feat(api): PDF + DOCX extraction in folder indexer

Add pypdf/python-docx deps, _extract_pdf_text/_extract_docx_text helpers,
and summarize_pdf/summarize_docx wrappers that delegate to summarize_text.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

tests/test_folder_indexer.py

@@ -51,3 +51,33 @@ async def test_summarize_image_uses_vision_content_blocks():
         isinstance(p, dict) and p.get("type") == "image_url"
         for p in (last.content if isinstance(last.content, list) else [])
     )
+
+
+async def test_summarize_pdf_extracts_then_summarizes(monkeypatch):
+    # pypdf.PdfReader returns text from pages
+    from app.core import folder_indexer
+    class FakePage:
+        def extract_text(self): return "PDF page content with project info."
+    class FakeReader:
+        pages = [FakePage(), FakePage()]
+    monkeypatch.setattr(folder_indexer, "PdfReader", lambda buf: FakeReader())
+    mock_resp = AsyncMock(); mock_resp.content = "Project info doc."; mock_resp.usage_metadata = {"total_tokens": 50}
+    async def fake_llm(messages): return mock_resp
+    with patch("app.core.folder_indexer._llm_text", new=fake_llm):
+        result = await folder_indexer.summarize_pdf(pdf_b64="SGVsbG8=", name="doc.pdf")
+    assert "Project info" in result.summary
+    assert result.tokens_used == 50
+
+
+async def test_summarize_docx_extracts_then_summarizes(monkeypatch):
+    from app.core import folder_indexer
+    class FakePara:
+        def __init__(self, t): self.text = t
+    class FakeDoc:
+        paragraphs = [FakePara("Heading"), FakePara("Body paragraph one.")]
+    monkeypatch.setattr(folder_indexer, "DocxDocument", lambda buf: FakeDoc())
+    mock_resp = AsyncMock(); mock_resp.content = "Heading and body."; mock_resp.usage_metadata = {"total_tokens": 30}
+    async def fake_llm(messages): return mock_resp
+    with patch("app.core.folder_indexer._llm_text", new=fake_llm):
+        result = await folder_indexer.summarize_docx(docx_b64="UEsDBBQ=", name="doc.docx")
+    assert result.summary == "Heading and body."