source: TI02-CSML/trunk/parser/parser_extra.py @ 1147

#parser_extra.py
# DL 12 april 2006
# 'extra' functions to enable parsing.
#contains:
#. Namespace fix
# Function to resolve references in CSML doc.

import Parser
import sys
import elementtree.ElementTree as etree


#some xml/string manipulation functions. may as well go in this file for now:


def PrettyPrint(elem,indent='',html=0,space='   '):
        '''Lightweight pretty printing of elementTree elements'''
        def estrip(elem):
                ''' Just want to get rid of unwanted whitespace '''
                if elem is None:
                        return ''
                else:
                        return elem.strip()
        strAttrib=''
        for att in elem.attrib:
                strAttrib+=' %s="%s"'%(att,elem.attrib[att])
        result='%s<%s%s>%s'%(indent,elem.tag,strAttrib,estrip(elem.text))
        children=len(elem)
        if children:
                for item in elem:
                        result+='\n'+PrettyPrint(item,indent=indent+space)
                result+='\n%s%s</%s>'%(indent,estrip(item.tail),elem.tag)
        else:
                result+='</%s>'%(elem.tag)
        return result

# fixing up namespaces:
def removeInlineNS(csmlstring):
        #  removeInlineNS: function removes "inline" namespaces and declares them as part of the Dataset element.
        #TODO: This whole thing is hardcoded and needs reviewing.
        
        #first need to update the ElementTree namespace map:
        etree._namespace_map.update({
        'http://www.opengis.net/om': 'om',  'http://www.opengis.net/gml': 'gml','http://ndg.nerc.ac.uk/csml' : 'csml', 'http://www.w3.org/1999/xlink':'xlink'})
    
        #replace any fully qualified namespaces
        csmlstring=csmlstring.replace('{http://www.opengis.net/gml}', 'gml:')
        csmlstring=csmlstring.replace('{http://ndg.nerc.ac.uk/csml}','')
        csmlstring=csmlstring.replace('{http://www.w3.org/1999/xlink}','xlink:')
        csmlstring=csmlstring.replace('{http://www.opengis.net/om}','om:')        
        csmlstring=csmlstring.replace('{http://ndg.nerc.ac.uk/moles}','moles:')
        #remove cmsl: prefixes
        csmlstring=csmlstring.replace('<csml:','<')
        csmlstring=csmlstring.replace('</csml:','</')
        
        #add namespace declarations at top of document
        csmlstring=csmlstring.replace('<Dataset', '<Dataset xmlns="http://ndg.nerc.ac.uk/csml"  xmlns:gml="http://www.opengis.net/gml" xmlns:om="http://www.opengis.net/om" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xmlns:moles="http://ndg.nerc.ac.uk/moles"  xmlns:xlink="http://www.w3.org/1999/xlink" xsi:schemaLocation="http://ndg.nerc.ac.uk/csml/XMLSchemas/CSMLAppSchema.xsd"')
        
        #this is only used when creating a new csml document
        csmlstring=csmlstring.replace('ns0', 'gml')
        csmlstring=csmlstring.replace('xmlns:ns1', 'xmlns:xlink')
        csmlstring=csmlstring.replace('ns1:href', 'xlink:href')
        csmlstring=csmlstring.replace('ns1:moles', 'xlink:moles')
        csmlstring=csmlstring.replace('ns1:', '')  #the rest of the ns1s are CSML namespaces... due to the way it has been parsed.        
        return csmlstring


class ParserPostProcessor:
    def __init__(self,dataset):
        if isinstance(dataset,Parser.Dataset):
            self.dataset=dataset
        else: 
            print 'Error: Could not post-process CSML document'
            sys.exit()
    
    def resolveReferences(self):
        #if there are any phenomenon defintions then 
         #for every <parameter> element, need to resolve any hrefs.
        if hasattr(self.dataset, 'phenomenonDefinitions'):
            for feature in self.dataset.featureCollection.members:
                if hasattr(feature, 'parameter'):
                    if hasattr(feature.parameter,'href'):
                        #search phenomenon definitions for a match.
                        for definition in self.dataset.phenomenonDefinitions.definitionMembers:
                            if definition.id == feature.parameter.href[1:]:  #remove '#'
                                #remove href and replace with object
                                feature.parameter.href =None
                                feature.parameter=definition
                                break
        
        #if there are embedded reference system definitions then 
        #for every DomainReference and DomainComplement need to resolve any srsNames
        if hasattr(self.dataset, 'referenceSystemDefinitions'):
            for feature in self.dataset.featureCollection.members:
                if hasattr(feature, 'domain'):  #Trajectory features don't have domains..
                    if hasattr(feature.domain.domainReference, 'srsName'):
                        #search reference definitions for a match.
                        for definition in self.dataset.referenceSystemDefinitions.definitionMembers:
                            if definition.id == feature.domain.domainReference.srsName[1:]:
                                #remove srsName attribute and replace with Object.
                                pass #Leave this for now.
                    if hasattr(feature.domain, 'domainComplement'):
                        if hasattr(feature.domain.domainComplement, 'srsName'):
                            #search reference definitions for a match.
                            for definition in self.dataset.referenceSystemDefinitions.definitionMembers:
                                if definition.id == feature.domain.domainComplement.srsName[1:]:
                                        #remove srsName attribute and replace with Object.
                                    pass #Leave this for now.
       
       #NetCDFExtracts etc may be referenced e.g. as <axisvalues> element. Need to fix these too.
        if hasattr(self.dataset, 'arrayDescriptors'):
            for feature in self.dataset.featureCollection.members:
                if hasattr (feature, 'domain'):
                    if hasattr(feature.domain, 'domainComplement'):
                        if isinstance(feature.domain.domainComplement, Parser.Grid):
                                for ordinate in feature.domain.domainComplement.ordinates:
                                    #search arrayDescriptors for a match.
                                    for arrayDescriptor  in self.dataset.arrayDescriptors:
                                        if arrayDescriptor.id == ordinate.axisValues[1:]:  #remove '#'
                                            ordinate.axisValues  = arrayDescriptor
                                            break
                                        
        #uom may reference UnitDefinitions. Replace these too
        if hasattr(self.dataset,'arrayDescriptors'):
            for feature in self.dataset.featureCollection.members:
                if hasattr (feature, 'rangeSet'):
                    if hasattr (feature.rangeSet, 'aggregatedArray'):
                         if hasattr(feature.rangeSet.aggregatedArray, 'uom'):
                             if hasattr(self.dataset, 'unitDefinitions'):
                                for definition in self.dataset.unitDefinitions.definitionMembers:
                                    if definition.id== feature.rangeSet.aggregatedArray.uom[1:]:
                                        feature.rangeSet.aggregatedArray.uom=definition
                                        break
        #TODO need to also take into account the fact that UnitDefinitions could be externally defined.
        
        return self.dataset