MCPcopy Create free account

hub / github.com/HKUDS/Vibe-Trading / functions

Functions26,887 in github.com/HKUDS/Vibe-Trading

Functiontest_a_malformed_line_fails_the_draft
It reaches the verdict, with the offending text in the message.
agent/tests/test_grounding_figures_contract.py:169
Functiontest_a_malformed_line_in_a_sealed_segment_is_reported_not_raised
verify_chain already reports malformed JSON as a clean ChainBreak instead of letting json.JSONDecodeError escape; the whole-history verifier m
agent/tests/test_governance.py:795
Functiontest_a_malformed_line_in_the_active_file_is_reported_not_raised
(tmp_path)
agent/tests/test_governance.py:816
Functiontest_a_malformed_line_with_nothing_to_cut_releases_the_checked_text
(tmp_path: Path)
agent/tests/test_grounding_shape_hardening.py:612
Functiontest_a_manifest_failure_never_breaks_the_run
(trace_dir, monkeypatch)
agent/tests/test_run_manifest_wiring.py:118
Functiontest_a_market_answer_with_no_price_and_no_analysis_is_still_refused
(tmp_path: Path)
agent/tests/test_grounding_backtest_outputs.py:659
Functiontest_a_market_data_rows_own_volume_is_evidence
A row is one observation, and its volume is part of that row. Restricting the pool to price fields alone rejected every report that quoted th
agent/tests/test_grounding_figures_contract.py:365
Functiontest_a_materially_different_figure_is_rejected_even_inside_the_old_half_percent_band
( tmp_path: Path, prose: str )
agent/tests/test_grounding_numeric_locale_rounding.py:279
Functiontest_a_measurement_declared_count_is_checked_not_refused
It is checked as the observation it is, so a real print still passes.
agent/tests/test_grounding_role_hardening.py:170
Functiontest_a_measurement_under_a_structural_column_is_checked
(header: str, cell: str)
agent/tests/test_grounding_shape_hardening.py:369
Functiontest_a_metadata_count_leaf_has_no_metric_kind
(leaf: str)
agent/tests/test_grounding_role_hardening.py:547
Functiontest_a_metric_backed_by_a_risk_tool_passes_in_prose_and_in_a_table
The other side: kind-mapped leaves ARE evidence, at either scale.
agent/tests/test_grounding_release.py:1751
Functiontest_a_metric_cell_restating_observed_closes_is_not_an_invented_metric
| 峰值→当前 | 1.180 → 1.110 |" states two observed closes, not a drawdown figure. The prose path exempts an observed value from the analysis check; t
agent/tests/test_grounding_release.py:1419
Functiontest_a_metric_leaf_keeps_its_kind
(leaf: str, kind: str)
agent/tests/test_grounding_role_hardening.py:567
Functiontest_a_metric_name_in_another_sense_registers_no_claim
(abstract, forbidden)
agent/tests/test_research_papers_tool.py:803
Functiontest_a_metric_needs_metric_evidence_or_its_own_arithmetic
A percentage is never answered by a price, however close the number is. A run holds hundreds of observed values spanning the instrument's range,
agent/tests/test_grounding_release.py:1732
Functiontest_a_mid_run_system_message_does_not_replace_the_system_prompt
The loop appends system messages mid-run; each one used to become ``instructions`` and every later request lost the whole system prompt.
agent/tests/test_openai_codex_replay.py:151
Methodtest_a_minute_is_still_a_minute
(self)
agent/tests/test_weekly_monthly_bars.py:182
Functiontest_a_missing_5m_loader_names_the_loader_not_the_provider
(monkeypatch)
agent/tests/test_paper_trade_5m_scope.py:225
Functiontest_a_missing_factor_is_excluded_not_read_as_zero
()
agent/tests/quantlib/test_combination.py:43
Functiontest_a_missing_file_inside_a_root_is_not_found
(run_dir: Path)
agent/tests/test_read_file_refusals.py:51
Functiontest_a_missing_total_value_is_an_error_not_a_zero_account
(tmp_path, broker: _Broker)
agent/tests/test_robinhood_portfolio_source.py:269
Functiontest_a_model_asking_to_compact_a_small_prompt_is_declined
(tmp_path, monkeypatch)
agent/tests/test_context_budget.py:327
Functiontest_a_model_authored_release_note_is_stripped_before_redacting
The marker and the footnote mean only what the gate put there. A draft carrying its own "※ 略去 0 处……" shipped two contradictory footnotes, and
agent/tests/test_grounding_release.py:1845
Methodtest_a_model_that_rejects_effort_is_not_sent_it
Haiku 4.5 answers `This model does not support the effort parameter`. This is the case that matters in a swarm: a per-agent model split puts
agent/tests/test_llm_reasoning_effort.py:491
Functiontest_a_model_whose_chat_endpoint_accepts_tools_is_left_alone
()
agent/tests/test_llm_responses_self_heal.py:259
Functiontest_a_module_outside_the_allowlist_is_unreachable
(tool)
agent/tests/test_quantlib_tool.py:30
Functiontest_a_money_figure_is_derived_from_money
( tmp_path: Path, markets: tuple[Any, ...], prose: str, row: str )
agent/tests/test_grounding_role_hardening.py:976
Methodtest_a_month_is_a_calendar_month_and_a_partial_last_one_stays_inside_it
(self)
agent/tests/test_weekly_monthly_bars.py:53
Functiontest_a_mostly_unlabelled_panel_keeps_the_tag_off
One "unknown" bucket is the global fallback wearing a sector label.
