing /approve cancel in a comment