feat(pdf): #74 au complet — fix auth 500 stream + endpoint pdf/info + HTTP Range 206 + indexation incrementale PDF + config taille/timeout

tests/test_pdf.py

@@ -124,6 +124,144 @@ class TestPdfReader:
         assert isinstance(toc, list)
 
 
+class TestPdfIncrementalIndexing:
+    """_index_single_file_sync (watcher/update path) must handle binary PDFs.
+
+    Regression: it used to read_text() every file, producing garbage for PDFs.
+    """
+
+    def test_index_single_file_sync_extracts_pdf_text(self, pdf_dir: Path):
+        from backend.indexer import _index_single_file_sync
+
+        info = _index_single_file_sync("v", str(pdf_dir), str(pdf_dir / "simple.pdf"))
+        assert info is not None
+        assert info["extension"] == ".pdf"
+        assert "ObsiGate test PDF" in info["content"]
+        assert "uniqueword0" in info["content"]
+
+    def test_index_single_file_sync_pdf_title_from_metadata(self, pdf_dir: Path):
+        from backend.indexer import _index_single_file_sync
+
+        info = _index_single_file_sync("v", str(pdf_dir), str(pdf_dir / "simple.pdf"))
+        # title metadata was set at generation time
+        assert info["title"] == "Simple Test"
+
+
+class TestPdfSizeLimit:
+    def test_oversized_pdf_skips_text_extraction(self, pdf_dir: Path, monkeypatch):
+        import backend.pdf_reader as pr
+
+        monkeypatch.setattr(pr, "PDF_MAX_SIZE_MB", 0)  # everything is "too large"
+        text = pr.extract_pdf_text(pdf_dir / "simple.pdf")
+        assert text == ""
+
+    def test_normal_pdf_within_limit_extracts(self, pdf_dir: Path, monkeypatch):
+        import backend.pdf_reader as pr
+
+        monkeypatch.setattr(pr, "PDF_MAX_SIZE_MB", 50)
+        text = pr.extract_pdf_text(pdf_dir / "simple.pdf")
+        assert "ObsiGate test PDF" in text
+
+
+# ── API: /pdf/stream (206 Range) + /pdf/info ───────────────────────────────
+
+
+@pytest.fixture
+def pdf_client():
+    """TestClient (auth disabled) over a vault containing one generated PDF."""
+    tmp = Path(tempfile.mkdtemp())
+    vault = tmp / "PdfVault"
+    vault.mkdir()
+    make_simple_pdf(vault / "doc.pdf", pages=2, title="Doc Test", author="Bruno")
+
+    os.environ["VAULT_1_NAME"] = "PdfVault"
+    os.environ["VAULT_1_PATH"] = str(vault)
+    os.environ["OBSIGATE_AUTH_ENABLED"] = "false"
+    os.environ["OBSIGATE_WATCHER_ENABLED"] = "false"
+
+    import backend.main
+    backend.main._load_config = lambda: {"watcher_enabled": False}
+
+    from backend.indexer import build_index, index
+    for key in list(index.keys()):
+        del index[key]
+
+    import asyncio
+    loop = asyncio.new_event_loop()
+    asyncio.set_event_loop(loop)
+    loop.run_until_complete(build_index())
+
+    from fastapi.testclient import TestClient
+    client = TestClient(backend.main.app)
+    yield client
+    client.close()
+    shutil.rmtree(str(tmp), ignore_errors=True)
+    for k in ["VAULT_1_NAME", "VAULT_1_PATH", "OBSIGATE_AUTH_ENABLED", "OBSIGATE_WATCHER_ENABLED"]:
+        os.environ.pop(k, None)
+
+
+class TestPdfStreamApi:
+    def test_stream_returns_200_application_pdf(self, pdf_client):
+        r = pdf_client.get("/api/file/PdfVault/pdf/stream?path=doc.pdf")
+        assert r.status_code == 200
+        assert r.headers["content-type"] == "application/pdf"
+        assert r.headers.get("accept-ranges") == "bytes"
+        assert r.content[:4] == b"%PDF"
+
+    def test_stream_full_range_returns_whole_file(self, pdf_client):
+        r = pdf_client.get("/api/file/PdfVault/pdf/stream?path=doc.pdf",
+                           headers={"Range": "bytes=0-99999999"})
+        assert r.status_code == 206
+        assert r.headers["content-range"].startswith("bytes 0-")
+        assert r.content[:4] == b"%PDF"
+
+    def test_stream_partial_range(self, pdf_client):
+        full = pdf_client.get("/api/file/PdfVault/pdf/stream?path=doc.pdf").content
+        r = pdf_client.get("/api/file/PdfVault/pdf/stream?path=doc.pdf",
+                           headers={"Range": "bytes=10-19"})
+        assert r.status_code == 206
+        assert r.content == full[10:20]
+        assert len(r.content) == 10
+
+    def test_stream_open_ended_range(self, pdf_client):
+        full = pdf_client.get("/api/file/PdfVault/pdf/stream?path=doc.pdf").content
+        r = pdf_client.get("/api/file/PdfVault/pdf/stream?path=doc.pdf",
+                           headers={"Range": "bytes=100-"})
+        assert r.status_code == 206
+        assert r.content == full[100:]
+
+    def test_stream_bad_range_returns_416(self, pdf_client):
+        r = pdf_client.get("/api/file/PdfVault/pdf/stream?path=doc.pdf",
+                           headers={"Range": "bytes=999999999-"})
+        assert r.status_code == 416
+        assert "bytes */" in r.headers.get("content-range", "")
+
+    def test_stream_rejects_non_pdf(self, pdf_client):
+        r = pdf_client.get("/api/file/PdfVault/pdf/stream?path=doc.md")
+        assert r.status_code == 404 or r.status_code == 400
+
+
+class TestPdfInfoApi:
+    def test_info_returns_metadata_without_content(self, pdf_client):
+        r = pdf_client.get("/api/file/PdfVault/pdf/info?path=doc.pdf")
+        assert r.status_code == 200
+        body = r.json()
+        assert body["pages"] == 2
+        assert body["title"] in ("Doc Test", "doc.pdf")
+        assert body["size_bytes"] > 0
+        assert "path" in body and body["vault"] == "PdfVault"
+        # no heavy content in the payload
+        assert "html" not in body
+
+    def test_info_missing_file_404(self, pdf_client):
+        r = pdf_client.get("/api/file/PdfVault/pdf/info?path=nope.pdf")
+        assert r.status_code == 404
+
+    def test_info_missing_non_pdf_404(self, pdf_client):
+        r = pdf_client.get("/api/file/PdfVault/pdf/info?path=readme.txt")
+        assert r.status_code == 404  # missing file checked before extension
+
+
 # ── backend/indexer.py ─────────────────────────────────────────────────────