RJ3vans commited on
Commit
7480d5c
·
verified ·
1 Parent(s): 4f0d378

Update app.py

Browse files
Files changed (1) hide show
  1. app.py +11 -5
app.py CHANGED
@@ -1278,15 +1278,21 @@ def CompoundSimplify(xml_sent):
1278
  s = re.sub(r"\s+", " ", s).strip()
1279
  s = re.sub(r"\s+([.,!?;:])", r"\1", s)
1280
  s = re.sub(r"([.,!?;:])\s+", r"\1 ", s)
 
1281
  # decimals: 8 . 2 → 8.2
1282
  s = re.sub(r"(\d)\s*\.\s*(\d)", r"\1.\2", s)
1283
- s = re.sub(r",+\s*\.", ".", s)
1284
- s = re.sub(r",+\s*$", "", s)
 
 
 
 
1285
  if s and not s.endswith((".", "!", "?")):
1286
  s += "."
1287
- if s:
1288
- s = s[0].upper() + s[1:]
1289
- return s
 
1290
 
1291
  sent1 = tidy(sent1)
1292
  sent2 = tidy(sent2)
 
1278
  s = re.sub(r"\s+", " ", s).strip()
1279
  s = re.sub(r"\s+([.,!?;:])", r"\1", s)
1280
  s = re.sub(r"([.,!?;:])\s+", r"\1 ", s)
1281
+
1282
  # decimals: 8 . 2 → 8.2
1283
  s = re.sub(r"(\d)\s*\.\s*(\d)", r"\1.\2", s)
1284
+
1285
+ # list-coordinator debris: "43;." / "43,." / "43;;." → "43."
1286
+ s = re.sub(r"[,;:]+\s*\.", ".", s)
1287
+ # dangling separators at end before a period is added
1288
+ s = re.sub(r"[,;:]+\s*$", "", s)
1289
+
1290
  if s and not s.endswith((".", "!", "?")):
1291
  s += "."
1292
+ if s:
1293
+ s = s[0].upper() + s[1:]
1294
+ return s
1295
+
1296
 
1297
  sent1 = tidy(sent1)
1298
  sent2 = tidy(sent2)