## Proxy Web Apps

You can host many web applications by running them on the command line and then accessing them through the jupyter-proxy plugin.

Underneath the hood we are using the [Jupyterhub Proxy Plugin](https://jupyter-server-proxy.readthedocs.io/en/latest/arbitrary-ports-hosts.html) in order to access web applications through jupyterhub.

Before we start you'll need your:

1. Jupyterhub URL - copy and paste this from the browser
2. The service prefix, if any
3. Your username

```
env |grep JUPYTERHUB_SERVICE_PREFIX
env |grep JUPYTERHUB_BASE_URL
```

### Proxy URLs

The official [docs](https://jupyter-server-proxy.readthedocs.io/en/latest/arbitrary-ports-hosts.html). Many docs are embedded directly from the source docs.

### Starting and Proxying Processes

[Starting and Proxying Processes](https://jupyter-server-proxy.readthedocs.io/en/latest/server-process.html)

<iframe width="100%" height="200" scrolling="yes" frameborder="0" 
src="https://jupyter-server-proxy.readthedocs.io/en/latest/server-process.html">
</iframe>

### Accessing Arbitrary Ports

[Accessing Arbitrary Ports](https://jupyter-server-proxy.readthedocs.io/en/latest/arbitrary-ports-hosts.html)

<iframe width="100%" height="200" scrolling="yes" frameborder="0" 
src="https://jupyter-server-proxy.readthedocs.io/en/latest/arbitrary-ports-hosts.html">
</iframe>

### More Examples

Typically, you can grab the url from your browser.

For example, the URL for this browser is:

```
https://jhub1.bioanalyzedev.io/user/jillian/lab/tree/bioanalyze-hpc/docs/web-apps/proxy-web-apps.md
```

The base jupyterhub url is:

```
https://jhub1.bioanalyzedev.io/user/jillian/
```

The proxy path is:

```
https://jhub1.bioanalyzedev.io/user/jillian/proxy/${PORT}/
```

### Gotchas

You always want to make sure that you're serving your application from `0.0.0.0`, *not* localhost. Usually this option is available through the CLI to run the program either as `--host` or `--ip`.

### Example - ExCellxGene

Let's proxy the ExCellxGene in Jupyterhub.

Excellxgene needs to be installed to it's own conda environment because it doesn't play nice with the existing conda env.

* Grab the `excellxgene-environment.yaml` from this repo.
* Create the excellxgene environment
* Register the excellxgene environment as a Python Kernel.


```bash
git clone https://github.com/dabble-of-devops-bioanalyze/single-cell-saturn-cloud
git clone https://github.com/scverse/scanpy-tutorials

cd single-cell-saturn-cloud
mamba env create -f ./excellxgene-environment.yaml -p ~/.conda/envs/excellxgene
source activate conda activate excellxgene
```

If you are running on a local system or [BioAnalyze HPC](https://hpc.bioanalyze.io) I recommend using [these](https://hpc.bioanalyze.io/software/install-software.html) settings.

Now that we have some data let's visualize some clusters!


```bash
git clone https://github.com/czbiohub/excellxgene
cd excellxgene
cd example-dataset
wget https://raw.githubusercontent.com/dabble-of-devops-bioanalyze/single-cell-saturn-cloud/main/prepare.py
source activate excellxgene
python ./prepare.py pbmc3k.h5ad --output ./pbmc3k-preprocessed.h5ad --overwrite
```

### Launch the ExCellxGene Web App

```bash
excellxgene launch --host 0.0.0.0 --port 8000 ./pbmc3k-preprocessed.h5ad
```

The proxy path is:

```
https://{JHUB_URL}/user/${USERNAME}/proxy/${PORT}/
```