Pipeline is a library + CLI tool to load programmable pipelines, execute jobs and get detailed results per stage for reporting/debugging.
- Pipeline: A sequential list of stages
- Stage: An command template with a unique name within the pipeline.
- Job: A job that will be executed on a pipeline with an optional set of input variables
- JobResult: The output of a Job that has been executed on a pipeline, contains status and a StageResult for each stage in the pipeline
- StageResult: The exact command that was executed, it's exitcode, stdout and stderr.
Load a pipeline YML file and execute it with input variable topic=technology
bin/pipeline run examples/bbc-news-demo.pipeline.yml -d topic=technology
Load a pipeline YML file and execute it with input from STDIN
bin/pipeline run examples/bbc-news-stdin-demo.pipeline.yml < examples/bbc-news-demo.rss.xml
Output a JSON result of the JobResult + StageResults to result.json
bin/pipeline run -o result.json examples/bbc-news-demo.pipeline.yml -d topic=technology
Load a pipeline YML file and intentially cause an error by requesting a non-existant topic. This will output the errors.
bin/pipeline run examples/bbc-news-demo.pipeline.yml -d topic=this-topic-does-not-exist
You can suppress output and debugging output by using --quiet
. You can use this in combination with -o to read the details from a .json file.
In all cases the exit code of the pipeline command will match the exit code of the last stage result: 0 on success, or other on failure.
Please refer to examples/example-code.php
on how to use the pipeline package as a library.
pipeline will use environment variables in the inputs of pipelines. This way you don't need to define all required variables using -d
.
When pipeline runs, it first checks if a .env
file exists in the current directory and update the environment variabels based on it's contents.
MIT. Please refer to the license file for details.
Check out our other projects at linkorb.com/engineering.
Btw, we're hiring!