ulysses/tests
lekt8 8c376d2b0e
feat: adapt agent_input_token_budget to the model context window (#1170) (#1230)
The agent soft-trims input context to `agent_input_token_budget` (default 6000).
The old computation `min(context_length or budget, budget)` made the 6000 default
a hard ceiling for every model, so 128K/1M context models were silently capped at
6000 input tokens — now that num_ctx is sent correctly (#1056), this was the last
barrier to actually using a long context window.

This derives the default budget from the model's discovered context window
(~85%, capped at a generous hard max) while honouring an explicit user setting
exactly (clamped to the window). When the window is unknown it falls back to the
previous value, so behaviour is unchanged for that case.

- src/context_budget.py: pure `compute_input_token_budget()` (unit-testable)
- src/settings.py: `is_setting_overridden()` to tell an explicit user value from
  the merged default (load_settings merges DEFAULT_SETTINGS, so equality alone
  can't distinguish them)
- src/agent_loop.py: use the helper in the soft-trim path

Covered by tests/test_context_budget.py (6 cases).

Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-06-03 00:13:53 +09:00
..
bombadil-spec.ts Odysseus v1.0 2026-05-31 23:58:26 +09:00
conftest.py Odysseus v1.0 2026-05-31 23:58:26 +09:00
test_action_intents.py Route calendar action requests to tools 2026-06-01 14:32:41 +09:00
test_admin_wipe_gallery.py Admin: wipe gallery albums with images 2026-06-02 20:35:57 +09:00
test_agent_loop.py fix(agent): extract web search sources from output key 2026-06-02 13:06:09 +09:00
test_api_key_manager_resilience.py API keys: skip undecryptable entries on load 2026-06-02 20:28:26 +09:00
test_api_token_routes.py tests: cover API token CRUD routes 2026-06-02 20:42:32 +09:00
test_app.py Fix fresh checkout test failures 2026-06-01 02:22:17 +00:00
test_app_static_mime.py fix: normalize JS static MIME types on Windows 2026-06-02 01:32:00 +02:00
test_auth_event_loop.py Stabilize security regression tests 2026-06-02 05:48:59 +09:00
test_auth_regressions.py Add tests for open-signup endpoint 2026-06-02 20:42:10 +09:00
test_auth_session_revocation.py Stabilize auth session revocation tests 2026-06-02 06:02:49 +09:00
test_backup_cli_security.py Harden backup restore tar extraction 2026-06-02 05:55:03 +09:00
test_builtin_memory_consolidation.py Scope memory consolidation by owner group 2026-06-02 12:40:28 +09:00
test_calendar_event_contrast.py Improve calendar event text contrast (#1184) 2026-06-02 23:14:52 +09:00
test_calendar_owner_scope.py tests: cover calendar route owner gates 2026-06-02 20:42:37 +09:00
test_calendar_recurrence.py Fix YEARLY recurring CalDAV events only showing on DTSTART year (#179) 2026-06-01 13:42:44 +09:00
test_chat_attachment_picker.py Chat attachments: allow picker to choose any file type 2026-06-02 20:55:30 +09:00
test_chat_cached_model_normalization.py Chat: use cached endpoint model ids before probing 2026-06-02 21:00:58 +09:00
test_chat_image_routing.py Chat: route image sessions only to matching image endpoints 2026-06-02 20:52:03 +09:00
test_chat_metrics.py Chat metrics: surface backend generation speed 2026-06-02 20:52:08 +09:00
test_chat_stream_scope.py Fix chat stream recovery and PDF library indexing (#468) 2026-06-01 22:33:35 +09:00
test_chroma_client.py fix: ChromaDB unreachable blocks app startup for 30-60s (#326) (#476) 2026-06-01 22:22:41 +09:00
test_cleanup_owner_scope.py tests: cover cleanup owner scope 2026-06-02 20:42:21 +09:00
test_companion_pairing.py Companion: fix pairing admin guard import 2026-06-02 20:29:37 +09:00
