Changeset 35127

Timestamp:

2021-04-27T14:11:45+12:00 (3 years ago)

Author:

davidb

Message:

Code to extract ESC Logo for year

File:

• main/trunk/model-sites-dev/eurovision-lod/collect/eurovision/prepare/errata-categories/escwikipedia.py (modified) (1 diff)

main/trunk/model-sites-dev/eurovision-lod/collect/eurovision/prepare/errata-categories/escwikipedia.py

@@ -302 +302 @@
     ## debug_output_country_year_recs(country_year_recs)
 
+    # Splice in logo image for that year into each country_year_rec
+    #infobox_table = esc_year_soup.find("table",{"class": "infobox"})
+    #infobox_td = infobox_table.find("td",{"class": "infobox-image"})
+    #infobox_img = infobox_td.find("img")
+
+    #infobox_img = esc_year_soup.select("table.infobox td.infobox-image img")
+
+    #infobox_img = esc_year_soup.select("table.infobox td.infobox-image img")
+    #infobox_img = esc_year_soup.select("table.infobox td.infobox-image")
+    # infobox_img = esc_year_soup.select("table.infobox")
+
+    #infobox_table = esc_year_soup.find("table",{"class": "infobox"})
+    #infobox_img = infobox_table.tbody.tr[0].td[0].a.img
+
+    # Looks like Wikipedia has changed some of its infobox CSS
+    # If processing newer downloads, then the following is probably the select
+    # statement to use
+
+    # new school!
+    #infobox_img = esc_year_soup.select("table.infobox td.infobox-image img")
+
+    # old school
+    infobox_imgs = esc_year_soup.select("table.infobox tr td a.image img")
+    
+    if (len(infobox_imgs) == 0):
+        print("****")
+        print("****!!! No ESC Logo image found!!!")
+        print("****")
+    else:
+        # Some pages include addition image graphics, such as a map showing country entrants
+        # => Want the first one
+        infobox_logo_img = infobox_imgs[0];
+
+        for country_year_key in country_year_recs.keys():
+            country_year_rec = country_year_recs.get(country_year_key)
+            
+            country_year_rec["YearLogoImg"] = str(infobox_logo_img)
+
+            
     return country_year_recs