UV

Using UV#

The easiest method for developing PBM is to use the tool uv which combines handling of python dependencies, virtual environment, deployment and versioning.

Installing PBM with uv#

uv can be installed using pip:

pip install --user uv

you can now install the development version of PBM by running:

uv venv

within the repository directory. This will create a virtual environment containing all Python modules required to run the project as defined within the pyproject.toml file.

Any changes made within the repository will be picked up by uv whenever the code is run.

Running PBM with uv#

Run the powerbalance command from within the virtual environment by executing:

uv run powerbalance

Troubleshooting#

If you encounter issues with numpy as a dependency, this can usually fixed by installing it manually beforehand:
```
uv pip install numpy
```
Another problem encountered on Windows systems is with the installation of PyTables, the error usually states that HDF5 libraries could not be located and is due to there being no built wheels in the PyPi database for the user's Python version and system architecture (when this is the case Python attempts to build the module itself). A working solution is to download the relevant pytables wheels for Windows from here, then install them with pip:
```
uv pip install <path-to-wheels-file>
```
If you are using uv and the command uv is not recognised after it has been installed, make sure the location of your uv installation is added to PATH, to find the installation try running
```
python -m uv run which uv
```
and noting the address given during the initialisation. For example for the case of a prefix of C:\Users\<user>\AppData\Local\, uv was found in C:\Users\<user>\AppData\Local\Packages\PythonSoftwareFoundation.<python_version_string>\LocalCache\local-packages\Python<version-num>\Scripts. This location is then added to the PATH variable as described here or temporarily

In the case of Windows:
```
set "PATH=%PATH%;<location-of-uv.exe>"
```
for Linux/macOS:
```
export PATH=<location-of-uv>:$PATH
```
If the command powerbalance is not available after install, you can also run the program via Python as normal:
```
uv run python <path-to-cloned-repo>/power_balance/cli/__init__.py
```