feat: report indexing progress with page counts in status panel

backend/app/indexer.py

@@ -38,6 +38,9 @@ class IndexStatus:
     model: str = ""
     last_indexed_at: str | None = None
     errors: list[str] = field(default_factory=list)
+    pages_processed: int = 0
+    pages_total: int = 0
+    current_file: str | None = None
 
 
 @dataclass
@@ -78,6 +81,10 @@ class Indexer:
         self._lock = threading.Lock()
         self._last_indexed_at: str | None = None
         self._errors: list[str] = []
+        self._indexing = False
+        self._pages_processed = 0
+        self._pages_total = 0
+        self._current_file: str | None = None
         self._load_manifest()
 
     # ------------------------------------------------------------------ #
@@ -154,6 +161,21 @@ class Indexer:
     # Indexação
     # ------------------------------------------------------------------ #
 
+    def mark_indexing(self) -> None:
+        """Sinaliza que uma indexação está em andamento."""
+        self._indexing = True
+        self._pages_processed = 0
+        self._pages_total = 0
+        self._current_file = None
+
+    def report_progress(
+        self, pdf_name: str, pages_done: int, pages_total: int
+    ) -> None:
+        """Registra o progresso da extração de um arquivo."""
+        self._current_file = pdf_name
+        self._pages_processed = pages_done
+        self._pages_total = pages_total
+
     def discover_pdfs(self) -> list[Path]:
         """Lista os PDFs do diretório configurado, ordenados por nome."""
         directory = self.settings.pdf_dir
@@ -165,20 +187,25 @@ class Indexer:
         """Indexa os PDFs novos ou alterados."""
         with self._lock:
             self._errors = []
+            self.mark_indexing()
             pdfs = self.discover_pdfs()
 
-            for pdf in pdfs:
-                try:
-                    self._index_file(pdf, force=force)
-                except Exception as exc:  # pragma: no cover - falha inesperada
-                    logger.exception("Falha ao indexar %s", pdf.name)
-                    self._manifest[pdf.name] = _ManifestEntry(
-                        sha256="", state=STATE_ERROR, indexed_at=_now()
-                    )
-                    self._errors.append(f"{pdf.name}: {exc}")
+            try:
+                for pdf in pdfs:
+                    try:
+                        self._index_file(pdf, force=force)
+                    except Exception as exc:  # pragma: no cover - falha inesperada
+                        logger.exception("Falha ao indexar %s", pdf.name)
+                        self._manifest[pdf.name] = _ManifestEntry(
+                            sha256="", state=STATE_ERROR, indexed_at=_now()
+                        )
+                        self._errors.append(f"{pdf.name}: {exc}")
+
+                self._save_manifest()
+                self._last_indexed_at = _now()
+            finally:
+                self._indexing = False
 
-            self._save_manifest()
-            self._last_indexed_at = _now()
             return self.status()
 
     def _index_file(self, pdf: Path, force: bool) -> None:
@@ -190,7 +217,15 @@ class Indexer:
             return
 
         logger.info("Indexando %s", pdf.name)
-        pages = list(self.extractor.extract_document(pdf, file_hash=file_hash))
+        pages = []
+        page_count = _pdf_page_count(pdf)
+        for position, page in enumerate(
+            self.extractor.extract_document(pdf, file_hash=file_hash), start=1
+        ):
+            pages.append(page)
+            if position % 10 == 0 or position == page_count:
+                self.report_progress(pdf.name, position, page_count)
+
         failed_pages = [page.page for page in pages if page.failed]
 
         book_starts = self._book_starts(pages)
@@ -229,16 +264,36 @@ class Indexer:
             for entry in self._manifest.values()
             if entry.state in (STATE_COMPLETE, STATE_PARTIAL)
         )
+        if self._indexing:
+            state = "indexing"
+        elif self._errors and indexed == 0:
+            state = "error"
+        else:
+            state = "ready"
         return IndexStatus(
-            state="error" if self._errors and indexed == 0 else "ready",
+            state=state,
             pdfs_total=len(self.discover_pdfs()),
             pdfs_indexed=indexed,
             chunks=self.store.count(),
             model=self.settings.embedding_model,
             last_indexed_at=self._last_indexed_at,
             errors=list(self._errors),
+            pages_processed=self._pages_processed,
+            pages_total=self._pages_total,
+            current_file=self._current_file,
         )
 
 
+def _pdf_page_count(pdf: Path) -> int:
+    """Número de páginas do PDF, ou 0 se não for possível determinar."""
+    try:
+        import pymupdf
+
+        with pymupdf.open(pdf) as document:
+            return len(document)
+    except Exception:  # pragma: no cover - PDF malformado
+        return 0
+
+
 def _now() -> str:
     return datetime.now(timezone.utc).isoformat()