wehub-resource-sync
09e9f3545f
Test / Code Quality (push) Has been cancelled
Test / Test (macos-latest, Python 3.10) (push) Has been cancelled
Test / Test (macos-latest, Python 3.11) (push) Has been cancelled
Test / Test (macos-latest, Python 3.12) (push) Has been cancelled
Test / Test (macos-latest, Python 3.13) (push) Has been cancelled
Test / Test (macos-latest, Python 3.14) (push) Has been cancelled
Test / Test (ubuntu-latest, Python 3.10) (push) Has been cancelled
Test / Test (ubuntu-latest, Python 3.11) (push) Has been cancelled
Test / Test (ubuntu-latest, Python 3.12) (push) Has been cancelled
Test / Test (ubuntu-latest, Python 3.13) (push) Has been cancelled
Test / Test (ubuntu-latest, Python 3.14) (push) Has been cancelled
Test / Test (windows-latest, Python 3.10) (push) Has been cancelled
Test / Test (windows-latest, Python 3.11) (push) Has been cancelled
Test / Test (windows-latest, Python 3.12) (push) Has been cancelled
Test / Test (windows-latest, Python 3.13) (push) Has been cancelled
Test / Test (windows-latest, Python 3.14) (push) Has been cancelled
CodeQL / Analyze (push) Has been cancelled
dependency-audit / pip-audit (push) Has been cancelled
222 lines
9.1 KiB
Python
222 lines
9.1 KiB
Python
"""Tests for the WIZ_global_data token-extraction regex helper.
|
|
|
|
Covers ``extract_wiz_field`` (the unified helper introduced in PR) and
|
|
verifies that all extraction variants observed on real NotebookLM responses
|
|
parse correctly, while drift produces a typed diagnostic
|
|
(:class:`AuthExtractionError`) with a sanitized HTML preview.
|
|
"""
|
|
|
|
from __future__ import annotations
|
|
|
|
from pathlib import Path
|
|
|
|
import pytest
|
|
|
|
from notebooklm._auth.extraction import extract_wiz_field
|
|
from notebooklm.exceptions import AuthExtractionError, RPCError
|
|
|
|
FIXTURE = Path(__file__).parent.parent / "fixtures" / "notebooklm_html.html"
|
|
|
|
|
|
# ---------------------------------------------------------------------------
|
|
# Fixture-backed happy paths
|
|
# ---------------------------------------------------------------------------
|
|
|
|
|
|
@pytest.fixture(scope="module")
|
|
def real_shape_html() -> str:
|
|
"""Load the sanitized real-shape NotebookLM HTML fixture."""
|
|
return FIXTURE.read_text(encoding="utf-8")
|
|
|
|
|
|
def test_extracts_csrf_token_from_fixture(real_shape_html: str) -> None:
|
|
"""SNlM0e is extracted from a real-shape sanitized response."""
|
|
value = extract_wiz_field(real_shape_html, "SNlM0e")
|
|
assert value == "SCRUBBED_CSRF_AF1_QpN-abc123_def456"
|
|
|
|
|
|
def test_extracts_session_id_from_fixture(real_shape_html: str) -> None:
|
|
"""FdrFJe is extracted from the same fixture."""
|
|
value = extract_wiz_field(real_shape_html, "FdrFJe")
|
|
assert value == "1234567890123456"
|
|
|
|
|
|
def test_extracts_account_email_field_from_fixture(real_shape_html: str) -> None:
|
|
"""The helper is generic — it locates any WIZ_global_data field."""
|
|
value = extract_wiz_field(real_shape_html, "oM1Kwf")
|
|
assert value == "user@example.com"
|
|
|
|
|
|
# ---------------------------------------------------------------------------
|
|
# Variant matrix — single-quoted, HTML-escaped, whitespace, empty value
|
|
# ---------------------------------------------------------------------------
|
|
|
|
|
|
def test_single_quoted_variant_is_supported() -> None:
|
|
"""``'SNlM0e':'value'`` is parsed (observed in some debug renders)."""
