Merge pull request #111 from graphistry/dev/additional-python-path

Add env var ADDITIONAL_PYTHON_PATH to load custom python packages

README.md

@@ -166,6 +166,7 @@ For non-minimal installs, if GPUs are present, `graph-app-kit` leverages GPU clo
 
 * [Caddy](https://caddyserver.com/): Reverse proxy for custom URLs, [automatic LetsEncrypt TLS certificates](http://letsencrypt.org/), multiple sites on the same domain, pluggable authentication (see [integrations guide](docs/extend.md))
 
+* Also you can install your [custom Python Packages](docs/additional-packages.md).
 
 ## Contribute
 
@@ -176,4 +177,4 @@ We welcome all sorts of help!
 * Connectors: Examples for common databases and how to get a lot of data out
 * Demos!
 
-See [develop.md](develop.md) for more contributor information
+See [DEVELOP.md](DEVELOP.md) for more contributor information

docs/additional-packages.md

@@ -0,0 +1,33 @@
+# Adding Custom Python Packages to graph-app-kit
+
+## Install Custom Python Packages Locally
+
+Install the desired custom Python packages on your localhost.  For example, if you want to include a different version of `PyGraphistry` and an additional dependency like `faker`, run the following commands in your terminal:
+
+```bash
+mamba create --yes -n myenv python=3.8
+pip install pip install graphistry[all]
+pip install faker
+```
+
+If the environment is airgapped, copy the packages into your local directory (e.g. `/home/user/additional-python-packages`).  Ensure that you use the same system architecture and Python version.
+
+## Mount the Volume in Docker Compose
+
+Update your Docker Compose file to mount the volume to the GAK container.  Specify the path to your local directory and the desired mount point inside the container in the environment variable `ADDITIONAL_PYTHON_PATH` (`/mnt` in this example):
+
+```yaml
+services:
+  streamlit:
+    ...
+    volumes:
+      ...
+      - /home/user/additional-python-packages:/mnt:ro
+    environment:
+        ADDITIONAL_PYTHON_PATH: "/mnt"
+    ...
+```
+
+If using conda, replace `/home/user/additional-python-packages` with the full path of `$CONDA_PREFIX/lib/python3.8/site-packages`.
+
+That's it! You've successfully added custom Python packages to the graph-app-kit container.  You can now use these packages in your GAK environment for graph analytics and visualizations.

src/docker/entrypoint.sh

@@ -20,7 +20,10 @@ echo "Log level (LOG_LEVEL): $LOG_LEVEL"
 echo "Graphistry user (GRAPHISTRY_USERNAME): $GRAPHISTRY_USERNAME"
 echo ""
 
-
+if [ -n "$ADDITIONAL_PYTHON_PATH" ]; then
+    export PYTHONPATH=$ADDITIONAL_PYTHON_PATH:$PYTHONPATH
+    echo "Using ADDITIONAL_PYTHON_PATH=$ADDITIONAL_PYTHON_PATH"
+fi
 
 mkdir -p /root/.streamlit