parsing: Support dots in index

srt.py

@@ -22,7 +22,7 @@
 RGX_TIMESTAMP_PARSEABLE = r"^{}$".format(
     RGX_TIMESTAMP_MAGNITUDE_DELIM.join(["(" + RGX_TIMESTAMP_FIELD + ")"] * 4)
 )
-RGX_INDEX = r"-?[0-9]+"
+RGX_INDEX = r"-?[0-9.]+"
 RGX_PROPRIETARY = r"[^\r\n]*"
 RGX_CONTENT = r".*?"
 RGX_POSSIBLE_CRLF = r"\r?\n"
@@ -349,6 +349,13 @@ def parse(srt):
         # finditer and all match groups are mandatory in the regex.
         content = content.replace("\r\n", "\n")  # pytype: disable=attribute-error
 
+        try:
+            raw_index = int(raw_index)
+        except ValueError:
+            # Index 123.4. Handled separately, since it's a rare case and we
+            # don't want to affect general performance.
+            raw_index = int(raw_index.split(".")[0])
+
         yield Subtitle(
             index=int(raw_index),
             start=srt_timestamp_to_timedelta(raw_start),

tests/test_srt.py

@@ -588,6 +588,20 @@ def test_can_parse_index_trailing_ws(input_subs, whitespace):
     subs_eq(reparsed_subs, input_subs)
 
 
+@given(st.lists(subtitles()))
+def test_can_parse_index_with_dot(input_subs):
+    # Seen in Battlestar Galactica subs
+    out = ""
+
+    for sub in input_subs:
+        lines = sub.to_srt().split("\n")
+        lines[0] = lines[0] + "." + lines[0]
+        out += "\n".join(lines)
+
+    reparsed_subs = srt.parse(out)
+    subs_eq(reparsed_subs, input_subs)
+
+
 @given(st.lists(subtitles()), st.lists(st.just("0")))
 def test_can_parse_index_leading_zeroes(input_subs, zeroes):
     out = ""