pycaption 2.3.6__tar.gz → 2.3.7__tar.gz

{pycaption-2.3.6/pycaption.egg-info → pycaption-2.3.7}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: pycaption
-Version: 2.3.6
+Version: 2.3.7
 Summary: Closed caption converter
 Author: Joe Norton
 Author-email: joey@nortoncrew.com

{pycaption-2.3.6 → pycaption-2.3.7}/pycaption/sami/reader.py

@@ -7,10 +7,7 @@ styles, margin-based positioning, and multi-language content.
 from bs4 import BeautifulSoup, NavigableString
 
 from ..base import BaseReader, Caption, CaptionList, CaptionNode, CaptionSet
-from ..exceptions import (
-    CaptionReadNoCaptions,
-    CaptionReadTimingError,
-)
+from ..exceptions import CaptionReadNoCaptions, CaptionReadTimingError
 from ..geometry import Alignment, HorizontalAlignmentEnum, Layout, Padding, Size
 from .parser import SAMIParser
 
@@ -65,7 +62,8 @@ class SAMIReader(BaseReader):
             caption_dict[language] = lang_captions
 
         caption_set = CaptionSet(
-            caption_dict, layout_info=global_layout,
+            caption_dict,
+            layout_info=global_layout,
             visual_alignment_default=HorizontalAlignmentEnum.LEFT,
         )
 
@@ -244,7 +242,7 @@ class SAMIReader(BaseReader):
         attrs = {}
         css_attrs = tag.attrs
 
-        if "class" in css_attrs:
+        if "class" in css_attrs and css_attrs["class"]:
             attrs["class"] = css_attrs["class"][0].lower()
         if "id" in css_attrs:
             attrs["class"] = css_attrs["id"].lower()

{pycaption-2.3.6 → pycaption-2.3.7}/pycaption/sami/writer.py

@@ -42,9 +42,7 @@ class SAMIWriter(BaseWriter):
         """Serialize a CaptionSet into a SAMI document string."""
         caption_set = deepcopy(caption_set)
         source_default = self._get_visual_alignment_default(caption_set)
-        self._promote_center = (
-            source_default == HorizontalAlignmentEnum.CENTER
-        )
+        self._promote_center = source_default == HorizontalAlignmentEnum.CENTER
         sami = BeautifulSoup(SAMI_BASE_MARKUP, "lxml-xml")
 
         caption_set.layout_info = self._relativize_and_fit_to_screen(
@@ -310,7 +308,7 @@ class SAMIWriter(BaseWriter):
 
         style = ""
         klass = ""
-        if "classes" in content:
+        if "classes" in content and content["classes"]:
             klass += f' class="{" ".join(content["classes"])}"'
         elif "class" in content:
             klass += f' class="{content["class"]}"'

{pycaption-2.3.6 → pycaption-2.3.7}/pycaption/scc/writer.py

@@ -99,6 +99,7 @@ class SCCWriter(BaseWriter):
         codes = self._encode_all(classified)
         codes = self._adjust_timing(codes)
         codes = self._deduplicate_timestamps(codes)
+        codes = self._clamp_end_times(codes)
         output += self._render_mixed(codes)
 
         return output
@@ -195,6 +196,17 @@ class SCCWriter(BaseWriter):
             last_emitted_frame = self._microseconds_to_frame(start)
         return codes
 
+    @staticmethod
+    def _clamp_end_times(codes):
+        """Suppress end times that drift behind start after timing adjustments.
+        When dense cues push start forward beyond the original end, emit no
+        clear command (end=None) rather than producing an invalid negative-
+        duration cue."""
+        for index, (code, start, end, mode, depth) in enumerate(codes):
+            if end is not None and end <= start:
+                codes[index] = (code, start, None, mode, depth)
+        return codes
+
     def _render_mixed(self, codes):
         """Serialize all cues in chronological order, emitting mode-appropriate
         preambles inline. Handles pop-on (ENM+RCL...EDM+EOC), roll-up
@@ -225,12 +237,7 @@ class SCCWriter(BaseWriter):
         code_tokens = code.split()
 
         if len(code_tokens) <= max_payload:
-            return (
-                f"{ts}\t"
-                "94ae 94ae 9420 9420 "
-                f"{code}"
-                "942c 942c 942f 942f\n\n"
-            )
+            return f"{ts}\t" "94ae 94ae 9420 9420 " f"{code}" "942c 942c 942f 942f\n\n"
 
         output = ""
         offset = 0
@@ -243,11 +250,7 @@ class SCCWriter(BaseWriter):
             is_last = offset + max_payload >= len(code_tokens)
             if is_last:
                 line = line + ["942c", "942c", "942f", "942f"]
-            output += (
-                f"{self._format_timestamp(start)}\t"
-                + " ".join(line)
-                + "\n\n"
-            )
+            output += f"{self._format_timestamp(start)}\t" + " ".join(line) + "\n\n"
             offset += max_payload
             if not is_last:
                 start += MICROSECONDS_PER_CODEWORD
@@ -329,6 +332,7 @@ class SCCWriter(BaseWriter):
                 base_row = max(1, min(15, round((y.value - 5) / 90.0 * 15) + 1))
             else:
                 base_row = 15
+            base_row = max(1, min(base_row, 16 - num_lines))
             row = base_row + line_index
             return min(row, 15)
         return max(1, min(15, 16 - num_lines + line_index))

