Skip to content

Docker

Dockerfile 
FROM python

WORKDIR /home

RUN pip install kmviz gunicorn
ENV KMVIZ_CONF=/home/config.toml

ENTRYPOINT ["python", "-m", "kmviz", "app", "deploy"]
CMD ["-u 0.0.0.0", "-p 8000", "-w 4"]
Start the service using docker
docker pull tlemane/kmviz:latest
docker run -d --network="host" -v.:/home/ tlemane/kmviz -w 1 -u 0.0.0.0 -p 8000

  • Docker Options

    • --network="host": Use the host network
    • -p .:/home/: Mount . at /home/
    • -d: Detach the container

  • Server Options (See kmviz app deploy)

    • -w: The number of workers
    • -u: Host
    • -p: Port

By default, KMVIZ_CONF is set to /home/config.toml. To use another filename, use docker run --env KMVIZ_CONF=/home/my_config.yaml ....

About cache

3 caching systems are used in kmviz:

serverside and result configuration

Available backends

Disk
[cache.result]
type = "disk"
params.cache_dir = "result_cache"
params.threshold = 0
params.default_timeout = 1209600
Redis
[cache.result]
type = "redis"
params.host = "localhost"
params.port = 6379
params.db = 1
params.default_timeout = 86400

Note

redis is recommended for both serverside and result when deploying multi-user instances.

About timeout

  • result: The default_timeout field corresponds the time during which a user can access a result without recomputing the query.
  • serverside: The cache should always return and value, i.e. keys should not expire during a user session. Use a significant value, e.g. 24 hours.

manager configuration

Available backends

Disk
[cache.manager]
type = "disk"
params.directory = ".results/kmviz_manager_cache"

Complete example

config.toml
[cache]

[cache.serverside]
type = "redis"
[cache.serverside.params]
host = "localhost"
port = 6379
db = 0
default_timeout = 86400

[cache.manager]
type = "disk"
params.directory = "./kmviz_manager_cache"

[cache.result]
type = "disk"
[cache.result.params]
cache_dir = "./kmviz_result_cache"
threshold = 0
default_timeout = 1209600