cloudlab-cmu / cloudcompchem Goto Github PK

Follow the example of SinglePointEnergyResponse.

Add controller code and extend CLI for hessian calculations.

Validate the input to ensure correctness, reject the request properly if an invalid state is detected.

Handle any runtime errors in a graceful way*

• • graceful meaning, every possible runtime exception is caught and sent back to client with useful error code and message

Develop geometry opt input JSON payload (look at energy payload for reference)

Build out scaffolding for handling Hessian calculation response. For the request payload, you may re-use the same structure as the EnergyRequest. The response payload must include a value for a hessian which should be a square matrix.

Ask about using shared resources to route requests from CMU students to HPC systems

Goal is to start using pydantic models for automated type checking, validation, and (un)marshaling of json strings into structs. FastAPI supports this out of the box, while Flask doesn't, so we're going to switch now before the code gets too complicated.

We would like to enable async runtime for the webserver of cloudcompchem. This does not need to run locally; the requirement is that users may want to run a lot of a couple long running calculations that they don't want to wait around for.

Build the calculator for a geometry optimization in pyscf

Create the tests

Add initial scaffolding for interact with a deployed service, thus making it very simple for users to run calculations using cloud resources.

add orbital energies in addition to new changes to app structure

Need to define request (e.g. VibrationalEnergyRequest) and response (e.g. VibrationalEnergyResponse) structures that allow users to calculate vibrational frequencies and eigenvectors/modes. Request packet needs to contain enough information to run a vibrational calculation for a given molecule. The response packet needs to contain, at minimum, the vibrational energies and mode vectors.

Write the code required to serve vibrational calculation requests. Include all relevant tests to ensure proper handling of all edge cases

As soon as cloudcompchem.emeraldcloudlab.com is deployed, change the url key in the Client object to point to it.

Create endpoint for calculating the density matrix for a given molecule using the molecular orbital coefficients in a SCF calculation.

Quick validation function that checks the combination of charge and spin multiplicity for all DFT inputs.

See this for a minimal example of how to implement the logic.

Switch the energy endpoint from using dataclasses to pydantic base models

Write any server and controller code that handles incoming and outgoing JSON requests/responses. Include all relevant tests in the test_server.py file (or feel free to split into another file if necessary).

Think of a fancy branding name for our project!

Something like:
cloudcompchem.emeraldcloudlab.com
compchem.emeraldcloudlab.com
science.emeraldcloudlab.com