feat: troncature annoncee et lecture par fenetres des tableurs #153
CI / lint (push) Successful in 2m29s
CI / security (push) Failing after 1m39s
CI / test (push) Successful in 4m1s
CI / build (push) Successful in 1m37s
CI / e2e (push) Successful in 14m47s

- BUG-090 (#153 A8) : render_sheets() expose total_rows/total_cols,
  max_rows/max_cols et truncated ; la visionneuse affiche un bandeau
  « Feuille tronquée » (i18n FR/EN) au lieu de couper en silence, et la
  ligne d'en-têtes devient sticky (top:auto sur les numéros de ligne).
- #153 A9 : GET /api/file/{vault}/xlsx/sheet?sheet&offset&limit sert une
  fenêtre de 1 à 1000 lignes avec les vraies coordonnées A1, has_more de
  pagination et valeurs calculées A12 ; 404 feuille inconnue, 415 non-xlsx.
- Tests : TestXlsxTruncationNotice (4) + TestXlsxSheetWindow (11) avec
  contre-preuves, xlsx-viewer.test.mjs 14/14, E2E 7/7 (fixture
  sample-xlsx-large.xlsx 520 lignes), suite 1417 passed / 6 skipped,
  ruff/mypy 0, i18n parity.

🤖 Generated with Codebuff
Co-Authored-By: Codebuff <[email protected]>
This commit is contained in:
2026-09-28 10:38:24 -04:00
parent 06f8e63d06
commit 6ba04c4381
24 changed files with 726 additions and 47 deletions
+167
View File
@@ -519,6 +519,173 @@ class TestXlsxWriteLock:
assert openpyxl.load_workbook(xlsx_file)["Budget"]["A1"].value == "premier"
@pytest.fixture
def wide_xlsx(test_vault_dir: str) -> str:
"""Workbook whose sheet exceeds BOTH render caps (501 rows x 45 cols).
Sparse on purpose: a cell in A501 and one in AS1 are enough for openpyxl
to declare those dimensions, without writing 20 000 cells to disk.
"""
from openpyxl import Workbook
path = Path(test_vault_dir) / "grand.xlsx"
wb = Workbook()
ws = wb.active
ws.title = "Data"
ws["A1"] = "tête"
ws["A501"] = "dernière ligne"
ws["AS1"] = "colonne 45"
wb.save(path)
return str(path)
@pytest.fixture
def edge_xlsx(test_vault_dir: str) -> str:
"""Sheet exactly on the caps (500 rows x 40 cols) — must NOT be truncated."""
from openpyxl import Workbook
path = Path(test_vault_dir) / "limite.xlsx"
wb = Workbook()
ws = wb.active
ws.title = "Data"
ws["A1"] = "bord"
ws["A500"] = "ligne 500"
ws["AN1"] = "colonne 40"
wb.save(path)
return str(path)
# ── #153 A8 — silent truncation made visible ─────────────────────────────
class TestXlsxTruncationNotice:
"""A sheet bigger than the caps must SAY so instead of looking complete."""
def test_render_reports_the_real_dimensions(self, client, wide_xlsx):
resp = client.get(f"/api/file/{VAULT}", params={"path": "grand.xlsx"})
sheet = resp.json()["xlsx_sheets"][0]
assert (sheet["total_rows"], sheet["total_cols"]) == (501, 45)
assert sheet["truncated"] is True
# The caps are the coverage the banner announces — `rows`/`cols` are
# post-trim and would understate it on a sparse sheet.
assert (sheet["max_rows"], sheet["max_cols"]) == (500, 40)
assert (sheet["rows"], sheet["cols"]) == (1, 1) # only 3 filled cells
def test_a_sheet_on_the_caps_is_not_flagged(self, client, edge_xlsx):
"""Boundary: 500x40 is exactly what the renderer supports."""
resp = client.get(f"/api/file/{VAULT}", params={"path": "limite.xlsx"})
sheet = resp.json()["xlsx_sheets"][0]
assert sheet["truncated"] is False
assert (sheet["total_rows"], sheet["total_cols"]) == (500, 40)
def test_a_normal_sheet_is_not_flagged(self, client, xlsx_file):
resp = client.get(f"/api/file/{VAULT}", params={"path": "budget.xlsx"})
assert all(not s["truncated"] for s in resp.json()["xlsx_sheets"])
def test_blank_tail_is_not_reported_as_truncation(self, client, test_vault_dir):
"""A sheet with empty rows below its data fits in the caps."""
