XML parsers for GIS metadata that are designed to read in, validate, update and output a core set of properties that have been mapped between the most common standards, currently:
- FGDC
- ISO-19139 (and ISO-19115)
- ArcGIS (tested with ArcGIS format 1.0).
This library is compatible with Python versions 2.7 and 3.4 through 3.6.
Install with pip install gis-metadata-parser
.
Parsers can be instantiated from files, XML strings or URLs. They can be converted from one standard to another as well.
from gis_metadata.arcgis_metadata_parser import ArcGISParser
from gis_metadata.fgdc_metadata_parser import FgdcParser
from gis_metadata.iso_metadata_parser import IsoParser
from gis_metadata.metadata_parser import get_metadata_parser
# From file objects
with open(r'/path/to/metadata.xml') as metadata:
fgdc_from_file = FgdcParser(metadata)
with open(r'/path/to/metadata.xml') as metadata:
iso_from_file = IsoParser(metadata)
# Detect standard based on root element, metadata
fgdc_from_string = get_metadata_parser(
"""
<?xml version='1.0' encoding='UTF-8'?>
<metadata>
<idinfo>
</idinfo>
</metadata>
"""
)
# Detect ArcGIS standard based on root element and its nodes
iso_from_string = get_metadata_parser(
"""
<?xml version='1.0' encoding='UTF-8'?>
<metadata>
<dataIdInfo/></dataIdInfo>
<distInfo/></distInfo>
<dqInfo/></dqInfo>
</metadata>
"""
)
# Detect ISO standard based on root element, MD_Metadata or MI_Metadata
iso_from_string = get_metadata_parser(
"""
<?xml version='1.0' encoding='UTF-8'?>
<MD_Metadata>
<identificationInfo>
</identificationInfo>
</MD_Metadata>
"""
)
# Convert from one standard to another
fgdc_converted = iso_from_file.convert_to(FgdcParser)
iso_converted = fgdc_from_file.convert_to(IsoParser)
arcgis_converted = iso_converted.convert_to(ArcGISParser)
Finally, the properties of the parser can be updated, validated, applied and output:
with open(r'/path/to/metadata.xml') as metadata:
fgdc_from_file = FgdcParser(metadata)
# Example simple properties
fgdc_from_file.title
fgdc_from_file.abstract
fgdc_from_file.place_keywords
fgdc_from_file.thematic_keywords
# :see: gis_metadata.utils.get_supported_props for list of all supported properties
# Complex properties
fgdc_from_file.attributes
fgdc_from_file.bounding_box
fgdc_from_file.contacts
fgdc_from_file.dates
fgdc_from_file.digital_forms
fgdc_from_file.larger_works
fgdc_from_file.process_steps
fgdc_from_file.raster_info
# :see: gis_metadata.utils.get_complex_definitions for structure of all complex properties
# Update properties
fgdc_from_file.title = 'New Title'
fgdc_from_file.dates = {'type': 'single' 'values': '1/1/2016'}
# Apply updates
fgdc_from_file.validate() # Ensure updated properties are valid
fgdc_from_file.serialize() # Output updated XML as a string
fgdc_from_file.write() # Output updated XML to existing file
fgdc_from_file.write(out_file_or_path='/path/to/updated.xml') # Output updated XML to new file
There are a few unwritten (until now) rules about the way the metadata parsers are wired to work:
- Properties are generally defined by XPATH in each
parser._data_map
- Simple parser properties accept only values of
string
andlist
's ofstring
's - XPATH's configured in the data map support references to element attributes:
'path/to/element/@attr'
- Complex parser properties are defined by custom parser/updater functions instead of by XPATH
- Complex parser properties accept values of type
dict
containing simple properties, or a list of saiddict
's - Properties with leading underscores are parsed, but not validated or written out
- Properties that "shadow" other properties but with a leading underscore serve as backup values
- For backup properties, additional underscores indicate further backup options, i.e.
title
,_title
,__title
Some examples of existing backup properties are as follows:
# In the ArcGIS parser for distribution contact phone:
_agis_tag_formats = {
...
