MapofLife / MOL

Integrating information about species distributions in an effort to support global understanding of the world's biodiversity.
http://mol.org
BSD 3-Clause "New" or "Revised" License
25 stars 2 forks source link

Develop an analysis pipeline #5

Closed gaurav closed 12 years ago

gaurav commented 12 years ago

The goal for this issue is to construct an internal analysis pipeline so that calculated raster layers can be provided to the user. This involves (copied from @DanRosauer's notes from our meeting last weekend - see the "Map of Life - Meeting Notes" Google document for further information):

This is a high-priority task, since it is necessary to make it possible for the Jetz lab to work directly on Map of Life.

Future goals include replacing the sparse matrix with (1) a precomputed intersection of MoL polygons, and later (2) calculating it regularly as new data is added to MoL.

gaurav commented 12 years ago

Further information about this task:

  1. We do not have to worry about ingesting rasters at this point, only in allowing rasters to be generated by analysis programs internally and stored in the database.
  2. Quad id are unique to a grid, so cell 1201 in grid 360x114 will always point to the same lat/long coordinates. It might be easier to map quad ids to lat/long and render them using information about the grid, or to have a table mapping quad ids to polygons directly, which can then just be rendered by Windshaft. That's something we'll know more about as the polygon-ingest/Windshaft code comes together this week.
  3. @DanRosauer's suggestion is to have a table mapping values/results from an analysis to a 'theme' (the analysis class being executed) and 'analysis' (the specific instance of an analysis with the parameters used to execute the analysis).
gaurav commented 12 years ago

Is this going to be made redundant by our work on Google's Earth Engine?