pyTCPlotter

Michael S. Titus, Austin Hernandez, Victoria Tucker, Akhil Bejjipurapu, School of Materials Engineering, Purdue University, West Lafayette, IN 47907

calculate, plot, and save predictions from Thermo-Calc easily and efficiently

This code will work on the Purdue RCAC Bell or Negishi cluster.

Please follow one of the three methods to install the software:

1) Create anaconda environment and install packages manually.(Most Preffered. Method)

1a) open a jupyter notebook session:

• go to: https://www.rcac.purdue.edu/compute/bell or https://www.rcac.purdue.edu/compute/negishi

• login using your Purdue credentials

  • Gateway(Launch) => Top(Interactive Apps) => jupyter

  • To Launch Jupyter Notebook session on bell Cluster, select the queue from drop-down menu, Number of hours (up to user) and Processor Cores requested (choose 64)

  • Click launch, this starts a Jupyter Notebook session , then click on connect to jupyter

• open up a new terminal window by selecting "New" -> "Terminal" with the dropdown menu in the top right module load anaconda/2022.10-py39

conda create --name tcplotter24a

conda activate tcplotter24a

conda install pip=9.0.1 #check versioning

conda install matplotlib ipython jupyter pandas

1b) Clone the respository:

• enter: $ git clone https://github.itap.purdue.edu/michaeltitusgroup/pyTCPlotter

• move into the directory: $ cd pyTCPlotter

pip install -r requirements.txt

ipython kernel install --user --name tcplotter24a --display-name "tcplotter24a"

1c) Open the workflow notebook:

• return to the Jupyter notebook tab, enter into the tcplotter folder

• open the tc_plotter_workflow.ipynb notebook

• after opening your notebook or python code, from the top bar, select "Kernel" -> "Change kernel" -> "tcplotter24a"

• enjoy!

2) Automatically install a conda environment

2a) open a jupyter notebook session:

• go to: https://www.rcac.purdue.edu/compute/bell or https://www.rcac.purdue.edu/compute/negishi

• login using your Purdue credentials

  • Gateway(Launch) => Top(Interactive Apps) => jupyter

  • To Launch Jupyter Notebook session on bell Cluster, select the queue from drop-down menu, Number of hours (up to user) and Processor Cores requested (choose 64)

  • Click launch, this starts a Jupyter Notebook session , then click on connect to jupyter

• open up a new terminal window by selecting "New" -> "Terminal" with the dropdown menu in the top right

2b) Clone the respository:

• enter: $ git clone https://github.itap.purdue.edu/michaeltitusgroup/pyTCPlotter

• move into the directory: $ cd pyTCPlotter

2c) Change the setup file prefix to your username:

• enter: sed -i 's/\[USERNAME\]/YourUsername/g' environment.yml

2d) Make and activate the anaconda environment and ipython kernel:

• enter: $ conda env create -f environment.yml

• source the new environment: $ conda activate tcplotter24a

• enter: $ ipython kernel install --user --name tcplotter24a --display-name "tcplotter24a"

2e) Open the workflow notebook:

• return to the Jupyter notebook tab, enter into the tcplotter folder

• open the tc_plotter_workflow.ipynb notebook

• after opening your notebook or python code, from the top bar, select "Kernel" -> "Change kernel" -> "tcplotter24a"

• enjoy!

3) Install packages without anaconda environment (not recommended):

3a) Open up Bell terminal:

• go to https://gateway.bell.rcac.purdue.edu/pun/sys/dashboard or https://gateway.negishi.rcac.purdue.edu/pun/sys/dashboard

• Top middle, Cluster => >_Bell Shell Access

• Type “yes” to accept connection

• Type in your Purdue password (not boiler key)

3b) Load the correct modules version:

• Type in:

$ module load anaconda

• To make sure this worked, you can type:

$ pip --version

It should return the following line:

$ pip 9.0.1 from /apps/cent7/anaconda/5.1.0-py36-new/lib/python3.6/site-packages (python 3.6)

This means that the python 3.6 version and pip 9.0.1 is currently running.

3c) install the required python modules:

$ pip install --user /depot/coe-mse/apps/thermocalc/2024a/SDK/TC-Python/TC_Python-2024.1.30332-76-py3-none-any.whl

Install ternary-python module:

$ pip install --user python-ternary

Move the directory (as there is another commonly-used module named ternary):

$ mv .local/lib/python3.6/site-packages/ternary .local/lib/python3.6/site-packages/ternary_plotter

3d) Update your PYTHONPATH variable to include the local python3.6 libraries:

$ echo 'export PYTHONPATH=$PYTHONPATH:/home/<username>/.local/lib/python3.6/site-packages' >> ~/.bash_profile

where username is your Purdue career account name

3e) re-source your bash profile:

$ source .bash_profile

3f) test the installation by typing:

$ python import tc_python

If you receive no errors, then you are good to go!

3g) Clone the respository:

$ git clone https://github.itap.purdue.edu/michaeltitusgroup/pyTCPlotter

3h) open a jupyter notebook session and begin:

go to: https://notebook.brown.rcac.purdue.edu/

You should now be setup!

Please commit and push any changes made to methods in the tc_python_plotter.py file.

Funding Acknowledgements:

• NSF DMREF DMR2119619