Skip to content

Conversion cockpit

timrdf edited this page Aug 5, 2011 · 27 revisions

Conversion cockpit is a metaphor that refers to the directories from which you should perform conversions.

What does a conversion cockpit look like?

If SSS, DDD, and VVV are the identifiers that you establish to name the source, dataset, and version of the data files that collected and converted, then the following directory structure is the dataset's conversion cockpit:

source/SSS/DDD/version/VVV

What is in a conversion cockpit?

Conversion cockpit directories contain the following directories:

  • source/
  • manual/
  • automatic/
  • publish/

They also contain the conversion trigger script:

  • convert-*.sh

When should I be in a conversion cockpit?

(always)

A conversion cockpit directory should be your working directory every time you do work with the data.

The bin/publish/*.sh scripts should be run from conversion cockpit directories.

Conversion cockpit directories are situated within the source/, dataset/, version/ hierarchy that corresponds to how datasets are named. For more information, see conversion process phase: name. Example conversion cockpit directories include:

/source/epa-gov/dataset/air-quality-system/version/2011-Jan-31/
/Users/timrdf/Desktop/source/epa-gov/dataset/air-quality-system/version/2011-Jan-31/

/source/census-gov/dataset/census-2011/version/2011-Jan-31/
/Users/timrdf/Desktop/source/census-gov/dataset/census-2011/version/2011-Jan-31/

The goofy /Users/timrdf/Desktop examples are used to demonstrate that the stuff before source/ does not matter. You get to put it where you want, and you can have multiples hanging around different places.

See also Conversion process phase: retrieve, conversion trigger.

Clone this wiki locally