Python Dependencies¶

To prevent Git merge conflicts, the Python dependencies are split up into 3 files (note the .in. instead of .txt file extension):

requirements.in: Your application-specific top-level dependencies. This is usually the only file you should edit.
requirements-core.in: Dependencies required by xlwings Server.
requirements-dev.in: Additional dependencies only required during development.

Here’s an example of requirements.in:

-r requirements-core.txt  # Don't delete this line
pandas
numpy==1.26.4

When you first clone xlwings Server, there will be a few dependencies in requirements.in. They are there to make the examples work out of the box, but you should replace them with your own top-level dependencies. The .txt version of the requirements files need to be compiled, as we’ll see next.

Prerequisities¶

You need the package manager uv, which you can install either via

pip install uv

or by following the uv docs.

Add/remove dependencies¶

Edit requirements.in (note .in, not .txt) to add your application-specific dependencies.
Create the .txt versions of the requirements files by running:
```
python run.py deps compile
```
The .txt versions are the lock files where all dependencies incl. sub-dependencies are pinned.
Update your local dev environment by running:
```
uv pip sync requirements-dev.txt
```
If you use Docker, you need to rebuild your container instead:
```
docker compose build
```
Commit and push the changed requirements files to Git.

Upgrade dependencies¶

Run the following command:

python run.py deps upgrade

Note, however, that this command upgrades all the packages, including core and dev dependencies.

Conda environments¶

If you’d like to use a Conda env because e.g., you have dependencies that are only available via Conda, do the following:

Create a new Conda env. Make sure to replace myenv with the desired name of your environment:
```
conda create -n myenv python=3.12 -y
```
Activate the Conda env:
```
conda activate myenv
```
Install the uv package manager:
```
pip install uv
```
Install requirements-dev.txt (for development) or requirements.txt otherwise:
```
uv pip sync requirements-dev.txt
```
Install conda dependencies (instead of blas, provide your own packages):
```
conda install blas
```
Export the environment into environment.yml (call it environment-dev.yml if this is a development environment):
```
conda env export > environment.yml
```

Commit environment.yml to Git. You can now use it to recreate this specific Conda environment whenever you need it via conda env create -f environment.yml.