Changeset 34891

Timestamp:

2021-02-16T23:19:47+13:00 (3 years ago)

Author:

davidb

Message:

Some refactoring of where routines should be based

Location:

main/trunk/model-sites-dev/eurovision-lod/collect/eurovision/prepare/voting-excel

Files:

• util.py (added)
• xlsx-to-jsonmetadata.py (modified) (6 diffs)
• xlsxutil.py (modified) (3 diffs)

main/trunk/model-sites-dev/eurovision-lod/collect/eurovision/prepare/voting-excel/xlsx-to-jsonmetadata.py

@@ -11 +11 @@
 import openpyxl
 
+import util
 import xlsxutil
 
@@ -16 +17 @@
 tele_from_countries = {}
 
-def eprint(*args, **kwargs):
-    print(*args, file=sys.stderr, **kwargs)
-    
-def fileset_voting_for_esc_country_in_year(data_hashmap):
-
-    # Only generating a single fileset record here, however
-    # the Greenstone format allows for this to be an array
-    # of fileset entries => return [ fileset ]
-    
-    fileset = {}
-    
-    return [ fileset ]
-
-
-def filter_finalist_votes(entry):
-    return entry.get('(semi-) final') == "f" and entry.get('Duplicate') == None
-
-def sortkey_year_to_country(entry):
-    return str(entry.get('Year'))+entry.get('To country')
+#def fileset_voting_for_esc_country_in_year(data_hashmap):
+#
+#    # Only generating a single fileset record here, however
+#    # the Greenstone format allows for this to be an array
+#    # of fileset entries => return [ fileset ]
+#    
+#    fileset = {}
+#    
+#    return [ fileset ]
+
+
+#def filter_finalist_votes(entry):
+#    return entry.get('(semi-) final') == "f" and entry.get('Duplicate') == None
+#
+#def sortkey_year_to_country(entry):
+#    return str(entry.get('Year'))+entry.get('To country')
 
 
@@ -49 +47 @@
     
     # Filter down to just the voting results concerning finals
-    data_hashmap_array_finals = list(filter(filter_finalist_votes, data_hashmap_array))
+    data_hashmap_array_finals = list(filter(xlsxutil.filter_finalist_votes, data_hashmap_array))
 
     # Sort so array entries are grouped by the country receiving the votes in a given year
-    data_hashmap_array_finals.sort(key=sortkey_year_to_country)
+    data_hashmap_array_finals.sort(key=xlsxutil.sortkey_year_to_country)
 
 
@@ -114 +112 @@
     jury_metadata_vals = []
     tele_metadata_vals = []
+
+    to_country_jury_total = 0
+    to_country_tele_total = 0
     
     for to_country_year_vote in to_country_year_votes:
@@ -131 +132 @@
             jury_metadata_vals.append(id_from_country+"-J")
             jury_from_countries[id_from_country] = 1
+            to_country_jury_total = to_country_jury_total + points
             
         elif (vote_type == "T"):
             tele_metadata_vals.append(id_from_country+"-T")
             tele_from_countries[id_from_country] = 1
+            to_country_tele_total = to_country_tele_total + points
         else:            
-            eprint("Warning: Unrecognized voting type: " + vote_type)
+            util.eprint("Warning: Unrecognized voting type: " + vote_type)
 
     if (len(jury_metadata_vals)>0):
         metadata_array.append({ "name": "JuryVotesJSON", "content": json.dumps(jury_metadata_vals) })
+        metadata_array.append({ "name": "JuryVotesTotal", "content": to_country_jury_total})
     if (len(tele_metadata_vals)>0):
         metadata_array.append({ "name": "TeleVotesJSON", "content": json.dumps(tele_metadata_vals) })
+        metadata_array.append({ "name": "TeleVotesTotal", "content": to_country_tele_total})
                           
     id_to_country = to_country_year_votes[0].get('To country')
@@ -224 +229 @@
         print("  " + filename_id.ljust(28) + ": " + str(num_countries_voting_data) + " votes")
 
-
-    print("")
+ 
+    print("")
+    print("")
+    print("For, e.g., '<display><format>' section of collectionConfig.xml:")
+    print("")
+    print("  <gsf:headMetaTags>")
+
     for from_country in sorted(jury_from_countries.keys()):
-        print("<gsf:metadata name=\""+from_country+"-J\" />")
+        print("    <gsf:metadata name=\""+from_country+"-J\" />")
 
     print("")
     for from_country in sorted(tele_from_countries.keys()):
-        print("<gsf:metadata name=\""+from_country+"-T\" />")
+        print("    <gsf:metadata name=\""+from_country+"-T\" />")
+
+    print("  </gsf:headMetaTags>")
+    print("")
 
     greenstone_metadata_json = { "DirectoryMetadata": directory_metadata }

main/trunk/model-sites-dev/eurovision-lod/collect/eurovision/prepare/voting-excel/xlsxutil.py

@@ -1 +1 @@
 
 import os
-#import re
 import sys
 
 import openpyxl
-
-
-
+    
 ## Fix up names to match those used in DBpedia ESC
 
 # In Excel data:
-#   The Netherlands 
+#   The Netherlands (also in one case 'The Netherands')
 #   Serbia & Montenegro
 #   Bosnia & Herzegovina
@@ -21 +18 @@
 #   Serbia and Montenegro
 #   Bosnia and Herzegovina
-#
-# Some values noted in ESC that potentially need to be changed at SPARQL query point
+#   Macedonia
+
+#*****
+# TODO:
+# Some values noted in ESC that potentially need to be changed when queries via SPARQL query point
 #   Republic of Macedonia (2005)
 #   Federal Republic of Yugoslavia (1992)
@@ -34 +34 @@
     "F.Y.R. Macedonia":     "Macedonia"
 }
+
+def filter_finalist_votes(entry):
+    return entry.get('(semi-) final') == "f" and entry.get('Duplicate') == None
+
+def sortkey_year_to_country(entry):
+    return str(entry.get('Year'))+entry.get('To country')
+
 
 def load_xslx_sheet(excel_input_filename,sheetname):