refactor: strip ingest to bare minimum — PDF + story pages + one button

smoke_signal_tab.py

@@ -2815,8 +2815,8 @@ def smoke_signal_tab():
             with gr.TabItem("① Ingest", id="ss-ingest") as ingest_tab:
                 gr.HTML("""<div class="ss-panel-header" style="padding:20px 0 0 0">
                     <div class="ss-panel-icon">📥</div>
-                    <div><p class="ss-panel-title">Source Registry</p>
-                    <p class="ss-panel-sub">Drop PDF · Set pages · Click Run — everything else is automatic</p></div>
+                    <div><p class="ss-panel-title">Text Extraction</p>
+                    <p class="ss-panel-sub">Drop PDF · Define story pages · Extract clean text for author fingerprinting</p></div>
                 </div>""")
 
                 ingest_status = gr.HTML(_ingest_status_html("idle"))
@@ -2830,42 +2830,26 @@ def smoke_signal_tab():
                             type="filepath",
                         )
                     with gr.Column(scale=1):
-                        rights_dd = gr.Dropdown(
-                            label="② Rights Class",
-                            choices=["public-domain","licensed-owned","controlled-internal","unknown"],
-                            value="unknown",
-                        )
-                        book_code_input = gr.Textbox(
-                            label="Book Code (3 letters)",
-                            placeholder="e.g. grf",
+                        scope_include = gr.Textbox(
+                            label="② Story Pages (include)",
+                            placeholder="e.g. 7-27  or  7,8,9,11-27  or  all",
                             lines=1,
                         )
-                        pub_year_input = gr.Textbox(
-                            label="Publishing Year",
-                            placeholder="e.g. 1999",
+                        scope_exclude = gr.Textbox(
+                            label="Story Pages (exclude)",
+                            placeholder="e.g. 1,2,3,17,25  — title/copyright pages",
                             lines=1,
                         )
-                        ingest_notes = gr.Textbox(label="Notes", placeholder="Source, edition, etc.", lines=2)
 
-                with gr.Row():
-                    scope_include = gr.Textbox(
-                        label="③ Story Pages Include",
-                        placeholder="e.g. 7-27  or  7,8,9,11-27  or  all",
-                        scale=2,
-                    )
-                    scope_exclude = gr.Textbox(
-                        label="Story Pages Exclude",
-                        placeholder="e.g. 1,2,3,17,25",
-                        scale=2,
-                    )
-                    scope_title = gr.Textbox(
-                        label="Safe Book Title",
-                        placeholder="auto-filled from filename",
-                        scale=2,
-                    )
+                # Hidden fields — auto-populated, not shown to user
+                rights_dd       = gr.Dropdown(choices=["licensed-owned"], value="licensed-owned", visible=False)
+                book_code_input = gr.Textbox(visible=False, value="")
+                pub_year_input  = gr.Textbox(visible=False, value="")
+                ingest_notes    = gr.Textbox(visible=False, value="")
+                scope_title     = gr.Textbox(visible=False, value="")
 
                 run_pipeline_btn = gr.Button(
-                    "④ Register + OCR + Send to Review →",
+                    "③ Extract Text →",
                     elem_classes=["ss-btn-run"],
                 )
                 pipeline_done = gr.State(0)
@@ -2906,20 +2890,26 @@ def smoke_signal_tab():
                 )
 
                 # ── Auto-populate: PDF filename → safe book title ─────────────
-                def _safe_title_from_upload(files):
+                def _auto_fill_from_upload(files):
+                    """Auto-generate book code, year, and safe title from filename."""
+                    import re as _re
                     if not files:
-                        return ""
+                        return "", "", ""
                     first = files[0] if isinstance(files, list) else files
                     name = Path(first).stem if isinstance(first, str) else ""
-                    # slugify: lowercase, replace spaces/underscores with hyphens, strip non-alphanum
-                    import re as _re
                     slug = _re.sub(r"[^a-z0-9]+", "-", name.lower()).strip("-")
-                    return slug
+                    # Extract year if present in filename (e.g. "gruffalo-1999")
+                    year_match = _re.search(r"(19|20)\d{2}", name)
+                    year = year_match.group(0) if year_match else ""
+                    # Book code: first 3 alpha chars of filename
+                    alpha = _re.sub(r"[^a-z]", "", name.lower())
+                    code = alpha[:3] if len(alpha) >= 3 else alpha.ljust(3, "x")
+                    return code, year, slug
 
                 pdf_upload.change(
-                    _safe_title_from_upload,
+                    _auto_fill_from_upload,
                     inputs=[pdf_upload],
-                    outputs=[scope_title],
+                    outputs=[book_code_input, pub_year_input, scope_title],
                 )
 
                 # Main pipeline button — does everything