Skip to content

Node and Python FastAPI based RESTful APIs. The APIs are used by (1) RAMPVIS user interface: https://github.com/ScottishCovidResponse/rampvis-ui and (2) RAMPVIS Infrastructure management user interface: https://github.com/saifulkhan/rampvis-ontology-management-ui

License

Notifications You must be signed in to change notification settings

ScottishCovidResponse/rampvis-api

Repository files navigation

About

This project implements RESTful APIs for the RAMPVIS system. This repository consist of the following top-level folders.

  1. data-api

    • Implemented in Python, FastAPI, and other Python libraries.
    • APIs for all data.
    • API implements processing functions, e.g., analytical algorithms, propagation, scheduler agents, etc.
  2. infrastructure-api

    • Implemented in Typescript, Node.js, Express.js, and other JavaScript libraries.
    • APIs for ontology and database operations; authentication and user management; other infrastructure related services.
    • Thumbnail and search index services.

Getting Started

Prerequisites

  • This is tested in Ubuntu 22.04 and WSL.

  • Ensure docker is running.

  • We created multiple docker-compose scripts and each script handles a set of services. Sequentially start the services. Note that based on what version is installed either use docker compose or docker-compose.

  • Update docker virtual memory size. Sometimes, the Elasticsearch services in docker crash due to a known issue: max virtual memory areas vm.max_map_count [65530] is too low, increase to at least [262144]. Follow the instructions below to resolve this issue,

In WSL, run the following commnads in PowerShell. More information can be found here.

wsl -d docker-desktop cat /proc/sys/vm/max_map_count # current value
# 65530
wsl -d docker-desktop sysctl -w vm.max_map_count=262144
# vm.max_map_count = 262144

In Ubuntu, run the following command. More information can be found here.

sudo sysctl -w vm.max_map_count=262144

Stop & Clean (optional)

Stop & clean rampvis-ui and rampvis-ontology-management-ui

Following commands will stop the containers and and clean the images related to this repository.

# stop containers
docker-compose -f docker-compose-ext.yml stop
docker-compose -f docker-compose-int.yml stop
docker-compose -f docker-compose-seed.yml stop

# remove containers
docker-compose -f docker-compose-ext.yml rm
docker-compose -f docker-compose-int.yml rm
docker-compose -f docker-compose-seed.yml rm

# remove images
docker rmi mongo-setup seed-data rampvis-api_data-api rampvis-api_infrastructure-api

# remove volumes, e.g., database and search indices
docker volume rm rampvis-api_mongostatus rampvis-api_esdata01 rampvis-api_esdata02 rampvis-api_esdata03 rampvis-api_mongodata01 rampvis-api_mongodata02 rampvis-api_mongodata03

# remove all unused volumes
docker volume prune

# leave swarm network
docker swarm leave --force

# remove all unused networks
docker network prune

Start Development Instance

Create an overlay network to allow communication between docker applications:

docker swarm init
docker network create --driver overlay --attachable rampvis-api-network
# inspect the network
docker network inspect rampvis-api-network

[1] Start External Services

The infrastructure APIs are dependent on database and search engine- MongoDB, Elasticsearch, and Kibana.

# start the services
docker-compose -f docker-compose-ext.yml up -d

Wait for sometime to let all the services to start. Check if the Elasticsearch (es01, es02, es03), Kibana (kib01), and MongoDB (mongodb01, mongodb02, mongodb03) services are running properly.

# check the status
docker-compose -f docker-compose-ext.yml ps

# the output should look like
#  Name                  Command               State                 Ports
# ----------------------------------------------------------------------------------------
# es01          /bin/tini -- /usr/local/bi ...   Up       0.0.0.0:9200->9200/tcp, 9300/tcp
# es02          /bin/tini -- /usr/local/bi ...   Up       9200/tcp, 9300/tcp
# es03          /bin/tini -- /usr/local/bi ...   Up       9200/tcp, 9300/tcp
# kib01         /bin/tini -- /usr/local/bi ...   Up       0.0.0.0:5601->5601/tcp
# mongo-setup   docker-entrypoint.sh bash  ...   Exit 0
# mongodb01     /usr/bin/mongod --replSet  ...   Up       0.0.0.0:27017->27017/tcp
# mongodb02     /usr/bin/mongod --replSet  ...   Up       27017/tcp
# mongodb03     /usr/bin/mongod --replSet  ...   Up       27017/tcp

Run the following command to inspect any container log.

docker-compose -f docker-compose-ext.yml logs <container_name>

# for example, inspect es01
docker-compose -f docker-compose-ext.yml logs es01

[2] Start Internal Services

Start the data-api and infrastructure-api

# start the services
docker-compose -f docker-compose-int.yml up -d
# check the status
docker-compose -f docker-compose-int.yml ps

# the output should look like
#        Name                     Command               State           Ports
# ------------------------------------------------------------------------------------
# data-api             uvicorn app.main:app --rel ...   Up      0.0.0.0:4010->4010/tcp
# infrastructure-api   docker-entrypoint.sh yarn dev    Up      0.0.0.0:4000->4000/tcp

To display the list of the APIs and test test if the service internal services/APIs are started properly, open the following URLs:

Note: Not all the infrastructure APIs are documented and this UI will only show the few APIs that are documented.

Note In order to start and debug the services locally see the data-api README and infrastructure-api README files.

[3] Inject the Seed Data

This script will clear the databases and search indices, and seed the MongoDB data from the rampvis folder. Do not run this if there is already available data that you do not want to remove.

# start the services
docker-compose -f docker-compose-seed.yml up -d
# check the status
docker-compose -f docker-compose-seed.yml ps

Injecting data and creating index may take some time, to inspect the log, run:

docker-compose -f docker-compose-seed.yml logs seed-data

To inspect if the index are created properly, run:

# check the search index
curl localhost:9200/_cat/indices

# the following index must be present in the oupuput
# green open rampvis.onto_vis                9d4VfZyyRCursKTYtVP-vw 1 1    27    0  65.1kb  32.5kb
# green open rampvis.onto_page               nwrlo8-eR3OOSITUho4zRQ 1 1 11125    0     4mb     2mb
# green open rampvis.onto_data               beS7p-fWTmWan-FA_ndhVw 1 1 18635    0   5.1mb   2.5mb

Containers and Network

This figure illustrates the containers and their swarm network connections.

fishy

The rampvis-ui and rampvis-ontology-management-ui are in a different repository.

BibTex

@article{Khan:2022:TSC,
  author        = {Khan, Saiful and Nguyen, Phong H. and Abdul-Rahman, Alfie and Freeman, Euan and Turkay, Cagatay and Chen, Min},
  title         = {Rapid Development of a Data Visualization Service in an Emergency Response},
  journal       = {IEEE Transactions on Services Computing},
  volume        = {15},
  pages         = {1251-1264},
  year          = {2022},
  doi           = {10.1109/TSC.2022.3164146}
}
@article{Khan2022:IEEE-TVCG,,
   author = {Saiful Khan, Phong Nguyen, Alfie Abdul-Rahman, Benjamin Bach, Min Chen, Euan Freeman, and Cagatay Turkay},
   title = {Propagating Visual Designs to Numerous Plots and Dashboards},
   journal = {IEEE Transactions on Visualization and Computer Graphics},
   pages = {86-95},
   volume = {28},
   year = {2022},
   doi = {10.1109/TVCG.2021.3114828},
   arxiv = {https://arxiv.org/abs/2107.08882}
}

Contact

URL: https://sites.google.com/view/rampvis/teams

Email: khan.saiful@outlook.com