test_companion_readonly.py Tests: companion model JSON resilience 2026-06-02 13:15:22 +09:00
test_compare_js.py Fix duplicate compare modal on repeated clicks (#491) 2026-06-01 22:24:27 +09:00
test_context_budget.py feat: adapt agent_input_token_budget to the model context window (#1170) (#1230) 2026-06-03 00:13:53 +09:00
test_context_compactor.py Preserve large pasted messages in context 2026-06-01 12:38:35 +09:00
test_cookbook_dependency_completion_regression.py Fix Cookbook dependency install completion state 2026-06-02 12:59:29 +09:00
test_cookbook_endpoint_registration.py Fix Cookbook container-local model endpoints (#1223) 2026-06-03 00:09:48 +09:00
test_cookbook_helpers.py feat: add vllm kv cache dtype option (#1185) 2026-06-02 23:17:16 +09:00
test_cookbook_package_detection.py fix(cookbook): detect llama-cpp-python via its real distribution name (#1020) (#1167) 2026-06-02 22:52:37 +09:00
test_ddg_redirect_resolution.py Match host, not substring, when resolving DuckDuckGo redirects (#886) 2026-06-02 12:25:56 +09:00
test_deep_research_extraction_controls.py Allow longer deep research extraction timeouts (#651) 2026-06-02 11:50:03 +09:00
test_deep_research_search_error.py Research: report empty search provider results clearly 2026-06-02 20:34:25 +09:00
test_deleted_session_sidebar_regression.py Fix stale deleted sessions in sidebar (#1203) 2026-06-02 23:52:22 +09:00
test_dialog_aria.py Add dialog accessibility semantics 2026-06-02 12:41:25 +09:00
test_document_deeplink.py fix: open #document deep-links on refresh and surface load errors (#631) 2026-06-02 11:48:54 +09:00
test_document_pdf_marker.py Documents: strip PDF marker without corrupting text 2026-06-02 20:35:27 +09:00
test_document_tool_owner_scope.py Scope document tools to caller owner 2026-06-02 06:00:02 +09:00
test_email_imap_timeout.py Use shared IMAP timeout for account tests (#1088) 2026-06-02 23:11:04 +09:00
test_email_library_bulk_actions.py Email: persist bulk read state to provider 2026-06-02 20:28:01 +09:00
test_email_smtp_security.py Email: add explicit SMTP security mode 2026-06-02 13:15:06 +09:00
test_email_split_border_css.py fix(ui): contain email split divider (#1194) 2026-06-02 23:28:24 +09:00
test_endpoint_probing.py Fix Cookbook container-local model endpoints (#1223) 2026-06-03 00:09:48 +09:00
test_endpoint_resolver.py Never resolve to a disabled endpoint model (#861) 2026-06-02 11:10:43 +09:00
test_esc_menu_stack_js.py fix: make transient dropdown/popup menus close on Escape 2026-06-01 14:23:22 -04:00
test_extract_quotes.py fix: extract_quotes accepts mismatched opening/closing quotes (#1113) 2026-06-02 22:34:52 +09:00
test_extract_statistics.py fix: extract_statistics drops large numbers and trailing % signs (#1153) 2026-06-02 22:35:30 +09:00
test_forwarded_message_divider.py Email: recognize forwarded message dividers 2026-06-02 20:32:56 +09:00
test_gallery_endpoint_matching.py Gallery: match image endpoint URLs with exact v1 suffix 2026-06-02 20:34:05 +09:00
test_gallery_image_privileges.py Gate image editor AI endpoints by privilege (#447) 2026-06-01 22:35:24 +09:00
test_history_topics_owner_scope.py fix(history): scope topic analysis to authenticated owner only (#744) 2026-06-02 11:36:01 +09:00
test_hwfit_amd.py Cookbook fit: steer consumer AMD to GGUF recommendations 2026-06-02 21:01:42 +09:00
test_hwfit_macos.py fix: require GGUF sources for llama downloads (#368) 2026-06-01 22:47:47 +09:00
test_hwfit_manual_backend.py fix(hwfit): honor manual "metal" backend in the hardware simulator (#1090) 2026-06-02 23:12:34 +09:00
test_hwfit_quant_formats.py Fix native Cookbook quant classification 2026-06-02 13:07:20 +09:00