from openpyxl import Workbook
path = Path(test_vault_dir) / "blanc.xlsx"
wb = Workbook()
ws = wb.active
ws.title = "Data"
ws["A1"] = "seule ligne"
ws["A300"] = None # formatted-but-empty row inside the caps
wb.save(path)
resp = client.get(f"/api/file/{VAULT}", params={"path": "blanc.xlsx"})
sheet = resp.json()["xlsx_sheets"][0]
assert sheet["truncated"] is False
assert sheet["rows"] == 1 # trailing blanks dropped by _trim
# ── #153 A9 — lazy per-sheet loading ─────────────────────────────────────
class TestXlsxSheetWindow:
"""GET /api/file/{vault}/xlsx/sheet — one window of one sheet."""
def _get(self, client, path="budget.xlsx", **params):
return client.get(
f"/api/file/{VAULT}/xlsx/sheet", params={"path": path, **params}
)
def test_window_returns_rows_and_totals(self, client, xlsx_file):
resp = self._get(client, sheet="Budget", offset=0, limit=10)
assert resp.status_code == 200
data = resp.json()
assert data["sheet"] == "Budget"
assert (data["offset"], data["limit"]) == (0, 10)
assert data["rows"] == 2 and data["cols"] == 2
assert data["total_rows"] == 2 and data["truncated"] is False
assert (data["max_rows"], data["max_cols"]) == (500, 40)
assert data["has_more"] is False
assert 'data-cell="A1"' in data["html"]
def test_window_keeps_the_real_a1_coordinates(self, client, xlsx_file):
"""A window must be indistinguishable from a full render: the A1
references and the row numbers have to be the sheet's, not the
window's, or an edit would land on the wrong cell. The CONTENT matters
as much as the label — row 2 of "Budget" is "Total", not "Poste"."""
data = self._get(client, sheet="Budget", offset=1, limit=1).json()
assert data["rows"] == 1
assert 'data-cell="A2"' in data["html"]
assert 'data-cell="A1"' not in data["html"]
assert "<th class=\"xlsx-rownum\">2</th>" in data["html"]
assert "Total" in data["html"] and "Poste" not in data["html"]
def test_window_offsets_walk_the_whole_sheet(self, client, wide_xlsx):
first = self._get(client, path="grand.xlsx", sheet="Data", offset=0, limit=10).json()
last = self._get(client, path="grand.xlsx", sheet="Data", offset=500, limit=10).json()
assert first["truncated"] is True and first["has_more"] is True
assert 'data-cell="A501"' in last["html"] # the row the caps used to hide
assert "dernière ligne" in last["html"]
assert "dernière ligne" not in first["html"]
assert last["rows"] == 1 and last["has_more"] is False
def test_offset_past_the_end_is_empty_not_an_error(self, client, xlsx_file):
resp = self._get(client, sheet="Budget", offset=9999, limit=10)
assert resp.status_code == 200
data = resp.json()
assert data["rows"] == 0
assert data["has_more"] is False
def test_cached_result_survives_the_window(self, client, lossy_xlsx):
"""A12 must not be lost on the lazy path."""
data = self._get(client, path="risky.xlsx", sheet="Data", offset=0, limit=10).json()
assert "xlsx-cached" in data["html"]
def test_unknown_sheet_is_404(self, client, xlsx_file):
resp = self._get(client, sheet="Nope")
assert resp.status_code == 404
assert "Feuille introuvable" in resp.json()["detail"]
def test_missing_file_is_404(self, client, test_vault_dir):
assert self._get(client, path="absent.xlsx", sheet="Budget").status_code == 404
def test_non_xlsx_file_is_415(self, client, test_vault_dir):
(Path(test_vault_dir) / "note.md").write_text("# hi", encoding="utf-8")
resp = self._get(client, path="note.md", sheet="Budget")
assert resp.status_code == 415
def test_limit_above_the_server_cap_is_rejected(self, client, xlsx_file):
"""The cap is a contract, not a silent truncation of the request."""
assert self._get(client, sheet="Budget", limit=100_000).status_code == 422
def test_reader_clamps_a_hostile_limit(self, xlsx_file):
"""Belt and braces: the reader caps too, whoever calls it."""
from backend.xlsx_reader import MAX_WINDOW_ROWS, read_sheet_window
window = read_sheet_window(Path(xlsx_file), "Budget", offset=0, limit=10**9)
assert window["limit"] == MAX_WINDOW_ROWS
def test_reader_rejects_a_negative_offset(self, xlsx_file):
from backend.xlsx_reader import read_sheet_window
window = read_sheet_window(Path(xlsx_file), "Budget", offset=-5, limit=10)
assert window["offset"] == 0
# ── #153 A4 — formula injection ──────────────────────────────────────────