Tutorial - Installation

Installation Using Google Colab

Open the Jupyter Notebook file found here within Google Colab: https://bitbucket.org/ccsg-res/ticktalkpython/src/tutorial/CAVExamples.ipynb

First, run “Step 1” that will setup TTPython dependencies and external dependencies used in the tutorial.

Installation

TTPython runs and has been tested on Python >=3.8. We recommend using a virtual environment manager like Conda or python virtual environments

Clone the TTPython repository to your local directory

git clone https://bitbucket.org/ccsg-res/ticktalkpython.git

Navigate to the tutorial branch and follow the steps:

  1. Install the dependencies for TTPython

pip install -r requirements.txt

If you wish to use the graphviz package for graph visualization during compilation (when calling compile, setting use_graphviz=True), then graphviz must be installed with a few extra steps. The package “graphviz” will be installed by pip, but a corollary (and necessary) package pygraphviz requires a few additional steps. On Mac, we find ‘brew’ to be the best avenue for the initial build process, which requires more than just pip or conda installations.

  1. Basic Testing and Verification

In the top level of the repo, there is a jupyter notebook that will compile and simulate a simple TTPython example to help get you started. This is helpful for verifying the installation process worked correctly.

jupyter notebook TickTalkTest.ipynb

  1. Open the Jupyter notebook and run the first two blocks within the notebook

The first will run a set of basic regression tests and compile a TTPython program in the ‘examples’ directory called streaming_merge. This program simply produces two streams of values representing sinusoids sampled every 500 ms, and adds them together. It also computes a moving average on one of the sinusoid streams.

The second block in the notebook will interpret the graph within a simulated environment. Note that this may be quite slow within jupyter, taking up to 40 seconds to finish (running from terminal is about 2 orders of magnitude faster). If it finishes correctly, it will produce two output graphs depicting the added sinusoids and the moving average.

FAQ:

Can I run TTPython in Google Colab environment?

Yes, Colab uses python version 3.7 as default. We have tested TTPython with versions 3.7 to 3.9.

Why am I getting a ‘No module named’ error while running the TickTalkTest.ipynb notebook?

Make sure that all the requirements in the requirements.txt file are installed. If you are using a conda environment, you may want to recheck the python base version.

Also note that while using a conda environment, the ast_scope package requires pip since conda package is not available.

pip install ast_scope

Additionally, graphviz requires two additional ‘python-graphviz’ and ‘pydot’ packages if you are using conda.

conda install python-graphviz
conda install pydot