test_ics_escape.py fix: ICS export doesn't escape commas/semicolons in event fields (#1161) 2026-06-02 22:36:12 +09:00
test_ics_export_escaping.py fix: ICS export — escape X-WR-CALNAME and honour is_utc on DTSTART/DTEND (#1174) 2026-06-02 23:02:28 +09:00
test_keybind_altgr_js.py Ignore AltGr keystrokes in Ctrl+Alt keyboard shortcuts (#825) 2026-06-02 11:12:54 +09:00
test_llm_core_anthropic_cache.py Add Anthropic prompt caching to the agent loop (#812) 2026-06-02 11:14:31 +09:00
test_llm_core_concurrency.py Make LLM host health maps thread-safe 2026-06-02 05:54:23 +09:00
test_llm_core_fallback.py Surface silent model fallback instead of masking it (#868) 2026-06-02 11:37:25 +09:00
test_llm_core_ollama.py Ollama: pass discovered num_ctx in chat requests 2026-06-02 20:27:24 +09:00
test_llm_core_reasoning.py Support vLLM 0.20.2 / NIM reasoning-parser output end-to-end (surface + agent context + render) (#602) 2026-06-02 11:48:17 +09:00
test_llm_core_sanitize_tool_calls.py Chat: merge consecutive user messages for strict providers 2026-06-02 20:44:13 +09:00
test_llm_core_streaming.py Fix native tool-calling follow-up round on Gemini and Ollama (#867) 2026-06-02 11:39:40 +09:00
test_llm_core_temperature.py Providers: omit temperature for OpenAI reasoning models 2026-06-02 20:58:33 +09:00
test_lmstudio_discovery.py Discover LM Studio via host/port scanning and native-API fingerprint (#1126) 2026-06-02 23:04:58 +09:00
test_lmstudio_vision.py Use LM Studio-reported vision capability for image passthrough (#1130) 2026-06-02 23:01:04 +09:00
test_local_endpoint_api_key_js.py Models: allow API keys for local endpoints 2026-06-02 20:36:54 +09:00
test_local_endpoint_js.py fix: don't bill self-hosted models reached by a container/service hostname (#596) 2026-06-02 11:47:58 +09:00
test_markdown_rendering_js.py Fix ordered list rendering in markdown preview (#645) 2026-06-02 11:49:44 +09:00
test_markdown_table_row_js.py fix: markdown tables drop empty cells and misalign columns (#1164) 2026-06-02 22:41:27 +09:00
test_markitdown_runtime.py Add optional markitdown extraction for Office/EPUB documents (#766) 2026-06-02 11:28:52 +09:00
test_mcp_manager.py fix: add Browser MCP connection diagnostics (#662) 2026-06-02 11:50:17 +09:00
test_memory_bullet_extraction.py Fix AttributeError on bullet lines in extract_memory_from_chat (#873) 2026-06-02 11:46:06 +09:00
test_memory_extractor_vector_degraded.py fix: data integrity — deep-research result parsing + memory-extraction durability (#808) 2026-06-02 11:27:31 +09:00
test_modal_dock_composer_clearance.py fix(ui): keep minimized windows above composer (#1197) 2026-06-02 23:31:09 +09:00
test_model_context.py Models: prefer longest known context match 2026-06-02 20:33:09 +09:00
test_model_routes.py fix(models): clear deleted endpoint fallback refs (#1207) 2026-06-02 23:41:04 +09:00
test_new_chat_model_preference.py Chat: prefer active model for new desktop chats 2026-06-02 21:00:50 +09:00
test_notes_update_due_date.py Notes: parse natural-language due dates on update 2026-06-02 20:51:16 +09:00
test_null_owner_gates.py Security: owner-scope v1 chat endpoint fallback 2026-06-02 20:31:35 +09:00
test_og_image_extraction.py fix: source thumbnails dropped for http-only og:image URLs (#667) 2026-06-02 11:41:33 +09:00
test_ollama_port_detection.py Add Ollama port path detection regressions (#883) 2026-06-02 12:24:18 +09:00
test_pdf_runtime.py Show a clear message when PyMuPDF is missing 2026-06-01 18:27:17 +09:00
test_personal_docs_exclusions.py Docs: respect path boundary when clearing exclusions 2026-06-02 20:35:44 +09:00
