Build Feast Jupyter image and clean up examples

[woop]

What this PR does / why we need it:

• Builds a base image for Jupyter (gcr.io/kf-feast/feast-jupyter:latest) so that it's faster for new users to start the Jupyter container for Docker Compose. Feast dependencies are still installed upon startup, but most of them should now already be available.
• Cleaned up the Basic example a little bit.
• Fixed entity_df bug in Telco example and added more details

Fixes: #742

[algattik]

Funny, I've done exactly the same thing today on our fork :)

Here's my dockerfile (maximizing layer reuse). It seems to me that there is no need anymore for the startup script to download and install stuff, and downloading release versions in the dev cycle makes config management harder?

FROM jupyter/minimal-notebook:619e9cc2fc07

USER root

WORKDIR /feast

# Init empty .git settings (required by SDK setup.py)
RUN git init .

COPY sdk/python sdk/python
COPY Makefile Makefile
COPY protos protos
COPY tests tests
COPY README.md README.md

# Install Python dependencies
RUN make compile-protos-python

# Install CI requirements (only needed for running tests)
RUN pip install -r sdk/python/requirements-ci.txt

# Install Feast SDK
RUN pip install -e sdk/python -U

# Switch back to original user and workdir
USER $NB_UID
WORKDIR $HOME

CMD start-notebook.sh --NotebookApp.token=''

[woop]

@algattik neat!

I might reuse some of yours. I will still be cloning the repository since the image I am building is not meant for development, it's meant to have the latest stable dependencies preinstalled, and I guess we can have the actual SDK also preinstalled.

I don't like mapping in the examples directly from the repo because master is normally not consistent with the stable branch, and if we pin versions then its very hard to backport tutorials or make small tweaks.

[woop]

@algattik Made some changes to find a better middle ground. Hope it looks better :)

I didn't attribute the Dockerfile to you, but normally people don't care 🎖️

[woop]

/test test-end-to-end-batch

[mrzzy]

Is there a reason why can't we bake in the Feast repo into the docker image?

[woop]

The latest image will only be built on a release, which means that if we want to roll out the latest tutorials then we have to cut a release. Often I find myself updating docs/tutorials between releases, which is why I have pinned this clone to the branch instead. In theory we could build the latest image on every commit, but that image will go stale very quickly on the user's side (they have to repull the whole time).

[woop]

/test test-end-to-end-batch

[algattik]

.git and README.md are only needed by " pip install -e sdk/python" IIRC, so if you move them down you can reuse layer with requirements installed. (.git folder changes with every commit)

[woop]

Corrected, thanks.

[algattik]

@algattik Made some changes to find a better middle ground. Hope it looks better :)

I didn't attribute the Dockerfile to you, but normally people don't care 🎖️

My ego shall survive :)

[zhilingc]

/approve

[feast-ci-bot]

[APPROVALNOTIFIER] This PR is APPROVED

This pull-request has been approved by: woop, zhilingc

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 [woop,zhilingc]

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

[feast-ci-bot]

@woop: Updated the config configmap in namespace default at cluster default using the following files:

• key config.yaml using file .prow/config.yaml

In response to this:

What this PR does / why we need it:

• Builds a base image for Jupyter (gcr.io/kf-feast/feast-jupyter:latest) so that it's faster for new users to start the Jupyter container for Docker Compose. Feast dependencies are still installed upon startup, but most of them should now already be available.
• Cleaned up the Basic example a little bit.
• Fixed entity_df bug in Telco example and added more details

Fixes: #742

Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes/test-infra repository.