MCPcopy Create free account

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

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

Functiontest_an_identical_run_reproduces_the_same_manifest_hash
(trace_dir)
agent/tests/test_run_manifest_wiring.py:84
Functiontest_an_incomplete_catch_up_keeps_the_gp_short
Only 109 distributable: the catch-up tier is truncated at 1 of its 2.
agent/tests/quantlib/test_fundmath.py:526
Functiontest_an_index_column_is_recognised_by_its_sequence
(table: str)
agent/tests/test_grounding_figures_contract.py:872
Methodtest_an_index_is_not_adjusted
(self)
agent/tests/test_tushare_loader.py:242
Functiontest_an_indicator_reading_is_checked_against_the_call_it_names
An RSI reading is a number like any other: declared, and then verified. It used to be masked by NAME — "RSI 46.7" was exempt because "rsi" was in
agent/tests/test_agent_grounding.py:1572
Methodtest_an_infinite_price_is_never_scaled_into_the_series
A non-finite leg is not a price, so the bar cannot be adjusted.
agent/tests/test_fmp_loader.py:248
Functiontest_an_integer_keeps_the_previous_behaviour
An integer's precision is unknown ("6,700" may be rounded to hundreds): the flat band still applies.
agent/tests/test_grounding_numeric_locale_rounding.py:378
Functiontest_an_integer_price_of_an_instrument_quoted_in_the_thousands_is_checked
最新收盘 1520" was unchecked for 600519.SH; "200 日均线" and a 0.6-yuan ETF's "20 日" stay so.
agent/tests/test_grounding_role_hardening.py:1071
Functiontest_an_invented_value_declared_from_a_real_run_is_rejected
Naming the run is not a pass: the value must be one it wrote.
agent/tests/test_grounding_backtest_outputs.py:238
Functiontest_an_invented_value_is_a_mismatch_not_a_missing_ref
The rule keys on the value the figure matches, not on the session merely holding tail risk: 2.11% is the ES 95% and needs a ref, while 3.10% is in
agent/tests/test_grounding_role_hardening.py:765
Functiontest_an_iso_code_glued_to_digits_is_a_currency_mark
The identifier lookbehind used to turn "CNY0.888" into a bare "888".
agent/tests/test_grounding_shape_hardening.py:162
Functiontest_an_observed_value_is_never_cut
A correctly quoted close beside an ungrounded metric survives the cut.
agent/tests/test_grounding_release.py:1830
Methodtest_an_offset_past_a_section_is_rejected
(self, tool: LoadSkillTool)
agent/tests/test_agent_output_discipline.py:877
Methodtest_an_offset_past_the_end_is_rejected
(self, tool)
agent/tests/test_load_skill_paging.py:78
Functiontest_an_old_serialized_session_with_no_owner_key_still_loads
()
agent/tests/test_principal.py:114
Functiontest_an_omitted_non_equity_value_is_unknown_not_zero
()
agent/tests/test_robinhood_mcp_mapping.py:54
Functiontest_an_oos_split_inside_the_prices_but_outside_the_ic_skips_the_alpha
( monkeypatch: pytest.MonkeyPatch, split: str )
agent/tests/factors/test_bench_strict.py:694
Functiontest_an_oos_split_leaving_two_ic_observations_is_still_measured
( monkeypatch: pytest.MonkeyPatch, )
agent/tests/factors/test_bench_strict.py:711
Functiontest_an_open_with_negative_cash_is_skipped_and_reported_in_hold_mode
With cash below zero no scale fits, not even the empty plan. The search only replaced the full-scale plan when a candidate fitted, so the ful
agent/tests/test_plan_rejected.py:427
Functiontest_an_order_symbol_with_no_bar_at_all_is_halted_not_data_not_ready
()
agent/tests/test_paper_trade_5m_tool.py:293
Functiontest_an_ordinal_needs_the_space_after_its_punctuation
A row number glued to the pipe is a table cell, like "| 1 |" — not a list marker.
agent/tests/test_grounding_shape_hardening.py:683
Functiontest_an_ordinary_response_without_usage_is_still_announced
(tmp_path)
agent/tests/test_llm_cost_accounting.py:349
Functiontest_an_over_long_field_is_refused_with_the_documented_error
csv raises its own error for an over-long field; callers must not see it.
