Getting started

Installation

Recommended Configuration: Tested on Ubuntu 22.04 with Python 3.12.

1. Clone the repository from the main branch:

$ git clone https://github.com/jpcurbelo/torchHydroNodes.git

2. Prerequisites and Setup (Tested with Python 3.12.13)

2.1. Set Up the Virtual Environment

• Navigate to the torchHydroNodes directory:

  $ cd torchHydroNodes
  

• Create the virtual environment:

  $ python3.12 -m venv venv-thn
  

• Activate the virtual environment:

  $ source venv-thn/bin/activate
  

  2.2. Install Package Dependencies

  • Install the required packages using pip:

    (venv-thn) $ pip install .
    

    This command will handle the installation of all dependencies specified in the pyproject.toml file.

  • If you encounter any issues, you can manually install the dependencies listed in the requirements.txt file using:

    (venv-thn) $ pip install -r requirements.txt
    

3. Data Preparation

• Download and place your dataset in a convenient directory.

  Note: For the CAMELS US dataset, we recommend referring to the entries on meteorological time series, streamflow data, and catchment attributes in the Tutorial on Data Prerequisites within the NeuralHydrology documentation.

• Define or update the corresponding data_dir entry in the src/utils/data_dir.yml file.

  Note: In the default configurations of our tutorials, the key data_dir is defined as:

  data_dir_camelsus: /gladwell/hydrology/SUMMA/summa-ml-models/CAMELS_US
  

  This path may vary depending on your setup.

• Ensure that the shapefiles (.shp) required for plotting are specified in the data_dir.yml file. These files are dataset-specific and are used for geographic visualization of catchment boundaries or other spatial attributes.

  For example:

  # Relative path after the data_dir
  map_shape_file: basin_set_full_res/usa-states-census-2014.shp
  hm_catchment_file: basin_set_full_res/HCDN_nhru_final_671.shp