Replace live spinner with periodic progress prints

console.status doesn't flush to a redirected log, making long background
extract/embed/summarize runs look silent. Print periodic progress lines that
flush instead, so background jobs are observable via their output file.

Co-Authored-By: Claude Opus 4.8 (1M context) <[email protected]>

clawlibrary/__main__.py

@@ -154,20 +154,20 @@ def cmd_extract(args: argparse.Namespace) -> int:
         vlm = lambda png: oc.caption_image(png, VLM_OCR_PROMPT)  # noqa: E731
 
     done = skipped = failed = vlm_pages = 0
-    with console.status("Extracting text...") as status:
+    total = len(pdfs)
-        for i, pdf in enumerate(pdfs, 1):
+    for i, pdf in enumerate(pdfs, 1):
-            if not args.force and already_extracted(pdf):
+        if not args.force and already_extracted(pdf):
-                skipped += 1
+            skipped += 1
-                continue
+            continue
-            try:
+        try:
-                stats = extract_one(pdf, vlm=vlm)
+            stats = extract_one(pdf, vlm=vlm)
-                done += 1
+            done += 1
-                vlm_pages += stats.vlm_pages
+            vlm_pages += stats.vlm_pages
-            except Exception as e:  # noqa: BLE001
+        except Exception as e:  # noqa: BLE001
-                failed += 1
+            failed += 1
-                console.print(f"[red]✗[/red] {pdf.name}: {e}")
+            console.print(f"[red]✗[/red] {pdf.name}: {e}")
-            if i % 25 == 0:
+        if i % 100 == 0:
-                status.update(f"Extracting... {i}/{len(pdfs)} (done {done}, skipped {skipped})")
+            console.print(f"  extract {i}/{total}  (done {done}, skipped {skipped})")
     console.print(
         f"[green]Extract complete[/green]  extracted={done} skipped={skipped} "
         f"failed={failed} vlm_pages={vlm_pages}"
@@ -193,17 +193,18 @@ def cmd_summarize(args: argparse.Namespace) -> int:
         pdfs = pdfs[: args.limit]
 
     done = skipped = 0
-    with console.status("Summarizing...") as status:
+    total = len(pdfs)
-        for i, pdf in enumerate(pdfs, 1):
+    for i, pdf in enumerate(pdfs, 1):
-            side = pdf.with_suffix(".json")
+        side = pdf.with_suffix(".json")
-            if not args.force and has_summary(side):
+        if not args.force and has_summary(side):
-                skipped += 1
+            skipped += 1
-                continue
+            continue
-            summary = summarize_one(pdf, oc)
+        summary = summarize_one(pdf, oc)
-            if summary is not None:
+        if summary is not None:
-                append_insight(insights, _json.loads(side.read_text()))
+            append_insight(insights, _json.loads(side.read_text()))
-                done += 1
+            done += 1
-            status.update(f"Summarizing... {i}/{len(pdfs)} (done {done})")
+        if i % 10 == 0:
+            console.print(f"  summarize {i}/{total}  (done {done}, skipped {skipped})")
     console.print(f"[green]Summaries complete[/green]  done={done} skipped={skipped}")
     console.print(f"Insight digest: {insights}")
     return 0
@@ -228,24 +229,24 @@ def cmd_embed(args: argparse.Namespace) -> int:
         pdfs = pdfs[: args.limit]
 
     added = skipped = 0
-    with console.status("Embedding...") as status:
+    total = len(pdfs)
-        for i, pdf in enumerate(pdfs, 1):
+    for i, pdf in enumerate(pdfs, 1):
-            data = json.loads(pdf.with_suffix(".json").read_text())
+        data = json.loads(pdf.with_suffix(".json").read_text())
-            doi = data.get("doi") or data.get("id") or pdf.stem
+        doi = data.get("doi") or data.get("id") or pdf.stem
-            if doi in seen:
+        if doi in seen:
-                skipped += 1
+            skipped += 1
-                continue
+            continue
-            text = pdf.with_suffix(".txt").read_text()
+        text = pdf.with_suffix(".txt").read_text()
-            vecs, rows = add_document(
+        vecs, rows = add_document(
-                oc, text=text, doi=doi, title=data.get("title", ""),
+            oc, text=text, doi=doi, title=data.get("title", ""),
-                topic=data.get("topic", ""), text_path=str(pdf.with_suffix(".txt")),
+            topic=data.get("topic", ""), text_path=str(pdf.with_suffix(".txt")),
-            )
+        )
-            store = merge(store, vecs, rows)
+        store = merge(store, vecs, rows)
-            seen.add(doi)
+        seen.add(doi)
-            added += 1
+        added += 1
-            if i % 20 == 0:
+        if i % 20 == 0:
-                store.save(root)  # periodic checkpoint for resumability
+            store.save(root)  # periodic checkpoint for resumability
-                status.update(f"Embedding... {i}/{len(pdfs)} (added {added})")
+            console.print(f"  embed {i}/{total}  (added {added}, vectors {store.vectors.shape[0]})")
     store.save(root)
     console.print(
         f"[green]Embedding complete[/green]  papers_added={added} skipped={skipped} "

embed.out

@@ -0,0 +1 @@
+Embedding complete  papers_added=1199 skipped=3 vectors=40978