Pcpatch to gdal raster

goal

This project proposes python library to

• convert from pcpatch to numpy structured array (custom dtype)
• convert a numpy structured array to one or several gdal rasters

##architecture the mainarchitecture is :

• a class to store pc_schema and tools to parse xml to file it;
• a class to store the numpy structured array plus metadata associated to it
• function to read write pcpoint/pcpatch
• function to convert numpy structured array plus metadata to gdal rasters

there are several distinct utilities :

• reading and parsing a xml schema
  • pcpatch and pcpoint can be understood thanks to an xml pcschema
  • this xml schema is parsed and translated as a custom pcschema python class
• reading writing pcpoint / pcpatch in wkb format
  • uncompressed pcpatch can be read and write in wkb thanks to python struct module.
  • both read and write use numpy structured array as interface
• creating gdal rasters from numpy array
  • on pcpatch may with several attributes may be converted to several monoband rasters.
  • the user define the function that create the pixel values, so it can be the result of any computing between attributes

feature

• full schema parsing
• convert to numpy structured array
• cache for schema
• work inside and outside db (transparently, plpyhton or python)
• automatic src transmission
• automatic nodata creation
• user defined function to create rasters

##limitation

• a pixel value can only be determined from point with min_z in this pixel
• need a way to easily transfer attributes without modification
• need a way to propagate scale and offset

##todo

• vrt dataset from all output rasters
• scale and offset
• cleaning
• better doc (for users)