#7 add keywords and contacts readers

isogeo_xml_toolbelt/list_themes_inspire.txt

@@ -0,0 +1,34 @@
+Altitude
+Référentiels de coordonnées
+Unités statistiques
+Répartition de la population — démographie
+Zones de gestion, de restriction ou de réglementation et unités de déclaration
+Zones à risque naturel
+Conditions atmosphériques
+Caractéristiques géographiques météorologiques
+Caractéristiques géographiques océanographiques
+Régions maritimes
+Régions biogéographiques
+Habitats et biotopes
+Répartition des espèces
+Bâtiments
+Occupation des terres
+Systèmes de maillage géographique
+Sources d'énergie
+Ressources minérales
+Dénominations géographiques
+Ortho-imagerie
+Sols
+Géologie
+Unités administratives
+Usage des sols
+Adresses
+Santé et sécurité des personnes
+Parcelles cadastrales
+Services d'utilité publique et services publics
+Installations de suivi environnemental
+Réseaux de transport
+Hydrographie
+Lieux de production et sites industriels
+Installations agricoles et aquacoles
+Sites protégés

isogeo_xml_toolbelt/reader_iso19139.py

@@ -25,11 +25,17 @@
 from lxml import etree
 
 # submodules
+try:
+    from .xml_19139_fields import Contact
+except (ImportError, ValueError, SystemError):
+    from xml_19139_fields import Contact
+
 try:
     from .xml_utils import XmlUtils
 except (ImportError, ValueError, SystemError):
     from xml_utils import XmlUtils
 
+
 # #############################################################################
 # ########## Globals ###############
 # ##################################
@@ -40,6 +46,7 @@
 # utils
 utils = XmlUtils()
 
+
 # #############################################################################
 # ########## Classes ###############
 # ##################################
@@ -147,12 +154,13 @@ def __init__(self, xml: Path):
             "gco:DateTime/text()",
             self.namespaces)
         self.md_date = utils.parse_string_for_max_date(md_dates_str)
-        self.contact = {
-            "mails": self.md.xpath(
-                "/gmd:MD_Metadata/gmd:contact/gmd:CI_ResponsibleParty/gmd:contactInfo/"
-                "gmd:CI_Contact/gmd:address/gmd:CI_Address/gmd:electronicMailAddress/gco:CharacterString/text()",
-                namespaces=self.namespaces)
-        }
+
+        #contacts
+        self.list_contacts = self.get_md_contacts()
+
+        #keywords
+        self.keywords = self.get_md_keywords()     
+
         # bounding box
         self.bbox = []
         try:
@@ -186,9 +194,6 @@ def __init__(self, xml: Path):
             self.latmin = -90
             self.latmax = 90
 
-        #Vector geometry
-        # self.geometry = self.get_vector_geometry(self.md)
-
         self.geometry = utils.xmlGetTextTag(
             self.md,
             "gmd:spatialRepresentationInfo/gmd:MD_VectorSpatialRepresentation/"
@@ -226,22 +231,112 @@ def __repr__(self):
     def __str__(self):
         return self.fileIdentifier
 