|
|
html = "<script>window.WIZ_global_data = {'SNlM0e':'tok_single_quote'};</script>"
|
|
assert extract_wiz_field(html, "SNlM0e") == "tok_single_quote"
|
|
|
|
|
|
def test_html_escaped_variant_is_supported() -> None:
|
|
"""``"key":"value"`` works when the script block
|
|
is rendered inside an attribute or escaped fragment."""
|
|
html = '<div data-init="{"SNlM0e":"tok_escaped"}"></div>'
|
|
assert extract_wiz_field(html, "SNlM0e") == "tok_escaped"
|
|
|
|
|
|
def test_whitespace_around_colon_is_tolerated() -> None:
|
|
"""Real responses sometimes have a space after the colon for readability."""
|
|
html = '{"SNlM0e" : "tok_spaced"}'
|
|
assert extract_wiz_field(html, "SNlM0e") == "tok_spaced"
|
|
|
|
|
|
def test_empty_value_returns_empty_string_not_none() -> None:
|
|
"""``"SNlM0e":""`` is a legitimate empty token — not drift."""
|
|
html = '{"SNlM0e":"","FdrFJe":"abc"}'
|
|
assert extract_wiz_field(html, "SNlM0e") == ""
|
|
|
|
|
|
def test_canonical_pattern_takes_priority_over_single_quoted() -> None:
|
|
"""When both variants appear, the canonical double-quoted form wins.
|
|
|
|
This guarantees deterministic extraction on pages that include both a
|
|
canonical embedding and a stray single-quoted reference (e.g. inline
|
|
JavaScript debug logs)."""
|
|
html = "'SNlM0e':'wrong'... \"SNlM0e\":\"right\""
|
|
assert extract_wiz_field(html, "SNlM0e") == "right"
|
|
|
|
|
|
# ---------------------------------------------------------------------------
|
|
# Negative paths — drift produces AuthExtractionError with preview
|
|
# ---------------------------------------------------------------------------
|
|
|
|
|
|
def test_missing_key_raises_in_strict_mode() -> None:
|
|
"""Strict drift path raises AuthExtractionError, not generic ValueError."""
|
|
with pytest.raises(AuthExtractionError) as excinfo:
|
|
extract_wiz_field("<html></html>", "SNlM0e", strict=True)
|
|
assert excinfo.value.key == "SNlM0e"
|
|
|
|
|
|
def test_strict_mode_is_default() -> None:
|
|
"""``strict`` defaults to True so missing keys cannot pass silently."""
|
|
with pytest.raises(AuthExtractionError):
|
|
extract_wiz_field("<html></html>", "SNlM0e")
|
|
|
|
|
|
def test_non_strict_returns_none_on_drift() -> None:
|
|
"""``strict=False`` returns None so callers can fall back."""
|
|
assert extract_wiz_field("<html></html>", "SNlM0e", strict=False) is None
|
|
|
|
|
|
def test_non_strict_returns_value_when_present() -> None:
|
|
"""``strict=False`` still returns the value when the field exists."""
|
|
html = '{"SNlM0e":"tok"}'
|
|
assert extract_wiz_field(html, "SNlM0e", strict=False) == "tok"
|
|
|
|
|
|
# ---------------------------------------------------------------------------
|
|
# Preview / diagnostic behavior
|
|
# ---------------------------------------------------------------------------
|
|
|
|
|
|
def test_preview_included_in_error_message() -> None:
|
|
"""The error's ``__str__`` includes the preview so it surfaces in logs."""
|
|
payload = "this is the response body that should appear in the preview"
|
|
with pytest.raises(AuthExtractionError) as excinfo:
|
|
extract_wiz_field(payload, "SNlM0e", strict=True)
|
|
rendered = str(excinfo.value)
|
|
assert "this is the response body" in rendered
|
|
assert "SNlM0e" in rendered
|
|
|
|
|
|
def test_preview_is_truncated_to_200_chars() -> None:
|
|
"""Long payloads are truncated to keep error messages compact."""
|
|
payload = "A" * 5000
|
|
with pytest.raises(AuthExtractionError) as excinfo:
|
|
extract_wiz_field(payload, "SNlM0e", strict=True)
|
|
# The preview attribute itself caps at 200; sanity-check both that bound
|
|
# and that the rendered string never carries the full giant payload.
|
|
assert len(excinfo.value.payload_preview) <= AuthExtractionError.PREVIEW_LENGTH
|
|
assert len(str(excinfo.value)) < 1000
|
|
|
|
|
|
def test_preview_collapses_whitespace() -> None:
|
|
"""Newlines and runs of whitespace are collapsed to a single space so
|
|
the preview is readable when surfaced through ``str(exc)``."""