test_personal_docs_office_index.py Add optional markitdown extraction for Office/EPUB documents (#766) 2026-06-02 11:28:52 +09:00
test_personal_docs_pdf_index.py Fix chat stream recovery and PDF library indexing (#468) 2026-06-01 22:33:35 +09:00
test_personal_upload_isolation.py Scope personal RAG uploads by owner (#446) 2026-06-01 22:36:53 +09:00
test_platform_compat.py Windows: improve Git Bash detection 2026-06-02 20:45:48 +09:00
test_preset_fill_missing_defaults.py Presets: fill missing built-in defaults on load 2026-06-02 20:32:08 +09:00
test_prompt_bar_manual_resize.py fix(ui): allow manual prompt bar resize (#1201) 2026-06-02 23:43:53 +09:00
test_provider_classification.py Add provider classification and upstream-error tests 2026-06-02 20:42:43 +09:00
test_provider_detection.py Provider detection: match by hostname instead of substring (re #768) (#815) 2026-06-02 11:11:17 +09:00
test_provider_endpoints.py Add provider endpoint resolver tests 2026-06-02 12:53:50 +09:00
test_rag_vector_id_stability.py fix(rag): use a stable hash for document IDs so dedup survives restarts (#1098) 2026-06-02 22:42:23 +09:00
test_rate_limiter.py Odysseus v1.0 2026-05-31 23:58:26 +09:00
test_readiness.py feat: add /api/ready readiness probe (DB, data dir, local-first) (#1200) 2026-06-02 23:33:22 +09:00
test_rename_user_case_insensitive.py fix(auth): case-insensitive owner migration on username rename (#1183) 2026-06-02 23:18:15 +09:00
test_reply_recipients_js.py fix: reply-all Cc's the user's own other addresses (multi-account) (#672) 2026-06-02 11:42:20 +09:00
test_research_owner_scope_routes.py tests: cover research route owner gates 2026-06-02 20:42:15 +09:00
test_research_probe_errors.py Surface deep research probe errors (#1086) 2026-06-02 22:51:25 +09:00
test_research_query_fallback.py Deep research: don't treat a bare 'yes' as the research topic (#858) 2026-06-02 11:30:53 +09:00
test_research_service.py fix: data integrity — deep-research result parsing + memory-extraction durability (#808) 2026-06-02 11:27:31 +09:00
test_research_session_id_validation.py fix(research): validate session_id to block path traversal 2026-06-01 23:25:38 +01:00
test_research_utils.py fix: deep research discards valid sources mentioning cookies/copyright (#481) 2026-06-01 22:26:37 +09:00
test_reserved_username_admin_escalation.py Reserve internal sentinel usernames 2026-06-02 05:58:58 +09:00
test_resolve_endpoint_fallbacks.py Add resolve_endpoint fallback chain regressions (#890) 2026-06-02 12:24:50 +09:00
test_review_regressions.py Models: avoid hidden models in default fallback 2026-06-02 20:37:14 +09:00
test_scheduler_restart_doublefire.py fix(scheduler): push next_run forward on startup to stop restart double-fire (#708) 2026-06-02 11:43:30 +09:00
test_search_cache_invalidation.py Fix invalidate_search_cache using a key that never matches stored entries (#852) 2026-06-02 10:53:33 +09:00
test_search_config_provider_key.py Report provider-specific search API keys correctly (#1202) 2026-06-02 23:37:15 +09:00
test_search_content_extraction_parity.py Search: align service content extraction 2026-06-02 20:53:07 +09:00
test_search_module_consolidation.py Search: consolidate core and provider implementations 2026-06-02 21:02:26 +09:00
test_search_query.py Fix year extraction in research queries 2026-06-01 23:09:41 +09:00
test_search_ranking.py Odysseus v1.0 2026-05-31 23:58:26 +09:00
test_security_regressions.py Security: sanitize export and gallery filenames 2026-06-02 20:29:56 +09:00
test_select_dropdown_theme_css.py Normalize native select option theming (#1178) 2026-06-02 23:09:15 +09:00
test_serve_profiles.py Cookbook serve profiles and engine filter 2026-06-02 12:34:42 +09:00
