Fix speaker ID regex and page break/line spacing in transcript DOCX

scraibe/email_sender.py

@@ -457,7 +457,7 @@ def create_transcript_docx(text: str, filename: str, summary_text: str = ""):
     """
     from . import docx_styles
 
-    # Step 1: Prepare transcript into pages of 29 lines each
+    # Step 1: Prepare transcript into pages of 30 lines each
     prepared_pages = []
     current_page = []
     line_count = 0
@@ -485,7 +485,7 @@ def create_transcript_docx(text: str, filename: str, summary_text: str = ""):
         if current:
             segments.append(current)
 
-        # Add segments to pages, enforcing 29 lines per page
+        # Add segments to pages, enforcing 30 lines per page
         for seg in segments:
             if line_count == 30:
                 prepared_pages.append(current_page)
@@ -523,21 +523,41 @@ def create_transcript_docx(text: str, filename: str, summary_text: str = ""):
 
     # Step 4: Write prepared pages into DOCX
     for page_idx, page_lines in enumerate(prepared_pages):
-        # Insert page break between pages
+        # Write each line with its number (1-30)
-        if page_idx > 0:
+        for line_num, line_text in enumerate(page_lines, start=1):
+            p = doc.add_paragraph()
+            _add_transcript_paragraph(doc, line_text, line_number=line_num)
+            # Remove the extra paragraph added by add_paragraph (we already added runs)
+            # _add_transcript_paragraph already creates its own paragraph, so we need to avoid duplication.
+            # Correct approach: _add_transcript_paragraph should add to doc, not p.
+            # To avoid breaking existing behavior, keep _add_transcript_paragraph as-is
+            # and remove the temporary paragraph we just added.
+            # This is an internal fix to ensure page break logic is clean.
+            # We'll rely on _add_transcript_paragraph creating its own paragraph.
+            # To avoid an extra blank paragraph, we remove p if it has no runs.
+            if not p.runs:
+                body.remove(p._p)
+
+        # After each page except the last, add a page break via next paragraph
+        if page_idx < len(prepared_pages) - 1:
             p_break = doc.add_paragraph()
+            p_break.paragraph_format.page_break_before = True
+            # Ensure it’s empty and doesn’t show as a visible line
+            if p_break.runs:
+                for r in p_break.runs:
+                    r.text = ""
+            # Remove any formatting that would show as a visible line
             pPr = p_break._p.get_or_add_pPr()
             for child in list(pPr):
                 tag = child.tag.split("}")[-1] if "}" in child.tag else child.tag
                 if tag in ("tabs", "spacing", "ind"):
                     pPr.remove(child)
-            page_break = OxmlElement("w:pageBreak")
+            spacing = OxmlElement("w:spacing")
-            page_break.set("{http://schemas.openxmlformats.org/wordprocessingml/2006/main}val", "1")
+            _set_element_attr(spacing, "before", "0")
-            pPr.append(page_break)
+            _set_element_attr(spacing, "after", "0")
-
+            _set_element_attr(spacing, "line", "240")
-        # Write each line with its number (1-29)
+            _set_element_attr(spacing, "lineRule", "auto")
-        for line_num, line_text in enumerate(page_lines, start=1):
+            pPr.append(spacing)
-            _add_transcript_paragraph(doc, line_text, line_number=line_num)
 
     # Step 5: If summary_text provided, append it with a hard page break
     if summary_text and summary_text.strip():

scraibe/tasks.py

@@ -322,8 +322,8 @@ def _apply_speaker_map(transcript_text: str, segments: list, speaker_map: dict)
 
     # Replace in transcript lines
     def replace_in_line(line: str) -> str:
-        # Pattern: timestamp bracket + speaker label + colon
+        # Match: [MM:SS] or [HH:MM:SS] then speaker label then colon
-        # More flexible: allow any non-empty label before the colon
+        # Use a lazy match for the label up to the first colon
         def repl(m):
             prefix = m.group(1)  # e.g. "[00:12] "
             label = m.group(2).strip()
@@ -331,9 +331,10 @@ def _apply_speaker_map(transcript_text: str, segments: list, speaker_map: dict)
             new_label = speaker_map.get(normalized, label)
             return f"{prefix}{new_label}: "
         return re.sub(
-            r"(\[\d+:\d+(?::\d+)?\]\s*)([^\]]+?):\s*",
+            r"(\[\d+:\d+(?::\d+)?\]\s*)(.+?):\s*",
             repl,
             line,
+            count=1,
         )
 
     updated_transcript = "\n".join(