Update conversation_storyline/io.py

conversation_storyline/io.py

@@ -1,49 +1,97 @@
-import re
-from typing import List
-from .schemas import Interaction
-
-
-SPEAKER_PATTERNS = [
-    # Speaker A: ...
-    re.compile(r"^(?P<speaker>Speaker\s+[A-Za-z0-9_\- ]{1,64})\s*:\s*(?P<text>.+)\s*$"),
-    # A: ...
-    re.compile(r"^(?P<speaker>[A-Za-zÁÉÍÓÚÜÑáéíóúüñ0-9_\- ]{1,32})\s*:\s*(?P<text>.+)\s*$"),
-]
-
-
-def parse_transcript(text: str) -> List[Interaction]:
-    """
-    Robusto para texto pegado:
-    - Cada línea que matchee "SPEAKER: ..." crea nuevo mensaje.
-    - Líneas sin speaker se anexan al texto del último mensaje (continuación).
-    """
-    lines = [l.rstrip() for l in (text or "").splitlines()]
-    interactions: List[Interaction] = []
-    cur = None
-
-    for raw in lines:
-        line = raw.strip()
-        if not line:
-            continue
-
-        matched = None
-        for pat in SPEAKER_PATTERNS:
-            m = pat.match(line)
-            if m:
-                matched = m
-                break
-
-        if matched:
-            speaker = matched.group("speaker").strip()
-            msg = matched.group("text").strip()
-            cur = Interaction(message_id=len(interactions), speaker=speaker, text=msg)
-            interactions.append(cur)
-        else:
-            # continuation line
-            if cur is None:
-                cur = Interaction(message_id=0, speaker="Unknown", text=line)
-                interactions.append(cur)
-            else:
-                cur.text = (cur.text + " " + line).strip()
-
-    return interactions
+import re
+from typing import List
+from .schemas import Interaction
+
+
+SPEAKER_PATTERNS = [
+    # Speaker A: ...
+    re.compile(r"^(?P<speaker>Speaker\s+[A-Za-z0-9_\- ]{1,64})\s*:\s*(?P<text>.+)\s*$"),
+    # A: ...
+    re.compile(r"^(?P<speaker>[A-Za-zÁÉÍÓÚÜÑáéíóúüñ0-9_\- ]{1,32})\s*:\s*(?P<text>.+)\s*$"),
+]
+
+
+def parse_transcript(text: str) -> List[Interaction]:
+    """
+    Robusto para texto pegado:
+    - Cada línea que matchee "SPEAKER: ..." crea nuevo mensaje.
+    - Líneas sin speaker se anexan al texto del último mensaje (continuación).
+    """
+    lines = [l.rstrip() for l in (text or "").splitlines()]
+    interactions: List[Interaction] = []
+    cur = None
+
+    for raw in lines:
+        line = raw.strip()
+        if not line:
+            continue
+
+        matched = None
+        for pat in SPEAKER_PATTERNS:
+            m = pat.match(line)
+            if m:
+                matched = m
+                break
+
+        if matched:
+            speaker = matched.group("speaker").strip()
+            msg = matched.group("text").strip()
+            cur = Interaction(message_id=len(interactions), speaker=speaker, text=msg)
+            interactions.append(cur)
+        else:
+            # continuation line
+            if cur is None:
+                cur = Interaction(message_id=0, speaker="Unknown", text=line)
+                interactions.append(cur)
+            else:
+                cur.text = (cur.text + " " + line).strip()
+
+    return interactions
+def load_messages_from_text(transcript_text: str) -> List[RawMessage]:
+    """
+    Parse transcript from a raw string blob, same logic as TXT loader.
+
+    Supported formats:
+      - [00:01] Ana: texto
+      - 00:01 Ana: texto
+      - Ana: texto
+
+    Lines that don't match any pattern are treated as continuation lines.
+    """
+    lines = [ln.strip() for ln in (transcript_text or "").splitlines() if ln.strip()]
+    if not lines:
+        return []
+
+    parsed = []
+    speakers = []
+
+    for ln in lines:
+        m = None
+        for pat in LINE_PATTERNS:
+            m = pat.match(ln)
+            if m:
+                break
+        if not m:
+            if parsed:
+                parsed[-1]["content"] += "\n" + ln
+            else:
+                parsed.append({"ts": None, "speaker": "Unknown", "content": ln})
+            continue
+
+        gd = m.groupdict()
+        speaker = normalize_speaker(gd.get("speaker") or "Unknown")
+        speakers.append(speaker)
+        parsed.append({"ts": gd.get("ts"), "speaker": speaker, "content": (gd.get("content") or "").strip()})
+
+    speakers_norm, _ = normalize_speakers(speakers)
+
+    msgs: List[RawMessage] = []
+    sp_i = 0
+    for i, r in enumerate(parsed):
+        sp = r["speaker"]
+        if sp != "Unknown" and sp_i < len(speakers_norm):
+            sp = speakers_norm[sp_i]
+            sp_i += 1
+        msgs.append(RawMessage(id=i, speaker=sp, content=r["content"], timestamp=r.get("ts")))
+    return msgs
+