STGraph
STGraph copied to clipboard
π Vertex Centric approach for building GNN/TGNNs
I'm trying to build and install STGraph with some local changes. I'm following the README.md, with the command ### python3 -m build && pip uninstall stgraph -y && pip install...
Shall revamp the stgraph.benchmark_tools so that it contains more utility tools the developer can use while training their models. Shall rename it to stgraph.tools module.
Not all of our dataloaders have the functionality to return the node features - which is very important for training a model on that dataset. So we should decide to...
As of now the graph abstraction in STGraph has the following **_"base"_** classes 1. `STGraphBase` 2. `StaticGraph` 3. `DynamicGraph` `StaticGraph`and `DynamicGraph` inherits from `STGraphBase`. Whereas `NaiveGraph`, `PCSRGraph` and `GPMAGraph` inherits...
Need to add script tests for EnglandCovid dataset. Add it to the _dynamic_tgcn_dataloaders_ testpack for version _1_1_0_. This testpack is not created yet.
Add all the docstrings and documentation for the core stgraph.compiler sub-module
Add all the docstrings and linting fixes for the classes present inside the `stgraph.graph` submodule.
The initial page is added, but it lacks depth. Enhancements are needed to highlight STGraph's strengths. We can take inspiration from our repo README and other library documentation homepages
A comprehensive tutorial would be helpful for first time users of STGraph. Should be present inside our Read The Docs documentation page.