Docker.md

ScienceBeam Docker Containers

This section details using or building the docker images. You will need Docker and Docker Compose.

Run Latest Docker Container

Make sure you have the latest docker image by running:

docker pull elifesciences/sciencebeam-pipelines

Note: add the develop tag to use the local image after building it (docker-compose build)

Clone this repository and run:

docker-compose -f docker-compose.latest.yml up

That will start GROBID and ScienceBeam docker containers. The ScienceBeam API will be available on port 8075.

Alternatively run (the --add-host parameter is used to prevent Crossref lookups):

docker run --rm --add-host api.crossref.org:127.0.0.1 -p 8070:8070 lfoppiano/grobid:0.5.3

and:

docker run --rm -i -t -p 8075:8075 --net=host elifesciences/sciencebeam-pipelines \
  ./server.sh --host=0.0.0.0 --port=8075 --grobid-url http://localhost:8070/api

Run Science Parse and ScienceBeam Docker Container

Run the Science Parse and ScienceBeam docker container:

docker run --rm -p 8071:8080 allenai-docker-public-docker.bintray.io/s2/scienceparse:1.3.2

and:

docker run --rm -i -t -p 8075:8075 --net=host elifesciences/sciencebeam-pipelines \
  ./server.sh --host=0.0.0.0 --port=8075 --pipeline=scienceparse --science-parse-url \
  http://localhost:8071/v1

Run Science Parse V2 and ScienceBeam Docker Container

Run the Science Parse V2 and ScienceBeam docker container:

docker run --rm -p 8073:8081 allenai-docker-public-docker.bintray.io/s2/spv2:2.10

and:

docker run --rm -i -t -p 8075:8075 --net=host elifesciences/sciencebeam-pipelines:develop \
  ./server.sh --host=0.0.0.0 --port=8075 --pipeline=scienceparse \
  --science-parse-url http://localhost:8073/v1/json/pdf --no-science-parse-xslt

Run CERMINE and ScienceBeam Docker Container

Run the CERMINE and ScienceBeam docker container:

docker run --rm -p 8072:8080 elifesciences/cermine:1.13

and:

docker run --rm -i -t -p 8075:8075 --net=host elifesciences/sciencebeam-pipelines \
  ./server.sh --host=0.0.0.0 --port=8075 --pipeline=cermine --cermine-url \
  http://localhost:8072/extract.do

Run meTypeset and ScienceBeam Docker Container

Run the meTypeset and ScienceBeam docker container:

docker run --rm -i -t -p 8074:8080 elifesciences/metypeset:latest

and:

docker run --rm -i -t -p 8075:8075 --net=host elifesciences/sciencebeam-pipelines \
  ./server.sh --host=0.0.0.0 --port=8075 --pipeline=metypeset --metypeset-url \
  http://localhost:8074/api/convert

Run ContentMine and ScienceBeam Docker Container

Run the contentmine (for ContentMine's svg2xml) and ScienceBeam docker container:

docker run --rm -i -t -p 8076:8080 elifesciences/contentmine

and:

docker run --rm -i -t -p 8075:8075 --net=host elifesciences/sciencebeam-pipelines \
  ./server.sh --host=0.0.0.0 --port=8075 --pipeline=contentmine --contentmine-url \
  http://localhost:8076/api/convert

Build and Run Docker Container with Docker Compose

docker-compose -f docker-compose.yml build

docker-compose -f docker-compose.yml up

That will start GROBID and ScienceBeam docker containers. The ScienceBeam API will be available on port 8075.

Build Docker Container Only

docker-compose build

GROBID and Crossref Consolidation

GROBID is optionally using Crossref consolidation to make the results more useful for archiving.

For this project, the preference is not to enable it and the docker-compose.yml* files explictly use a local network to prevent that.