openrefine-client/docker/README.md

84 lines
3.4 KiB
Markdown

## batch processing with python-client
There are some client libraries for OpenRefine that communicate with the [OpenRefine API](https://github.com/OpenRefine/OpenRefine/wiki/OpenRefine-API). I have prepared a docker container on top of the [Python Library from PaulMakepeace](https://github.com/PaulMakepeace/refine-client-py/) and extended the CLI with some options to create new OpenRefine projects from files.
### basic usage
**1) start server:**
> docker run -d --name=openrefine-server felixlohmeier/openrefine
**2) run client with one of the following commands:**
list projects:
> docker run --rm --link openrefine-server felixlohmeier/openrefine-client --list
create project from file:
> docker run --rm --link openrefine-server felixlohmeier/openrefine-client --create [FILE] [PROJECTID]
apply rules from json file:
> docker run --rm --link openrefine-server felixlohmeier/openrefine-client --apply [FILE.json] [PROJECTID]
export project to file:
> docker run --rm --link openrefine-server felixlohmeier/openrefine-client --export [PROJECTID] --output=FILE.tsv
check help screen for more options:
> docker run --rm --link openrefine-server felixlohmeier/openrefine-client --help
**3) cleanup:**
> docker stop openrefine-server && docker rm openrefine-server
### example for customized run commands in interactive mode (e.g. for usage in terminals)
**1) start server in terminal A:**
```docker run --rm --name=openrefine-server -p 80:3333 -v /home/felix/refine:/data:z felixlohmeier/openrefine -i 0.0.0.0 -m 4G -d /data```
* automatically remove docker container when it exits
* set name "openrefine" for docker container
* publish internal port 3333 to host port 80
* mount host directory /home/felix/refine as working directory
* make openrefine available in the network
* increase java heap size to 4 GB
* set refine workspace to /data
* OpenRefine should be available at http://localhost
**2) start client in terminal B (prints help screen):**
```docker run --rm --link openrefine-server -v /home/felix/refine:/data:z felixlohmeier/openrefine-client```
* automatically remove docker container when it exits
* build up network connection with docker container "openrefine"
* mount host directory /home/felix/refine as working directory
* apply history in file /home/felix/refine/history.json to project with id 1234567890123
### example for customized run commands in detached mode (e.g. for usage in shell scripts)
**1) define variables (bring your own example data)**
> workingdir=/home/felix/refine
> inputfile=example.csv
> jsonfile=test.json
**2) start server**
```docker run -d --name=openrefine-server -v ${workingdir}:/data:z felixlohmeier/openrefine -i 0.0.0.0 -m 4G -d /data```
**3) create project (import file)**
```docker run --rm --link openrefine-server -v ${workingdir}:/data:z felixlohmeier/openrefine-client --create $inputfile```
**4) get project id**
```project=($(docker run --rm --link openrefine-server -v ${workingdir}:/data felixlohmeier/openrefine-client --list | cut -c 2-14))```
**5) apply transformations from json file**
```docker run --rm --link openrefine-server -v ${workingdir}:/data felixlohmeier/openrefine-client --apply ${jsonfile} ${project}```
**6) export project to file**
```docker run --rm --link openrefine-server -v ${workingdir}:/data felixlohmeier/openrefine-client --export --output=${project}.tsv ${project}```
**7) cleanup**
```docker stop -t=500 openrefine-server && docker rm openrefine-server```