agent/tests/test_alpha_bench_sector_tag.py:96
Functiontest_a_movement_identifier_that_is_not_a_symbol_is_refused
A movement may carry a bare crypto symbol, but not arbitrary text.
agent/tests/test_extraetf_reader.py:864
Functiontest_a_movement_without_an_instrument_is_refused
A blank ISIN is a cash or fee row, not a movement this reader can place. Allowing it produced a movement with ``instrument_id=None`` — a movement
agent/tests/test_extraetf_reader.py:277
Functiontest_a_multi_year_window_is_annualised_geometrically
Two years of doubling annualises to sqrt(2) - 1.
agent/tests/quantlib/test_performance.py:468
Functiontest_a_naive_split_leaks_and_the_purged_split_does_not
()
agent/tests/quantlib/test_crossvalidation.py:35
Functiontest_a_name_with_unknown_volume_is_untradeable_not_infinitely_liquid
()
agent/tests/test_factor_costs.py:69
Functiontest_a_named_derivation_survives_prices_of_several_symbols
Several instruments' bars bar the session pools, not the evidence a ref names.
agent/tests/test_grounding_backtest_outputs.py:619
Methodtest_a_named_section_comes_back_on_its_own
(self, tool: LoadSkillTool)
agent/tests/test_agent_output_discipline.py:835
Functiontest_a_nat_label_end_purges_through_the_last_observation
NaT is the tail of ``shift(-h)``: a label resolving after the sample ends.
agent/tests/quantlib/test_crossvalidation.py:481
Functiontest_a_negative_mark_is_refused_rather_than_guessed
()
agent/tests/quantlib/test_fundmath.py:434
Functiontest_a_net_cash_peer_with_negative_ev_is_excluded_not_averaged_in
The gap the old one-sided rule left open. A net-cash-rich peer can carry a negative enterprise value with perfectly healthy EBITDA. EV/EBITDA
agent/tests/quantlib/valuation/test_comps.py:494
Methodtest_a_new_holdings_amendment_with_no_base_is_called_a_fragment
Nova Wealth Management's only 2026Q1 filing declares NEW HOLDINGS.
agent/tests/test_institutional_holdings_tool.py:1015
Methodtest_a_new_holdings_fragment_is_not_ranked_as_the_whole_portfolio
DZ BANK's NEW HOLDINGS amendment reports $1.17M against a $41.68B quarter.
agent/tests/test_institutional_holdings_tool.py:1454
Functiontest_a_no_break_space_before_the_percent_sign_is_still_a_percent
( tmp_path: Path, space: str )
agent/tests/test_grounding_shape_hardening.py:127
Functiontest_a_non_conclusive_re_resolution_cannot_retract_a_locked_symbol
A timeout, a refusal or an empty list is not evidence against the lock.
agent/tests/test_agent_grounding.py:323
Functiontest_a_non_conclusive_step_cannot_clear_a_conflict
A timeout after a contradiction leaves the contradiction standing.
agent/tests/test_agent_grounding.py:284
Functiontest_a_non_finite_adjusted_price_is_not_served
An adjusted leg of infinity is not a price: the bar is not adjusted.
agent/tests/test_tiingo_loader.py:233
Functiontest_a_non_finite_bound_is_refused_instead_of_dropped
A NaN threshold is not a bound the template can render, so it errors.