|
|
payload = "alpha\n\n\n\t\tbeta gamma"
|
|
with pytest.raises(AuthExtractionError) as excinfo:
|
|
extract_wiz_field(payload, "SNlM0e", strict=True)
|
|
assert "alpha beta gamma" in excinfo.value.payload_preview
|
|
|
|
|
|
def test_auth_extraction_error_is_rpc_error() -> None:
|
|
"""The exception subclasses RPCError so existing handlers cover it."""
|
|
with pytest.raises(RPCError):
|
|
extract_wiz_field("", "SNlM0e", strict=True)
|
|
|
|
|
|
def test_key_attribute_records_failing_field() -> None:
|
|
"""``.key`` lets callers route on which field drifted."""
|
|
with pytest.raises(AuthExtractionError) as excinfo:
|
|
extract_wiz_field("", "FdrFJe", strict=True)
|
|
assert excinfo.value.key == "FdrFJe"
|
|
|
|
|
|
# ---------------------------------------------------------------------------
|
|
# Key regex escaping — defensive
|
|
# ---------------------------------------------------------------------------
|
|
|
|
|
|
def test_key_with_regex_metacharacters_is_escaped() -> None:
|
|
"""Field names containing regex metacharacters are escaped, so the helper
|
|
matches them literally rather than treating them as regex syntax."""
|
|
# ``.`` would otherwise match any character — confirm it does not.
|
|
html = '{"a.b":"literal","aXb":"wildcard_would_match_this"}'
|
|
assert extract_wiz_field(html, "a.b") == "literal"
|
|
|
|
|
|
# ---------------------------------------------------------------------------
|
|
# Escape handling — JSON-style quote escapes inside values
|
|
# ---------------------------------------------------------------------------
|
|
|
|
|
|
def test_escaped_quote_inside_double_quoted_value() -> None:
|
|
"""JSON-style ``\\"`` inside the value is captured intact rather than
|
|
terminating the match early. Without escape-aware matching, the value
|
|
``a\\"b`` would be truncated to just ``a``."""
|
|
html = r'{"SNlM0e":"a\"b"}'
|
|
assert extract_wiz_field(html, "SNlM0e") == r"a\"b"
|
|
|
|
|
|
def test_escaped_quote_inside_single_quoted_value() -> None:
|
|
"""Same escape-awareness applies to the single-quoted variant."""
|
|
html = r"{'SNlM0e':'a\'b'}"
|
|
assert extract_wiz_field(html, "SNlM0e") == r"a\'b"
|
|
|
|
|
|
def test_html_escaped_value_can_contain_other_entities() -> None:
|
|
"""The HTML-escaped pattern terminates only on a literal ``"``, not
|
|
on any ``&`` — so embedded entities like ``&`` survive."""
|
|
html = 'data="{"SNlM0e":"a&b"}"'
|
|
assert extract_wiz_field(html, "SNlM0e") == "a&b"
|
|
|
|
|
|
# ---------------------------------------------------------------------------
|
|
# Preview slicing — performance guard on huge payloads
|
|
# ---------------------------------------------------------------------------
|
|
|
|
|
|
def test_preview_slices_before_regex_substitution() -> None:
|
|
"""The whitespace-collapse regex must operate on a bounded prefix, not
|
|
the full payload — otherwise a multi-MB response would do a huge
|
|
needless substitution on the failure path."""
|
|
# A multi-megabyte payload with the key buried far past the preview window.
|
|
payload = (" " * 5_000_000) + "tail"
|
|
with pytest.raises(AuthExtractionError) as excinfo:
|
|
extract_wiz_field(payload, "SNlM0e", strict=True)
|
|
# The preview itself must still be capped — confirms the slice is tight
|
|
# enough that the rendered diagnostic does not balloon.
|
|
assert len(excinfo.value.payload_preview) <= AuthExtractionError.PREVIEW_LENGTH
|