Fruit Fly Brain Hackathon 2019

Wednesday, March 20, 2019
750 CEPSR
Center for Neural Engineering and Computation
Columbia University, New York, NY, USA

Overview

The 4th Fruit Fly brain Hackathon (FFBH 2019) will be held on Wednesday, March 20, 2019. The goal of the hackathon is to bring together researchers interested in developing executable models of the fruit fly brain. In this year’s hackathon, we will focus on the FlyBrainLab (FBL), a newly developed interactive computing platform for studying the function of executable circuits constructed from fly brain data. The hackathon is aimed at three main groups of participants: neurobiologists, modelers and software engineers. We welcome researchers working on the fruit fly brain as well as those working on other model organisms to participate and broaden the discussion in the hackathon.

The Fruit Fly Brain Hackathon 2019 is organized in conjunction with the Columbia Workshop on Brain Circuit, Memory and Computation on March 21-22, 2019. Participants of the hackathon are welcome to attend the workshop.

Organizers

Tingkai Liu, Department of Electrical Engineering, Columbia University

Mehmet Kerem Turkcan, Department of Electrical Engineering, Columbia University

Chung-Heng Yeh, Department of Electrical Engineering, Columbia University

Yiyin Zhou, Department of Electrical Engineering, Columbia University

Registration

Registration is free but all participants have to register. To help us better organize the event, please provide in the appropriate registration block a brief description of your background and what you would like to learn/achieve during the hackathon. Thank you!

Lodging and Directions to Venue

Please follow this link for lodging details and directions to the hotel and venue.

Schedule

Wednesday, March 20th, 2019

Time	Topic
09:00 AM - 09:10 AM	Introduction to the Fruit Fly Brain Hackathon (Chung-Heng Yeh, Columbia University, USA).
09:10 AM - 09:15 AM	Self-introductions by Participants.
09:15 AM - 09:30 AM	Introduction to FlyBrainLab (Mehmet K. Turkcan, Columbia University, USA).
09:30 AM - 12:30 PM	Hacking: Basic Features of FBL.
12:30 PM - 01:30 PM	Lunch Break.
01:30 PM - 01:45 AM	Building Models of Fly Brain Neuropil Using FBL (Yiyin Zhou, Columbia University, USA).
01:45 AM - 02:00 PM	Discussions and Summary of Project Ideas.
02:00 PM - 06:00 PM	Hacking.

Computing Resources

Participants only need to bring a laptop to the hackathon to fully explore the capacity of the FFBO. We will host the main FFBO service on an Amazon EC2 instance with limited capacity. For attendees interested in installing FFBO on their own machines, or developing new features, all service packages will be available in the form of Docker images.

Docker Image

Docker image for a complete install of FFBO and FBL is available HERE. You can simply pull the image onto your system:

docker pull fruitflybrain/ffbh19

Running this docker image with GPU support requires nvidia-docker (Installation Instruction). The image has been tested on Ubuntu 16.04 with nvidia-docker v2.0. An Amazon Machine Image (ami-0221bb3d675af4924 in the US-East-1 region) with Docker and nvidia-docker installed is provided for convience.

To launch a Docker container using the image:

docker run -d --runtime=nvidia --name flybrainlab -p 10000-10002:10000-10002 -p 10003:22 -it fruitflybrain/ffbh19 bash -c "service ssh start; /opt/orientdb/bin/server.sh"

The container will be launched in the background. You can now SSH into it with

ssh -p 10003 ffbh@server-ip

and the password is Drosophila. For security reason, please change the password after your first login by

passwd

Launching FFBO Servers

To launch all FFBO servers, simply run the following shell script in /home/ffbh/run_scripts in the following order (we recommand running each script using a tmux session/window):

sh run_processor.sh
sh run_nlp.sh
sh run_neuroarch.sh
sh run_neurokernel.sh

Please go to http://server-ip:10000 and test a query before using FBL.

Launching FBL

To launch FBL, you can run the shell script /home/ffbh/run_script/run_fbl.sh. The jupyter lab will be launched at port 10004 by default. Since the port is not exposed by the Docker container, you can do one of the following:

1. (preferred) Use a SSH tunnel, i.e,:

   ssh -p 10003 -L local_port:localhost:10004 ffbh@server-ip

   Replace local_port by your choice of port number. You can access the FBL using your browser at http://localhost:local_port.

2. Expose the 10004 port when starting the Docker container, i.e,:

   docker run -d --runtime=nvidia --name flybrainlab -p 10000-10002:10000-10002 -p 10003:22 -p 10004:10004 -it fruitflybrain/ffbh19 bash -c "service ssh start; /opt/orientdb/bin/server.sh"

   and access FBL using your browser at http://server-ip:10004.

Using Supplied Server

You can install the FBL on your laptop using the instruction here: https://github.com/FlyBrainLab/flybrainlab

Please note that, for FBLClient, you need to use the branch called ffbh19.

After installation, you need to the following configuration before starting jupyter lab.

The IP address of our server is 54.243.17.111. Please put this into the field ip in FBLClient.ini in the FBLClient repository and change the [ID][digits] field to 0.

Finally, launch jupyter lab.