Add script to validate files

scripts/trad.py

@@ -361,8 +361,74 @@ def move_non_tgh_to_extend():
     save_trad_table(table)
 
 
+def validate_trad():
+    table = load_trad_table()
+
+    src = os.path.join(root, 'sts', 'data', 'scheme', 'ts_multi.tsv')
+    with open(src, encoding='UTF-8') as fh:
+        for line in fh:
+            line = line.rstrip('\n')
+            if not line:
+                continue
+            trad, simps, *_ = line.split('\t')
+            simps = [s for s in simps.split(' ') if s]
+            if trad not in table:
+                print(f'WARNING: *{trad}* (ts_multi.tsv) is not defined in traditional.tsv')
+
+    src = os.path.join(root, 'sts', 'data', 'scheme', 'st_multi.tsv')
+    with open(src, encoding='UTF-8') as fh:
+        for line in fh:
+            line = line.rstrip('\n')
+            if not line:
+                continue
+            simp, trads, *_ = line.split('\t')
+            trads = [s for s in trads.split(' ') if s]
+            for trad in trads:
+                if trad not in table:
+                    print(f'WARNING: {simp} => *{trad}* (st_multi.tsv) is not defined in traditional.tsv')
+
+    src = os.path.join(root, 'sts', 'data', 'dictionary', 'STCharacters.txt')
+    src_base = os.path.basename(src)
+    with open(src, encoding='UTF-8') as fh:
+        for line in fh:
+            line = line.rstrip('\n')
+            if not line:
+                continue
+            simp, trads, *_ = line.split('\t')
+            trads = [s for s in trads.split(' ') if s]
+            simp_found = False
+            for trad in trads:
+                try:
+                    entry = table[trad]
+                except KeyError:
+                    print(f'WARNING: {simp} => *{trad}* ({src_base}) is not defined in traditional.tsv')
+                else:
+                    if simp in (entry.cn + entry.cnx):
+                        simp_found = True
+            if not simp_found:
+                print(f'WARNING: *{simp}* ({src_base}) is not defined in traditional.tsv')
+
+    src = os.path.join(root, 'sts', 'data', 'dictionary', 'TSCharacters.txt')
+    src_base = os.path.basename(src)
+    with open(src, encoding='UTF-8') as fh:
+        for line in fh:
+            line = line.rstrip('\n')
+            if not line:
+                continue
+            trad, simps, *_ = line.split('\t')
+            simps = [s for s in simps.split(' ') if s]
+            try:
+                entry = table[trad]
+            except KeyError:
+                print(f'WARNING: *{trad}* ({src_base}) is not defined in traditional.tsv')
+            else:
+                for simp in simps:
+                    if simp not in (entry.cn + entry.cnx):
+                        print(f'WARNING: {trad} => *{simp}* ({src_base}) is not defined in traditional.tsv')
+
+
 def main():
-    tidy_trad_table()
+    validate_trad()
 
 
 if __name__ == '__main__':