test_service_search_provider_guards.py Search: consolidate core and provider implementations 2026-06-02 21:02:26 +09:00
test_session_ghost_delete.py Sessions: allow deleting memory-only ghost sessions 2026-06-02 20:51:26 +09:00
test_session_mode_helpers.py Fix database stubs in regression tests (#301) 2026-06-01 16:55:09 +09:00
test_session_owner_attribution.py Attribute API-token sessions to the token owner (effective_user) (#871) 2026-06-02 11:39:01 +09:00
test_settings_scrub.py Deep-scrub secrets from public settings 2026-06-01 23:11:50 +09:00
test_setup_admin_user.py Normalize setup admin username (#448) 2026-06-01 22:38:56 +09:00
test_shell_routes.py fix: distinguish external cookbook runtimes (#1188) 2026-06-02 23:20:00 +09:00
test_skill_index_prompt_injection.py fix(agent-loop): wrap matched skills + skill index in untrusted user-role message (#788) 2026-06-02 11:15:45 +09:00
test_skills_delete_owner.py Skills: delete owner-scoped skills with owner 2026-06-02 20:28:36 +09:00
test_skills_manager_owner_isolation.py fix(skills): scope skill reads to caller owner (#777) 2026-06-02 11:21:27 +09:00
test_speech_service_toggles.py Honor disabled speech service toggles (#814) 2026-06-02 10:44:39 +09:00
test_sqlite_foreign_keys.py DB: enable SQLite foreign key cascades 2026-06-02 20:36:13 +09:00
test_strip_think.py Text: strip dangling think blocks after visible text 2026-06-02 20:36:37 +09:00
test_stt_leak.py STT: clean temp audio files on transcription failure 2026-06-02 20:43:24 +09:00
test_task_scheduler_cancel.py Tasks: clean up queued cancellation state 2026-06-02 20:51:21 +09:00
test_task_scheduler_session_delivery.py Fix test suite: ESM module loading and stub isolation (#844) 2026-06-02 11:29:29 +09:00
test_tool_index_keyword_boundaries.py Tools: match keyword hints on word boundaries 2026-06-02 20:32:20 +09:00
test_tool_path_confinement.py fix(tools): strict path confinement with sensitive-subpath deny list (#1072) 2026-06-02 23:13:30 +09:00
test_tool_support_heuristic.py fix: deepseek-r1 on Ollama returns HTTP 400 when tool schemas are sent (#1169) 2026-06-02 23:22:57 +09:00
test_topic_analyzer.py Topics: hydrate session history before analysis 2026-06-02 20:44:27 +09:00
test_tts_cache_stats.py TTS: include mp3 files in cache stats 2026-06-02 20:43:29 +09:00
test_unknown_tool_calls.py fix(agent): map native google_search and surface empty rounds 2026-06-02 12:57:45 +09:00
test_upload_handler_atomicity.py Uploads: write uploads index atomically 2026-06-02 20:51:39 +09:00
test_upload_routes_owner_scope.py tests: cover upload route owner gates 2026-06-02 20:42:26 +09:00
test_url_safety.py fix: SSRF hardening for the custom embedding endpoint URL (#132) (#1206) 2026-06-02 23:46:33 +09:00
test_vault_password_not_in_argv.py fix(security): stop leaking the vault master password via process argv (#879) 2026-06-02 12:25:43 +09:00
test_venice_hosts.py Treat Venice as a tool-capable SOTA cloud provider (#1173) 2026-06-02 23:03:46 +09:00
test_vision_model_detection.py Recognize local vision models so their images aren't dropped (#185) 2026-06-01 13:09:21 +09:00
test_visual_report.py Fix visual report chapter navigation (#505) 2026-06-01 22:26:13 +09:00
test_webhook_ssrf_resilience.py Webhook: block IPv6 SSRF bypasses 2026-06-02 20:28:12 +09:00
test_webhook_trigger_auth_exempt.py Exempt task webhook trigger from session auth (#784) 2026-06-02 11:23:40 +09:00
test_windows_update_script.py Windows: add Docker update script 2026-06-02 20:45:32 +09:00
test_youtube_comments_timeout.py YouTube: enforce comment fetch timeout while waiting 2026-06-02 20:44:24 +09:00