+    def get_md_contacts(self) -> dict:
+
+        md_contact = list()
+
+        root = self.md.getroot() #get xml root
+
+        #get contacts in gmd:contact
+        for ct in root.findall("gmd:contact/", self.namespaces): 
+
+            md_contact.append(Contact(ct, self.namespaces).asDict())
+
+        #get contacts in gmd:pointOfContact
+        for pct in root.findall("gmd:identificationInfo/gmd:MD_DataIdentification/gmd:pointOfContact/", self.namespaces): 
+
+            md_contact.append(Contact(pct, self.namespaces).asDict()) 
+
+        return md_contact  
+        #     try:
+        #         name = contact.find("gmd:individualName/gco:CharacterString", self.namespaces).text   
+        #     except:
+        #         name = None
+        #     try:
+        #         organisation = contact.find("gmd:organisationName/gco:CharacterString", self.namespaces).text  
+        #     except:
+        #         organisation = None
+
+        #     adr_path = "gmd:contactInfo/gmd:CI_Contact/gmd:address/gmd:CI_Address/"
+
+        #     try:
+        #         rue = contact.find(adr_path + "gmd:deliveryPoint/gco:CharacterString", self.namespaces).text 
+        #     except:
+        #         rue = None
+        #     try:
+        #         ville = contact.find(adr_path + "gmd:city/gco:CharacterString", self.namespaces).text   
+        #     except:
+        #         ville = None 
+        #     try:
+        #         cp = contact.find(adr_path + "gmd:postalCode/gco:CharacterString", self.namespaces).text   
+        #     except:
+        #         cp = None
+        #     try:
+        #         country = contact.find(adr_path + "gmd:country/gco:CharacterString", self.namespaces).text 
+        #     except:
+        #         country = None
+        #     try:
+        #         mail = contact.find(adr_path + "gmd:electronicMailAddress/gco:CharacterString", self.namespaces).text
+        #     except:
+        #         mail = None
+
+        #     try:
+        #         telephone = contact.find("gmd:contactInfo/gmd:CI_Contact/"
+        #                                 "gmd:phone/gmd:CI_Telephone/gmd:voice/gco:CharacterString", self.namespaces).text
+        #     except:
+        #         telephone = None
+
+        #     try:
+        #         role = contact.find("gmd:role/gmd:CI_RoleCode", self.namespaces).get("codeListValue") 
+        #     except:
+        #         role = None
+
+        #     md_contact.append({"name": name,"organisation": organisation,"role": role,"rue": rue,"ville": ville,
+        #                 "cp": cp,"country": country,"mail": mail,"telephone": telephone})
+
+        # return md_contact
+
+
+    def get_md_keywords(self) -> list:
+
+        md_keywords = list()
+
+        root = self.md.getroot() #get xml root
+
+        #get contacts in gmd:contact
+        for kw in root.findall("gmd:identificationInfo/"\
+                                "gmd:MD_DataIdentification/"\
+                                    "gmd:descriptiveKeywords/"\
+                                        "gmd:MD_Keywords/gmd:keyword/gco:CharacterString"\
+                                        , self.namespaces):
+
+            #Test for exceptions like <gco:CharacterString>cycles ; circulations douces ; vélo ; aménagements cyclables ; transport ; véloroute ;</gco:CharacterString>
+            keyword = kw.text.split(";") 
+            if (len(keyword) > 1):
+                for k in keyword:
+                    md_keywords.append(k)
+            else:        
+                md_keywords.append(kw.text)
+
+
+        return md_keywords 
+
     def asDict(self) -> dict:
         """Retrun object as a structured dictionary key: value."""
         return {
             "filename": self.filename,
             "fileIdentifier": self.fileIdentifier,
             "MD_Identifier": self.MD_Identifier,
-            "md_date": self.md_date,
+            "type": self.storageType,
             "title": self.title,
-            "OrganisationName": self.OrganisationName,
             "abstract": self.abstract,
-            "parentidentifier": self.parentIdentifier,
-            "type": self.storageType,
+            "OrganisationName": self.OrganisationName,
+            "keywords": self.keywords,
             "formatName": self.formatName,
             "formatVersion": self.formatVersion,
+            "contacts": self.list_contacts,
+            "md_date": self.md_date,
             "date": self.date,
-            "contact": self.contact,
             "geometry": self.geometry,
             "srs": "{}:{}".format(self.srs_codeSpace, self.srs_code),
             "latmin": self.latmin,
@@ -250,21 +345,23 @@ def asDict(self) -> dict:
             "lonmax": self.lonmax,
             "featureCount": self.featureCount,
             "featureCatalogs": self.featureCatalogs,
-            "storageType": self.storageType
+            "storageType": self.storageType,
+            "parentidentifier": self.parentIdentifier
 
         }
 
 
+
 # #############################################################################
 # ### Stand alone execution #######
 # #################################
 
 if __name__ == "__main__":
     """Test parameters for a stand-alone run."""
-    li_fixtures_xml = sorted(Path(r"tests/fixtures/").glob("**/*.xml"))
-    # li_fixtures_xml = sorted(Path(r"input").glob("**/*.xml"))
+    # li_fixtures_xml = sorted(Path(r"tests/fixtures/").glob("**/*.xml"))
+    li_fixtures_xml = sorted(Path(r"input").glob("**/*.xml"))
     for xml_path in li_fixtures_xml:
         test = MetadataIso19139(xml=xml_path)
-        # print(test.asDict().get("title"), test.asDict().get("storageType"))
-        print(test.asDict())
+        print(test.asDict().get("title"), test.asDict().get("keywords"))
+        # print(test.asDict())
         # print(xml_path.resolve(), test.storageType)

