MetaList 0.11.0__tar.gz → 0.13.0__tar.gz

{metalist-0.11.0 → metalist-0.13.0}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: metalist
-Version: 0.11.0
+Version: 0.13.0
 Summary: A minimalist single-user note-taking app with SSR, fast in-memory tree operations, and efficient sync.
 Project-URL: Homepage, https://github.com/evolvingstuff/metalist
 Project-URL: Repository, https://github.com/evolvingstuff/metalist
@@ -49,6 +49,24 @@ A minimalist single-user note-taking app focused on server-side rendering (SSR),
 - Multi-tab search contexts with server-persisted scroll/search state (survives browser restarts)
 - Manual namespace backups/restores to a user-selected backup folder with retention controls
 
+## Changes in 0.13.0
+
+- Dictation: text from speech-to-text tools such as Superwhisper, pasted into the search bar or any tag field, becomes your existing tags ("neural network" gives `neural-network`). Say "quote … end quote" for text, "minus" to exclude, "at" before meta tags such as `@todo`. In a note, the words after "start tags" go to its tag bar. The phrases are set in Dictation Settings, and one undo restores what was pasted.
+- A quoted search with a space at its edge matches whole words: `" fat"` no longer finds "father".
+- The AI keeps a record of what it did for each answer (notes read, searches, web pages opened), so questions like "did you search the web for that?" are answered from what actually happened.
+- Dropping a `.txt`, `.md`, `.csv` or `.json` file stores it as before and adds its text as a child note, shown as Markdown, a table or formatted JSON. File pills show the extension (`.MD`, `.CSV`, …), and hovering a pill says that clicking downloads the file.
+- Typing one letter in the tag bar now narrows the suggestions you were already shown instead of reordering them, so the tag you want is in the top five far more often.
+
+## Changes in 0.12.0
+
+- AI chat is now a tool-using agent: in one answer it can look up MetaList help, read and search the notes in your view, open web pages, and open menus, instead of picking a single route up front. Answers cite notes and web pages with clickable references, including each item listed from a web page.
+- Nothing changes without your Yes: tag operations ask first with the exact change, and web addresses that carry words from your notes ask before opening. Summaries that fit in one pass start without a dialog, and summaries never block other edits while they run.
+- "How do I change…" questions open the menu that holds the setting; release notes are part of the help. AI prompts and skills are no longer editable, and a $ button shows or hides the spend estimate.
+- The AI is told only that a note is private, never which privacy setting hides it, and failed AI steps are explained in the chat.
+- Leaving edit mode keeps your place, with no highlight; a collapsed note you changed stays expanded. Clicking into a note keeps the clicked spot under the pointer, including in diagrams and formulas.
+- Dragging a note shows a ghost, a drop line and a direction cursor; the note lands where you dropped it without the page scrolling.
+- Searches that cannot be parsed are explained politely instead of failing.
+
 ## Changes in 0.11.0
 
 - Large namespaces stay fast while scrolling: only a band of root notes around the viewport is loaded, the server remembers what each tab shows so the browser no longer sends note hashes, and rendered notes are cached.

{metalist-0.11.0 → metalist-0.13.0}/README.md

@@ -12,6 +12,24 @@ A minimalist single-user note-taking app focused on server-side rendering (SSR),
 - Multi-tab search contexts with server-persisted scroll/search state (survives browser restarts)
 - Manual namespace backups/restores to a user-selected backup folder with retention controls
 
+## Changes in 0.13.0
+
+- Dictation: text from speech-to-text tools such as Superwhisper, pasted into the search bar or any tag field, becomes your existing tags ("neural network" gives `neural-network`). Say "quote … end quote" for text, "minus" to exclude, "at" before meta tags such as `@todo`. In a note, the words after "start tags" go to its tag bar. The phrases are set in Dictation Settings, and one undo restores what was pasted.
+- A quoted search with a space at its edge matches whole words: `" fat"` no longer finds "father".
+- The AI keeps a record of what it did for each answer (notes read, searches, web pages opened), so questions like "did you search the web for that?" are answered from what actually happened.
+- Dropping a `.txt`, `.md`, `.csv` or `.json` file stores it as before and adds its text as a child note, shown as Markdown, a table or formatted JSON. File pills show the extension (`.MD`, `.CSV`, …), and hovering a pill says that clicking downloads the file.
+- Typing one letter in the tag bar now narrows the suggestions you were already shown instead of reordering them, so the tag you want is in the top five far more often.
+
+## Changes in 0.12.0
+
+- AI chat is now a tool-using agent: in one answer it can look up MetaList help, read and search the notes in your view, open web pages, and open menus, instead of picking a single route up front. Answers cite notes and web pages with clickable references, including each item listed from a web page.
+- Nothing changes without your Yes: tag operations ask first with the exact change, and web addresses that carry words from your notes ask before opening. Summaries that fit in one pass start without a dialog, and summaries never block other edits while they run.
+- "How do I change…" questions open the menu that holds the setting; release notes are part of the help. AI prompts and skills are no longer editable, and a $ button shows or hides the spend estimate.
+- The AI is told only that a note is private, never which privacy setting hides it, and failed AI steps are explained in the chat.
+- Leaving edit mode keeps your place, with no highlight; a collapsed note you changed stays expanded. Clicking into a note keeps the clicked spot under the pointer, including in diagrams and formulas.
+- Dragging a note shows a ghost, a drop line and a direction cursor; the note lands where you dropped it without the page scrolling.
+- Searches that cannot be parsed are explained politely instead of failing.
+
 ## Changes in 0.11.0
 
 - Large namespaces stay fast while scrolling: only a band of root notes around the viewport is loaded, the server remembers what each tab shows so the browser no longer sends note hashes, and rendered notes are cached.

