DeepCell Kiosk: Deploying a Scalable Solution for Biological Image Analysis¶
This DeepCell distribution is designed to easily spin up an end-to-end DeepCell environment on Kubernetes. The repository contains shell scripts and configuration files that follow the infrastructure-as-code model to easily manage and scale a deep learning solution for biological image analysis.
Once running, the Kiosk iniates a web portal that has access to cluster of computers in the cloud. This distribution is managed automatically by Kubernetes, deployment software also launched by the Kiosk at run-time. The web portal allows multiple users to upload mulitple images (in the form of zip files) and download the resulting segmentation and classification of these images automatically.
Before starting the Kiosk¶
The scalability of the software is enabled by cloud computing. As such, the preparation steps vary depending on which cloud provider you plan on deploying with:
- Create an account at Google Cloud.
- Create a Google Cloud project.
- Make sure you have at least one account with the
- Make sure you have a storage bucket with public access.
- Make sure that any files you upload to the bucket (such as pre-trained Tensorflow models) have public access. (Google cloud uses a permission system that may be unfamiliar to Linux users. File permissions are not inherited from buckets; they are set individually on each file.)
- Once the Kiosk has started, select the configuration option for your chosen cloud provider, either Amazon or Google, and fill out the configuration values as needed. ( If using Google, follow the link provided during the configuration process in a web browser.) Once the Kiosk has been configured for a cloud provider, the word
(active)will appear next to that cloud provider’s configuration option in the Kiosk menu.
- With the Kiosk configured for the appropriate cloud provider, select the
Createoption from the Kiosk’s main menu to create the cluster on the chosen cloud provider. This may take up to 10 minutes. Cluster creation is done when you see
Cluster Createdfollowed by
---COMPLETE---printed to the terminal. If you see
---COMPLETE---with some error text immediately preceding it, cluster creation failed.
- Find the cluster’s web address by choosing the
Viewoption form the Kiosk’s main menu. (Depending on your chosne cloud provider and the cloud provider’s settings, your cluster’s address might be either a raw IP address, e.g., “123.456.789.012”, or a URL, e.g., “deepcellkiosk.cloudprovider.com”.
- Go to the cluster address in your web browser to find the Deepcell frontpage.
When you’re done using the cluster, you may want to shutdown the cluster, and perhaps the kiosk itself.
- To shutdown the cluster, select
Destroyfrom the Kiosk’s main menu. Cluster destruction is complete when you see
Cluster destroyedfollowed by
---COMPLETE---printed to the screen. However, if the screen shows error output immediately before the
---COMPLETE---indicator, cluster destruction failed and some components may remain. In this case, it may be best to delete resources manually though the cloud provider’s web interface.
- After shutting down the cluster, if you wish to shut down the kiosk, simply select
Exitform the Kiosk’s main menu. The Kiosk is now completely shut down.
- When using the
Predictfunctionality, the first image will take a while to process (up to 10 minutes) because the cluster will need to requisition more computing resources. (This is because the cluster is designed to use as few resources as possible in its resting state.)
- This repository is being actively developed. If you are experiencing issues with the Deepcell kiosk, please consult the Troubleshooting document in the
- Those interested in Kiosk developement should follow a different path to start the Kiosk:
- Clone this repo:
git clone firstname.lastname@example.org:vanvalenlab/kiosk.git
- Initialize the “build-harness”:
- Build the container:
- Install wrapper script:
- Start the kiosk.
- Clone this repo:
This software is licensed under a modified APACHE2.
See LICENSE for full details.
Licensed to the Apache Software Foundation (ASF) under one or more contributor license agreements. See the NOTICE file distributed with this work for additional information regarding copyright ownership. The ASF licenses this file to you under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with the License. You may obtain a copy of the License at https://www.apache.org/licenses/LICENSE-2.0 Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the specific language governing permissions and limitations under the License.
All other trademarks referenced herein are the property of their respective owners.