agent/tests/test_shadow_codegen_condition_bounds.py:205
Functiontest_a_non_finite_input_is_named_not_priced
Each range check is a comparison, and NaN fails none of them: it came out as a NaN price.
agent/tests/quantlib/test_fixedincome.py:594
Methodtest_a_non_http_exception_is_unchanged
(self)
agent/tests/test_mcp_client_adapter.py:863
Methodtest_a_non_integer_offset_is_rejected
(self, tool)
agent/tests/test_load_skill_paging.py:73
Methodtest_a_non_integer_offset_is_rejected
(self)
agent/tests/test_tool_result_paging.py:180
Methodtest_a_non_positive_factor_is_refused
(self)
agent/tests/test_alpha_bench_qfq.py:121
Functiontest_a_non_positive_invested_base_has_no_honest_return
An account at zero did not 'return' anything on the way back.
agent/tests/quantlib/test_performance.py:318
Methodtest_a_non_price_field_does_not_become_price_evidence
The P/E in the same payload must not license quoting 31.4 as a price.
agent/tests/test_agent_output_discipline.py:224
Functiontest_a_non_tail_risk_figure_is_untouched_by_the_ref_rule
A sharpe from another call still grounds off a call-scoped ref while the session holds four tail-risk identities -- the rule keys on the value the
agent/tests/test_grounding_role_hardening.py:781
Methodtest_a_non_us_market_suffix_is_not_stripped
(self)
agent/tests/test_sec_edgar_client.py:99
Methodtest_a_non_usd_futures_venue_is_not_called_usd
(self)
agent/tests/test_composite_currency_guard.py:44
Functiontest_a_note_that_is_not_arithmetic_is_not_a_derivation
A derivation needs an operation, not a restatement or a claim. "1.171" alone and "回测得出" are both refused, which is what makes it safe for the
agent/tests/test_grounding_release.py:249
Functiontest_a_note_token_is_a_cjk_pair_or_a_four_letter_word
()
agent/tests/test_grounding_role_hardening.py:209
Functiontest_a_null_bound_is_no_bound_in_the_generated_engine
``None`` means "no bound" here too, matching the scanner's ``_to_float``.
agent/tests/test_shadow_codegen_condition_bounds.py:196
Functiontest_a_number_behind_an_unrelated_word_is_not_linked_to_the_alias
()
agent/tests/test_research_papers_tool.py:902
Methodtest_a_number_that_tool_never_returned_is_still_rejected
(self, tmp_path: Path)
agent/tests/test_agent_output_discipline.py:200
Functiontest_a_number_the_paper_never_attached_to_the_metric_is_not_its_value
(abstract, metric)
agent/tests/test_research_papers_tool.py:872
Functiontest_a_numeric_glued_boundary_does_not_widen_the_evidence_quote
()
agent/tests/test_research_papers_tool.py:973
Functiontest_a_one_directional_stream_has_no_irr
Refusing is correct; there is no rate that zeroes an all-negative NPV.
agent/tests/quantlib/test_performance.py:422
Functiontest_a_one_sided_dict_bound_reaches_the_generated_engine
The scanner enforces `{"max": x}` alone; the engine must too.
agent/tests/test_shadow_codegen_condition_bounds.py:124
Functiontest_a_page_number_and_a_year_are_not_findings
Page 12 and the year 2025 are layout, not facts about the company.
agent/tests/test_grounding_document_figures.py:129
Methodtest_a_pair_no_venue_lists_resolves_to_nothing
(self, monkeypatch)
agent/tests/test_symbol_search_tool.py:755
Functiontest_a_parameter_without_a_currency_mark_is_a_count
A window, weight, threshold, multiplier or probability is the model's own choice.
agent/tests/test_grounding_role_hardening.py:151
Methodtest_a_partial_answer_names_the_full_portfolio_it_is_not
(self)
agent/tests/test_etf_holdings_tool.py:626
Functiontest_a_partial_catch_up_rate_still_lands_the_gp_on_its_carry
50% catch-up, 20% carry, pref 8, contributed 100, distributable 200. tier 1 return of capital 100 remaining 100 tier
agent/tests/quantlib/test_fundmath.py:494
Functiontest_a_partial_or_null_position_read_is_an_error_not_an_empty_account
( tmp_path, broker: _Broker, reply: dict, message: str )
agent/tests/test_robinhood_portfolio_source.py:217
Functiontest_a_partially_resolvable_batch_still_loads
One unresolvable symbol must not sink the symbols that do resolve.
