fix: support DOCX manuscript uploads in dashboard and codex extractor

app.py

@@ -2597,7 +2597,7 @@ def run_codex_extraction(
     """
     if file_obj is None:
         return (
-            "No file uploaded. Please upload a .txt or .pdf file.",
+            "No file uploaded. Please upload a .txt, .docx, or .pdf file.",
             "",
             "ERROR: No file uploaded.",
         )
@@ -2814,7 +2814,7 @@ def analyze_manuscript_for_dashboard(file_obj, active_path: str):
     """
     file_path = _extract_uploaded_path(file_obj)
     if not file_path:
-        raise gr.Error("Upload a manuscript (.txt or .pdf) first.")
+        raise gr.Error("Upload a manuscript (.txt, .docx, or .pdf) first.")
 
     stem = Path(file_path).stem.strip() or "Unknown Author"
     author_name = stem.replace("_", " ")[:120]
@@ -2922,8 +2922,8 @@ with gr.Blocks(title="TOTEM Studio") as demo:
                         "Upload manuscript here to run extraction and refresh workbook scoring context."
                     )
                     dashboard_manuscript_file = gr.File(
-                        label="Manuscript (.txt or .pdf)",
-                        file_types=[".txt", ".pdf"],
+                        label="Manuscript (.txt, .docx, or .pdf)",
+                        file_types=[".txt", ".docx", ".pdf"],
                         type="filepath",
                     )
                     dashboard_manuscript_analyze = gr.Button(
@@ -3015,7 +3015,7 @@ with gr.Blocks(title="TOTEM Studio") as demo:
                     )
                     codex_file = gr.File(
                         label="Upload Text or PDF",
-                        file_types=[".txt", ".pdf"],
+                        file_types=[".txt", ".docx", ".pdf"],
                         type="filepath",
                     )
                     with gr.Row():

src/codex_extractor.py

@@ -87,6 +87,7 @@ import math
 import os
 import re
 import string
+import zipfile
 from collections import Counter
 from pathlib import Path
 from shutil import which
@@ -473,14 +474,33 @@ def extract_text_from_file(
     end_page: int | None = None,
 ) -> tuple[str, str, list[dict]]:
     """
-    Extract text from PDF or plain text file.
+    Extract text from PDF, DOCX, or plain text file.
 
     Returns (story_text, raw_text, page_trace).
     For plain text files page_trace contains a single synthetic entry.
     """
     path = Path(file_path)
-    if path.suffix.lower() == ".pdf":
+    suffix = path.suffix.lower()
+    if suffix == ".pdf":
         return extract_text_from_pdf(path, start_page=start_page, end_page=end_page)
+    if suffix == ".docx":
+        with zipfile.ZipFile(path) as zf:
+            xml = zf.read("word/document.xml").decode("utf-8", errors="replace")
+        xml = re.sub(r"<w:tab\\s*/>", "\t", xml)
+        xml = re.sub(r"</w:p>", "\n", xml)
+        content = re.sub(r"<[^>]+>", "", xml)
+        content = content.replace("&amp;", "&").replace("&lt;", "<").replace("&gt;", ">")
+        content = re.sub(r"\n{3,}", "\n\n", content).strip()
+        wc = _word_count(content)
+        page_trace = [{
+            "page_number": 1,
+            "raw_word_count": wc,
+            "cleaned_word_count": wc,
+            "classification": "story",
+            "included": True,
+            "skip_reason": "",
+        }]
+        return content, content, page_trace
     else:
         content = path.read_text(encoding="utf-8", errors="replace")
         wc = _word_count(content)
@@ -1805,7 +1825,7 @@ def process_upload(
     Entry point for Gradio interface.
 
     Args:
-        file_path:           Path to uploaded PDF or text file.
+        file_path:           Path to uploaded PDF, DOCX, or text file.
         author_name:         Author's full name.
         author_id:           Codex author ID.
         works_sampled:       Title of the uploaded work (user-supplied only;
@@ -1913,4 +1933,4 @@ if __name__ == "__main__":
         debug_output_dir="/tmp/codex_debug",
     )
     print(format_fingerprint_report(fp))
-    print("\nDebug artefacts written to:", fp.get("Debug_Artefacts", {}))
+    print("\nDebug artefacts written to:", fp.get("Debug_Artefacts", {}))