Finaliza os ajustes para render de video

video_render/transcription.py

@@ -118,5 +118,75 @@ class TranscriptionService:
         with text_path.open("w", encoding="utf-8") as fp:
             fp.write(result.full_text)
 
-        logger.info("Transcrição salva em %s", destination)
+        logger.info("Transcricao salva em %s", destination)
+
+    @staticmethod
+    def load(source: Path) -> Optional[TranscriptionResult]:
+        json_path = source / "transcription.json"
+        if not json_path.exists():
+            return None
+
+        try:
+            with json_path.open("r", encoding="utf-8") as fp:
+                payload = json.load(fp)
+        except (OSError, json.JSONDecodeError) as exc:
+            logger.warning(
+                "Falha ao carregar transcricao existente de %s: %s", json_path, exc
+            )
+            return None
+
+        segments_payload = payload.get("segments", [])
+        if not isinstance(segments_payload, list):
+            logger.warning(
+                "Formato inesperado ao carregar transcricao de %s: 'segments' invalido",
+                json_path,
+            )
+            return None
+
+        segments: List[TranscriptSegment] = []
+        for idx, segment_data in enumerate(segments_payload):
+            if not isinstance(segment_data, dict):
+                logger.debug("Segmento invalido ignorado ao carregar: %s", segment_data)
+                continue
+            try:
+                segment_id = int(segment_data.get("id", idx))
+                start = float(segment_data["start"])
+                end = float(segment_data["end"])
+            except (KeyError, TypeError, ValueError):
+                logger.debug("Segmento sem dados obrigatorios ignorado: %s", segment_data)
+                continue
+
+            text = str(segment_data.get("text", "")).strip()
+            words_payload = segment_data.get("words", [])
+            words: List[WordTiming] = []
+
+            if isinstance(words_payload, list):
+                for word_data in words_payload:
+                    if not isinstance(word_data, dict):
+                        continue
+                    try:
+                        w_start = float(word_data["start"])
+                        w_end = float(word_data["end"])
+                    except (KeyError, TypeError, ValueError):
+                        logger.debug(
+                            "Palavra sem dados obrigatorios ignorada: %s", word_data
+                        )
+                        continue
+                    word_text = str(word_data.get("text", "")).strip()
+                    if not word_text:
+                        continue
+                    words.append(WordTiming(start=w_start, end=w_end, word=word_text))
+
+            segments.append(
+                TranscriptSegment(
+                    id=segment_id,
+                    start=start,
+                    end=end,
+                    text=text,
+                    words=words,
+                )
+            )
+
+        full_text = str(payload.get("full_text", "")).strip()
+        return TranscriptionResult(segments=segments, full_text=full_text)