agent/tests/test_fundamentals_pit.py:350
Functiontest_a_path_outside_the_readable_roots_is_refused_as_such
(run_dir: Path, path: str)
agent/tests/test_read_file_refusals.py:41
Methodtest_a_path_reaches_each_one_separately
(self, repeat: LoadSkillTool)
agent/tests/test_agent_output_discipline.py:972
Functiontest_a_percent_declared_as_a_fraction_is_named_in_the_correction
37% and 0.37 stay two assertions (spec §2), but the model is told which it wrote.
agent/tests/test_grounding_backtest_outputs.py:577
Functiontest_a_percent_figure_is_compared_in_percentage_points_only
约 0%" is not within half a unit of a 58% move — the half unit is in pp. A bare integer percent carries a half unit of 0.5, and running that band
agent/tests/test_grounding_release.py:1027
Functiontest_a_percent_is_never_a_proposed_level
(tmp_path: Path, row: str)
agent/tests/test_grounding_role_hardening.py:355
Functiontest_a_percent_is_never_grounded_by_a_price
(tmp_path: Path)
agent/tests/test_grounding_role_hardening.py:395
Functiontest_a_percentage_factor_and_a_return_ratio_are_not_offsets
(tmp_path: Path, prose: str, row: str)
agent/tests/test_grounding_role_hardening.py:261
Functiontest_a_period_after_a_number_with_no_space_still_splits
()
agent/tests/test_research_papers_tool.py:959
Functiontest_a_period_grouped_value_is_read_as_german_by_convention
``1.900`` is 1900, because the export is German and nothing says otherwise. This reading is a convention of the format rather than a deduction —
agent/tests/test_extraetf_reader.py:764
Methodtest_a_period_run_served_daily_bars_is_caught
The spacing check covers the new intervals too: daily bars under a weekly or monthly declaration annualise as daily, and say so.
agent/tests/test_cross_market_annualization.py:302
Methodtest_a_period_still_offering_more_rows_is_never_full
(self)
agent/tests/test_etf_holdings_tool.py:689
Functiontest_a_period_with_no_space_after_it_still_splits
()
agent/tests/test_research_papers_tool.py:952
Functiontest_a_phrase_that_finds_papers_is_never_widened
()
agent/tests/test_research_papers_tool.py:312
Functiontest_a_pinned_read_hides_a_snapshot_that_includes_a_disabled_source
(tmp_path)
agent/tests/test_portfolio_service.py:937
Functiontest_a_pinned_read_refuses_a_snapshot_valued_under_another_contract
(tmp_path, old_version)
agent/tests/test_portfolio_service.py:962
Methodtest_a_plain_filing_costs_exactly_one_cover_page_request
The filer index is only pulled when this quarter is actually amended.
agent/tests/test_institutional_holdings_tool.py:1492
Functiontest_a_plain_integer_is_a_count
(tmp_path: Path)
agent/tests/test_grounding_role_hardening.py:160
Functiontest_a_plain_list_stays_a_list
(tool)
agent/tests/test_quantlib_tool.py:193
Functiontest_a_point_mark_must_be_glued_and_whole
(text: str)
agent/tests/test_grounding_shape_hardening.py:546
Methodtest_a_portfolio_identical_to_the_benchmark_has_beta_one
(self)
agent/tests/test_metrics_tracking_error.py:66
Methodtest_a_portfolio_that_tracks_exactly_has_no_tracking_error
(self)
agent/tests/test_metrics_tracking_error.py:40
Functiontest_a_position_row_without_its_required_type_is_an_error
(broken: str | None)
agent/tests/test_robinhood_mcp_mapping.py:129
Functiontest_a_positions_read_without_a_positions_list_is_an_error_not_an_empty_source
(tmp_path)
agent/tests/test_portfolio_service.py:299
Functiontest_a_price_column_cell_is_not_grounded_by_a_metric
A close cell is compared with closes; a Sharpe of 0.888 is not one.
agent/tests/test_grounding_role_hardening.py:326
Functiontest_a_price_column_holds_observed_prints_only
(tmp_path: Path, row: str)
agent/tests/test_grounding_role_hardening.py:308
Functiontest_a_price_discount_and_a_drawdown_still_derive
(tmp_path: Path, prose: str, row: str)
agent/tests/test_grounding_role_hardening.py:992
Functiontest_a_price_under_the_date_column_is_refused
(tmp_path: Path)
agent/tests/test_grounding_shape_hardening.py:390
← previousnext →14,201–14,300 of 26,887, ranked by callers