agent/tests/test_extraetf_reader.py:572
Functiontest_an_overriding_subclass_still_gets_the_default_counting
A subclass that calls super() keeps both the hook and the metrics.
agent/tests/test_plan_rejected.py:297
Methodtest_an_oversized_body_is_truncated
(self)
agent/tests/test_mcp_client_adapter.py:857
Methodtest_an_oversized_result_says_it_was_cut
(self)
agent/tests/test_tool_result_paging.py:40
Methodtest_an_oversized_result_still_fits_the_budget
(self)
agent/tests/test_tool_result_paging.py:37
Methodtest_an_oversized_section_pages_within_itself
(self, tool: LoadSkillTool)
agent/tests/test_agent_output_discipline.py:861
Functiontest_an_oversized_trade_moves_part_way_and_reports_the_shortfall
()
agent/tests/test_factor_costs.py:40
Functiontest_an_unavailable_origin_cannot_carry_a_timestamp
Provenance has to agree with itself in both directions.
agent/tests/test_extraetf_reader.py:880
Methodtest_an_uncapped_list_is_not_flagged_as_truncated
(self)
agent/tests/test_institutional_holdings_tool.py:528
Functiontest_an_unchecked_measurement_never_streams_before_the_gate
An unbuffered draft stops streaming at its first measurement; a rejection resets it.
agent/tests/test_grounding_release.py:2269
Functiontest_an_unclassified_kind_is_refused_rather_than_assumed
Silently ignoring an unknown kind would return a plausible wrong number.
agent/tests/quantlib/test_performance.py:239
Methodtest_an_unclosed_info_cell_does_not_shift_the_columns
(self)
agent/tests/test_etf_holdings_tool.py:809
Methodtest_an_undated_bar_with_no_volume_is_not_filled_by_the_bar_before
Without dates, bars align by position: dropping the gap would report bar 28 as latest.
agent/tests/test_technical_indicator_tool.py:470
Functiontest_an_undeclared_delta_is_reported_in_every_spelling
The lower arm: a figure with no origin is named, not silently masked.
agent/tests/test_grounding_percentage_points.py:82
Functiontest_an_undeclared_figure_is_checked_exactly_as_before
The undeclared path still accepts only prices and metric kinds. Widening it to the whole metrics row accepted 16% of invented decimals on the
agent/tests/test_grounding_backtest_outputs.py:563
Functiontest_an_undeclared_figure_is_grounded_only_within_its_own_rounding
( tmp_path: Path, prose: str, valid: bool )
agent/tests/test_grounding_numeric_locale_rounding.py:317
Functiontest_an_undeclared_figure_on_the_last_draft_is_cut_not_fallen_back
(tmp_path: Path)
agent/tests/test_grounding_shape_hardening.py:642
Functiontest_an_undeclared_observed_value_is_not_refused
(tmp_path: Path)
agent/tests/test_grounding_role_hardening.py:516
Functiontest_an_undeclared_percentage_is_reported_not_masked
The lower side: a percentage with no declaration is named, not ignored. A percentage-range mask made "距阻力位仅 1–2%" invisible, and with it every
agent/tests/test_agent_grounding.py:1636
Functiontest_an_undeclared_tail_risk_percent_needs_a_field_ref
The undeclared half of the same rule: a percent a tool returned needs no declaration, so it was matched against every tail-risk value in the sessi
agent/tests/test_grounding_role_hardening.py:738
Functiontest_an_unevidenced_price_is_still_rejected_without_any_tool_call
Relaxing the identity check must not license a remembered quote.
agent/tests/test_agent_grounding.py:2270
Functiontest_an_unknown_flow_timing_token_is_rejected
A typo must not silently select the default convention.
agent/tests/quantlib/test_performance.py:288
Methodtest_an_unknown_section_errors_and_lists_the_real_ones
(self, tool: LoadSkillTool)
agent/tests/test_agent_output_discipline.py:855
Methodtest_an_unknown_skill_still_errors
(self, tool)
agent/tests/test_load_skill_paging.py:68
Methodtest_an_unknown_skill_still_errors
(self, tool: LoadSkillTool)
agent/tests/test_agent_output_discipline.py:797
Functiontest_an_unknown_tier_is_refused
(tools)
agent/tests/test_sdm_governance_wiring.py:162
Methodtest_an_unlisted_product_reports_both_defaults
(self)
agent/tests/test_global_futures_pricing_assumptions.py:58
Functiontest_an_unmet_hurdle_pays_the_gp_nothing
()
agent/tests/quantlib/test_fundmath.py:536
Functiontest_an_unnamed_source_is_still_reported
Accepting a localized provider name is not accepting no provider name.