isogeo_xml_toolbelt/xml_19139_fields.py

@@ -0,0 +1,129 @@
+# -*- coding: utf-8 -*-
+#! python3
+
+"""
+    Isogeo XML Fixer - Metadata
+
+    Purpose:     Read a metadata stored into XML ISO 19110 as an object
+    Authors:     Isogeo, inspired by the work did by GeoBretagne on mdchecker
+    Python:      3.6.x
+"""
+
+# #############################################################################
+# ########## Libraries #############
+# ##################################
+
+# standard library
+import datetime
+import logging
+import os
+from pathlib import Path
+from uuid import UUID
+
+# 3rd party library
+import arrow
+from lxml import etree
+
+# #############################################################################
+# ########## Globals ###############
+# ##################################
+
+# logging
+logging.basicConfig(level=logging.INFO)
+
+
+# #############################################################################
+# ########## Classes ###############
+# ##################################
+
+class Contact(object):
+    """Contact in metadata XML 19139
+    
+    :param lxml.etree._ElementTree contact: Element {http://www.isotc211.org/2005/gmd}CI_ResponsibleParty
+    :param dict namespaces: XML namespaces like `lxml.etree.getroot().nsmap`
+    """
+
+    def __init__(self, contact, namespaces):
+        """Instanciation."""
+        self.namespaces = namespaces
+        try:
+            self.name = contact.find("gmd:individualName/gco:CharacterString", self.namespaces).text   
+        except:
+            self.name = None
+        try:
+            self.organisation = contact.find("gmd:organisationName/gco:CharacterString", self.namespaces).text  
+        except:
+            self.organisation = None
+
+        self.adr_path = "gmd:contactInfo/gmd:CI_Contact/gmd:address/gmd:CI_Address/"
+
+        try:
+            self.rue = contact.find(self.adr_path + "gmd:deliveryPoint/gco:CharacterString", self.namespaces).text 
+        except:
+            self.rue = None
+        try:
+            self.ville = contact.find(self.adr_path + "gmd:city/gco:CharacterString", self.namespaces).text   
+        except:
+            self.ville = None 
+        try:
+            self.cp = contact.find(self.adr_path + "gmd:postalCode/gco:CharacterString", self.namespaces).text   
+        except:
+            self.cp = None
+        try:
+            self.country = contact.find(self.adr_path + "gmd:country/gco:CharacterString", self.namespaces).text 
+        except:
+            self.country = None
+        try:
+            self.mail = contact.find(self.adr_path + "gmd:electronicMailAddress/gco:CharacterString", self.namespaces).text
+        except:
+            self.mail = None
+
+        try:
+            self.telephone = contact.find("gmd:contactInfo/gmd:CI_Contact/"
+                                        "gmd:phone/gmd:CI_Telephone/gmd:voice/gco:CharacterString", self.namespaces).text
+        except:
+            self.telephone = None
+
+        try:
+            self.role = contact.find("gmd:role/gmd:CI_RoleCode", self.namespaces).get("codeListValue") 
+        except:
+            self.role = None
+
+
+    def asDict(self) -> dict:
+        """Return contact as a structured dictionary key: value."""
+
+        return {"name": self.name,"organisation": self.organisation,"role": self.role,"rue": self.rue,
+        "ville": self.ville,"cp": self.cp,"country": self.country,"mail": self.mail,"telephone": self.telephone}
+
+# #############################################################################
+# ### Stand alone execution #######
+# #################################
+if __name__ == "__main__":
+    """Test parameters for a stand-alone run."""
+    namespaces = {
+            "gts": "http://www.isotc211.org/2005/gts",
+            "gml": "http://www.opengis.net/gml",
+            "xsi": "http://www.w3.org/2001/XMLSchema-instance",
+            "gco": "http://www.isotc211.org/2005/gco",
+            "gmd": "http://www.isotc211.org/2005/gmd",
+            "gmx": "http://www.isotc211.org/2005/gmx",
+            "srv": "http://www.isotc211.org/2005/srv",
+            "xl": "http://www.w3.org/1999/xlink"}
+
+    li_fixtures_xml = sorted(Path(r"input").glob("**/*.xml"))
+
+    for xml_path in li_fixtures_xml:
+        # lxml needs a str not a Path
+
+        xml_path = str(xml_path.resolve())
+
+        md = etree.parse(xml_path)
+        root = md.getroot() #get xml root
+
+        #get contacts in gmd:contact
+        for ct in root.findall("gmd:contact/", namespaces): 
+
+            print(Contact(ct).asDict())
+
+