'dist_phone': 'distInfo/distributor/distorCont/rpCntInfo/cntPhone/voiceNum',
'_dist_phone': 'distInfo/distributor/distorCont/rpCntInfo/voiceNum', # If not in cntPhone
...
}
# In the FGDC parser for sub-properties in the contacts definition:
_fgdc_definitions = get_complex_definitions()
_fgdc_definitions[CONTACTS].update({
'_name': '{_name}',
'_organization': '{_organization}'
})
...
class FgdcParser(MetadataParser):
...
def _init_data_map(self):
...
ct_format = _fgdc_tag_formats[CONTACTS]
fgdc_data_structures[CONTACTS] = format_xpaths(
...
name=ct_format.format(ct_path='cntperp/cntper'),
_name=ct_format.format(ct_path='cntorgp/cntper'), # If not in cntperp
organization=ct_format.format(ct_path='cntperp/cntorg'),
_organization=ct_format.format(ct_path='cntorgp/cntorg'), # If not in cntperp
)
# Also see the ISO parser for backup sub-properties in the attributes definition:
_iso_definitions = get_complex_definitions()
_iso_definitions[ATTRIBUTES].update({
'_definition_source': '{_definition_src}',
'__definition_source': '{__definition_src}',
'___definition_source': '{___definition_src}'
})
Any of the supported parsers can be extended to include more of a standard's supported data. In this example we'll add two new properties to the IsoParser
:
metadata_language
: a simple string field describing the language of the metadata file itself (not the dataset)metadata_contacts
: a complex structure with contact info leveraging and enhancing the existing contact structure
This example will cover:
- Adding a new simple property
- Configuring a backup location for a property
- Referencing an element attribute in an XPATH
- Adding a new complex property
- Customizing the complex property to include a new sub-property
Also, this example is specifically covered by unit tests.
from gis_metadata.iso_metadata_parser import IsoParser
from gis_metadata.utils import CONTACTS, format_xpaths, get_complex_definitions, ParserProperty
class CustomIsoParser(IsoParser):
def _init_data_map(self):
super(CustomIsoParser, self)._init_data_map()
# Basic property: text or list (with backup location referencing codeListValue attribute)
lang_prop = 'metadata_language'
self._data_map[lang_prop] = 'language/CharacterString' # Parse from here if present
self._data_map['_' + lang_prop] = 'language/LanguageCode/@codeListValue' # Otherwise, try from here
# Complex structure (reuse of contacts structure plus phone)
# Define some basic variables
ct_prop = 'metadata_contacts'
ct_xpath = 'contact/CI_ResponsibleParty/{ct_path}'
ct_defintion = get_complex_definitions()[CONTACTS]
ct_defintion['phone'] = '{phone}'
# Reuse CONTACT structure to specify locations per prop (adapted only slightly from parent)
self._data_structures[ct_prop] = format_xpaths(
ct_defintion,
name=ct_xpath.format(ct_path='individualName/CharacterString'),
organization=ct_xpath.format(ct_path='organisationName/CharacterString'),
position=ct_xpath.format(ct_path='positionName/CharacterString'),
phone=ct_xpath.format(
ct_path='contactInfo/CI_Contact/phone/CI_Telephone/voice/CharacterString'
),
email=ct_xpath.format(
ct_path='contactInfo/CI_Contact/address/CI_Address/electronicMailAddress/CharacterString'
)
)
# Set the contact root to insert new elements at "contact" level given the defined path:
# 'contact/CI_ResponsibleParty/...'
# By default we would get multiple "CI_ResponsibleParty" elements under a single "contact"
# This way we get multiple "contact" elements, each with its own single "CI_ResponsibleParty"
self._data_map['_{prop}_root'.format(prop=ct_prop)] = 'contact'
# Use the built-in support for parsing complex properties (or write your own a parser/updater)
self._data_map[ct_prop] = ParserProperty(self._parse_complex_list, self._update_complex_list)
# And finally, let the parent validation logic know about the two new custom properties
self._metadata_props.add(lang_prop)
self._metadata_props.add(ct_prop)
with open(r'/path/to/metadata.xml') as metadata:
iso_from_file = CustomIsoParser(metadata)
iso_from_file.metadata_language
iso_from_file.metadata_contacts