agent/tests/test_agent_grounding.py:2177
Methodtest_an_unreachable_announcement_index_says_so_instead_of_going_quiet
(self)
agent/tests/test_etf_holdings_tool.py:723
Methodtest_an_unreadable_cover_page_falls_back_to_the_form_and_says_so
(self)
agent/tests/test_institutional_holdings_tool.py:753
Methodtest_an_unreadable_restatement_falls_back_even_when_new_holdings_stands_in
The fallback must key on 'no restatement applied', not on 'no answer at all'. DZ BANK's real chain is original + RESTATEMENT + NEW HOLDINGS.
agent/tests/test_institutional_holdings_tool.py:855
Methodtest_an_unreadable_restatement_falls_back_to_what_it_superseded
Skipping superseded tables must never turn one bad fetch into no answer.
agent/tests/test_institutional_holdings_tool.py:834
Methodtest_an_unreadable_table_leaves_the_prior_intact_and_says_so
When the check cannot run, the number must not silently change.
agent/tests/test_institutional_holdings_tool.py:1437
Methodtest_an_unrecognised_model_is_not_sent_it
The allowlist is positive: unknown means no, not maybe. Guessing wrong breaks the request outright, while a missing entry only leaves
agent/tests/test_llm_reasoning_effort.py:508
Functiontest_an_unregistered_tools_leaf_never_grounds_a_price
A price-shaped leaf name from an unregistered tool is not price evidence. ``_ingest_generic_numeric`` classifies every non-market tool result, so
agent/tests/test_grounding_release.py:1245
Functiontest_an_unrelated_400_is_not_retried
(monkeypatch, fresh_replay_state)
agent/tests/test_openai_codex_replay.py:256
Functiontest_an_unrelated_400_is_raised_untouched_and_nothing_is_remembered
()
agent/tests/test_llm_responses_self_heal.py:269
Methodtest_an_unrelated_xml_candidate_does_not_abort_the_filing
Only the root tag identifies the table, so a decoy XML must be skipped.
agent/tests/test_institutional_holdings_tool.py:671
Methodtest_an_unrepeated_title_still_takes_a_bare_name
(self, repeat: LoadSkillTool)
agent/tests/test_agent_output_discipline.py:981
Functiontest_an_unreported_non_equity_value_is_not_read_as_zero
(tmp_path, broker: _Broker)
agent/tests/test_robinhood_portfolio_source.py:199
Methodtest_an_unset_effort_stays_absent
(self)
agent/tests/test_llm_reasoning_effort.py:360
Functiontest_an_unsolvable_irr_is_reported_as_a_note_not_a_failure
A total loss with no distributions has no IRR; the rest still ships.
agent/tests/test_cashflow_analytics_tool.py:242
Functiontest_an_unsupported_5m_instrument_is_an_in_band_refusal_before_fetch
(symbol: str)
agent/tests/test_paper_trade_5m_scope.py:58
Functiontest_an_untagged_fence_is_prose_and_a_tagged_one_is_code
A trading plan set off in a bare fence was released unchecked.
agent/tests/test_grounding_figures_contract.py:835
Functiontest_an_unterminated_block_still_parses
A truncated answer must not silently un-declare its own figures.
agent/tests/test_grounding_figures_contract.py:190
Methodtest_an_untruncated_full_portfolio_reports_nothing_withheld
(self)
agent/tests/test_etf_holdings_tool.py:1095
Methodtest_an_untyped_amendment_alone_is_reported_but_labelled_unverified
(self)
agent/tests/test_institutional_holdings_tool.py:1033
Methodtest_an_untyped_amendment_is_reported_unresolved_not_guessed
AEGON files 13F-HR/A with no amendmentType; neither behaviour is assumed.
agent/tests/test_institutional_holdings_tool.py:979
Functiontest_an_unusable_band_is_refused
A band is a modelling choice; a meaningless one must not be assumed away.
agent/tests/test_rebalance_tolerance.py:102
Functiontest_analysis_completion_is_persisted_with_metric_provenance
The grounding artifact records why an analysis figure was accepted.
agent/tests/test_agent_grounding.py:3083
Functiontest_analysis_context_can_be_pinned_to_a_stored_snapshot
A later refresh must not move a read that names its snapshot.