{pycaption-2.3.6 → pycaption-2.3.7/pycaption.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: pycaption
-Version: 2.3.6
+Version: 2.3.7
 Summary: Closed caption converter
 Author: Joe Norton
 Author-email: joey@nortoncrew.com

{pycaption-2.3.6 → pycaption-2.3.7}/setup.py

@@ -20,7 +20,7 @@ transcript_dependencies = ["nltk==3.10.0"]
 
 setup(
     name="pycaption",
-    version="2.3.6",
+    version="2.3.7",
     description="Closed caption converter",
     long_description=open(README_PATH).read(),
     author="Joe Norton",

{pycaption-2.3.6 → pycaption-2.3.7}/tests/test_scc_writer.py

@@ -185,6 +185,30 @@ class TestSCCWriterOverlappingCues:
         # (only the last cue gets a clear-screen at its end time)
         assert len(clear_lines) <= 1
 
+    def test_dense_cues_never_produce_negative_duration(self):
+        """When many dense cues push start times forward via timing
+        adjustments, the end time must never be less than the start."""
+        lines = ["WEBVTT", ""]
+        for i in range(20):
+            start_ms = 1000 + i * 200
+            end_ms = start_ms + 150
+            s, e = start_ms / 1000, end_ms / 1000
+            lines.append(
+                f"00:00:{int(s):02d}.{int((s % 1) * 1000):03d} --> "
+                f"00:00:{int(e):02d}.{int((e % 1) * 1000):03d}"
+            )
+            lines.append(
+                f"Caption number {i + 1:02d} with a lot of extra "
+                f"text padding here for testing!!"
+            )
+            lines.append("")
+        captions = WebVTTReader().read("\n".join(lines))
+        scc_output = SCCWriter().write(captions)
+        reread = SCCReader().read(scc_output)
+        caps = reread.get_captions(reread.get_languages()[0])
+        for cap in caps:
+            assert cap.end >= cap.start
+
 
 class TestSCCWriterSplitLongCaption:
     def test_split_caption_exceeding_80_tokens(self):
@@ -319,6 +343,43 @@ class TestSCCWriterPositioning:
         pac_col_28 = WRITER_PAC_CODES[(15, 28, "plain")]
         assert pac_col_28 in output
 
+    def test_multiline_bottom_stacks_upward(self):
+        vtt = (
+            "WEBVTT\n\n"
+            "00:00:01.000 --> 00:00:04.000 line:100%\n"
+            "This is line one\n"
+            "This is line two\n"
+        )
+        captions = WebVTTReader().read(vtt)
+        scc_output = SCCWriter().write(captions)
+        reread = SCCReader().read(scc_output)
+        caps = reread.get_captions(reread.get_languages()[0])
+        assert caps[0].get_text() == "This is line one\nThis is line two"
+        pac_row_14 = WRITER_PAC_CODES[(14, 0, "plain")]
+        pac_row_15 = WRITER_PAC_CODES[(15, 0, "plain")]
+        assert pac_row_14 in scc_output
+        assert pac_row_15 in scc_output
+
+    def test_three_line_at_86_percent_stacks_upward(self):
+        vtt = (
+            "WEBVTT\n\n"
+            "00:00:01.000 --> 00:00:04.000 line:86%\n"
+            "First line\n"
+            "Second line\n"
+            "Third line\n"
+        )
+        captions = WebVTTReader().read(vtt)
+        scc_output = SCCWriter().write(captions)
+        reread = SCCReader().read(scc_output)
+        caps = reread.get_captions(reread.get_languages()[0])
+        assert caps[0].get_text() == "First line\nSecond line\nThird line"
+        pac_row_13 = WRITER_PAC_CODES[(13, 0, "plain")]
+        pac_row_14 = WRITER_PAC_CODES[(14, 0, "plain")]
+        pac_row_15 = WRITER_PAC_CODES[(15, 0, "plain")]
+        assert pac_row_13 in scc_output
+        assert pac_row_14 in scc_output
+        assert pac_row_15 in scc_output
+
 
 class TestSCCWriterStyles:
     def test_vtt_italic_emits_mid_row_code(self):

{pycaption-2.3.6 → pycaption-2.3.7}/tests/test_webvtt_conversion.py

@@ -672,3 +672,25 @@ class TestGlobalCueColorPropagation:
         style_opens = [n for n in nodes if n.type_ == CaptionNode.STYLE and n.start]
         assert len(style_opens) == 1
         assert style_opens[0].content.get("color") == "red"
+
+
+class TestYouTubeKaraokeEmptyClassRoundtrip:
+    def test_empty_class_vtt_to_sami_roundtrip(self):
+        vtt = (
+            "WEBVTT\n\n"
+            "00:00:01.000 --> 00:00:04.000\n"
+            "<c> Hello</c><c> world</c>\n"
+        )
+        caption_set = WebVTTReader().read(vtt)
+        sami_output = SAMIWriter().write(caption_set)
+
+        assert 'class=""' not in sami_output
+        assert "Hello" in sami_output
+        assert "world" in sami_output
+
+        reread_set = SAMIReader().read(sami_output)
+        captions = reread_set.get_captions(reread_set.get_languages()[0])
+        assert len(captions) == 1
+        text = captions[0].get_text()
+        assert "Hello" in text
+        assert "world" in text