agent/tests/test_portfolio_service.py:911
Functiontest_analysis_context_route_pins_snapshot_and_never_falls_back
(monkeypatch)
agent/tests/test_portfolio_routes.py:112
Functiontest_analysis_context_supplies_risk_xray_arguments
The portfolio feeds the existing risk x-ray; it does not reimplement it.
agent/tests/test_portfolio_service.py:599
Functiontest_analysis_definition_is_declared_not_phrased
A threshold convention is a citation, not a measurement of this session. "通常被认为" used to exempt the clause outright. The convention is now st
agent/tests/test_agent_grounding.py:3047
Functiontest_analysis_gate_still_rejects_unattributed_and_unsourced
The attribution exemption must not launder model-memory figures: the review's pinned reject pair and an invented metric with an ordinary prose
agent/tests/test_agent_grounding.py:3308
Functiontest_analysis_gates_accept_attributed_paper_restatement
An attributed figure is a citation, not an invented measurement (#1338 review): 'The paper reports a Sharpe ratio of 1.8' must pass with zero tool
agent/tests/test_agent_grounding.py:3289
Functiontest_analysis_mention_without_figures_is_allowed
#1336: refusal prose naming the gap is not a quantitative claim.
agent/tests/test_agent_grounding.py:2869
Functiontest_analysis_only_candidates_preserve_the_figure_kind
(tmp_path: Path, written: str)
agent/tests/test_grounding_unknown_call_id_refs.py:230
Functiontest_analyze_full_includes_profile_and_behavior
(allow_tmp: Path)
agent/tests/test_trade_journal.py:678
Functiontest_analyze_mc_results_on_a_simulated_run
()
agent/tests/quantlib/test_risk.py:432
Functiontest_analyze_mc_results_rejects_a_one_column_matrix
()
agent/tests/quantlib/test_risk.py:441
Functiontest_analyze_mc_results_reuses_the_var_convention
()
agent/tests/quantlib/test_risk.py:406
Functiontest_analyze_mc_results_tail_mask_is_not_fooled_by_an_all_negative_sample
()
agent/tests/quantlib/test_risk.py:421
Functiontest_analyze_missing_file
(allow_tmp: Path)
agent/tests/test_trade_journal.py:653
Functiontest_analyze_profile_only
(allow_tmp: Path)
agent/tests/test_trade_journal.py:723
Functiontest_analyze_reports_pnl_for_a_capitalized_aliased_journal
The tool must not report an empty, successful analysis over real fills. A journal headed ``Date,Symbol,Side,Qty,Price`` used to be detected as a
agent/tests/test_trade_journal.py:687
Functiontest_analyze_strategy_is_pending_placeholder
(allow_tmp: Path)
agent/tests/test_trade_journal.py:712
Functiontest_analyze_unsupported_extension
(allow_tmp: Path)
agent/tests/test_trade_journal.py:659
Functiontest_analyze_with_filter
(allow_tmp: Path)
agent/tests/test_trade_journal.py:732
Functiontest_analyze_with_inverted_date_filter_returns_error_envelope
An inverted date range must surface as an error envelope, not a raw raise.
agent/tests/test_trade_journal.py:748
Methodtest_ancestors_are_derived_from_heading_depth
(self)
agent/tests/test_agent_output_discipline.py:996
Methodtest_ann_date_normalized
(self, fsp)
agent/tests/test_fetch_sina_penalties.py:165
Functiontest_annotation_catches_write_map_missed
A tool absent from the map but annotated write → WRITE (not excused).
agent/tests/test_classification.py:47
Methodtest_announcement_index_contradicting_the_period_blocks_the_full_claim
(self)
agent/tests/test_etf_holdings_tool.py:697
Methodtest_announcement_index_rejects_the_request_without_crashing
(self)
agent/tests/test_etf_holdings_tool.py:740
Methodtest_announcements_rows_and_multiple_api_names
(self, monkeypatch)
agent/tests/test_gildata_research_tools.py:74
Functiontest_annual_dates_agree_with_textbook_irr
365-day spacing must reproduce the equal-period IRR. The reference is the real root of ``sum(cf_i * x**i) == 0`` with ``x = 1/(1+r)``, found
agent/tests/quantlib/test_fundmath.py:127
← previousnext →14,901–15,000 of 26,887, ranked by callers