{metalist-0.11.0 → metalist-0.13.0}/app/api/note_requests.py

@@ -106,6 +106,18 @@ class SearchSuggestionInteractionRequest(TypedDict):
 class TabSearchInteractionRequest(TypedDict):
     searchQuery: Text
 
+@with_config(ConfigDict(strict=True))
+class DictationPasteRequest(TypedDict):
+    """Text pasted into the search bar or the tag bar, and the field's current value."""
+    text: Text
+    target: Literal["search", "tags", "tag", "condition"]
+    current_value: Text
+    # The spoken phrases around a quoted phrase (Dictation Settings).
+    quote_open: Identifier
+    quote_close: Identifier
+    # Spoken before a search term to exclude it (Dictation Settings).
+    negate_phrase: Identifier
+
 @with_config(ConfigDict(strict=True))
 class TagSuggestionsRequest(TypedDict):
     note_id: Identifier
@@ -136,6 +148,14 @@ class CreateChildRequest(TypedDict):
     undoContext: UndoContext
     viewport: Viewport
 
+@with_config(ConfigDict(strict=True))
+class CreateFilePreviewChildRequest(TypedDict):
+    fileId: Identifier
+    search_query: Text | None
+    clientId: Identifier
+    undoContext: UndoContext
+    viewport: Viewport
+
 @with_config(ConfigDict(strict=True))
 class UpdateNoteRequest(TypedDict):
     clientId: Identifier

{metalist-0.11.0 → metalist-0.13.0}/app/api/routes/ai.py

@@ -33,21 +33,17 @@ from app.services.agent.openai_inference import OpenAIInferenceAdapter
 from app.services.agent.openai_inference import validate_openai_model
 from app.services.agent.openai_cost_tracking import OpenAICostSnapshot
 from app.services.agent.openai_cost_tracking import openai_cost_tracker
-from app.services.agent.permissions import AgentPermissionPolicy
 from app.services.agent.prompt_settings import AgentPromptSet
 from app.services.agent.skill_settings import AgentSkillSet
 from app.services.agent.retrieval_settings import AgentRetrievalSettings
 from app.services.agent.scope import ScopedSearchSnapshot
 from app.services.agent.prompt_settings import DEFAULT_AGENT_PROMPTS
-from app.services.agent.prompt_settings import resolve_agent_prompt_set
 from app.services.agent.retrieval_settings import resolve_agent_retrieval_settings
 from app.services.agent.scope import AgentScopeDescriptor
 from app.services.agent.scope import scoped_search_snapshot_factory
 from app.services.agent.runtime import AgentRuntime
 from app.services.agent.skill_settings import DEFAULT_AGENT_SKILLS
-from app.services.agent.skill_settings import resolve_agent_skill_set
 from app.services.agent.token_estimation import estimate_input_tokens
-from app.services.agent.tools import read_only_agent_tools
 from app.services.agent.trace import agent_trace_store
 from app.services.agent.menu_actions import MenuResult, menu_action_store
 from app.services.client_state_service import load_client_preferences
@@ -62,7 +58,7 @@ from app.services.agent.retrieval_settings import resolve_tagging_batch_tokens
 from app.services.agent.web_settings import AgentWebSettings
 from app.services.agent.web_settings import resolve_agent_web_settings
 from app.services.agent.web_evidence import web_evidence_store
-from app.services.agent.tagging import TAGGING_PROMPT_KEY, TAGGING_POLICY_KEY, DEFAULT_TAGGING_PROMPT
+from app.services.agent.tagging import TAGGING_POLICY_KEY
 from app.usecases.bulk_tag_proposals import apply_bulk_proposals, prepare_proposal_changes
 from app.services.bulk_operation import bulk_operation_guard
 from app.services.tab_state import tab_state_store
@@ -79,8 +75,6 @@ def _agent_runtime(*, inference: InferenceAdapter) -> AgentRuntime:
         context_builder=agent_context_builder,
         inference=inference,
         model_policy=SingleModelPolicy(),
-        permission_policy=AgentPermissionPolicy(),
-        tool_registry=read_only_agent_tools,
         trace_store=agent_trace_store,
         provider_label=inference.provider_label,
     )
@@ -136,23 +130,6 @@ class CloudPrivacyPreviewResponse(BaseModel):
     hidden_note_ids: list[str]
 
 
-class AiSkillDefaultResponse(BaseModel):
-    skill_id: str
-    title: str
-    description: str
-    trigger_action: str
-    preference_key: str
-    content: str
-    superseded_preference_keys: list[str]
-
-
-class AiPromptDefaultsResponse(BaseModel):
-    system_prompt: str
-    final_response_prompt: str
-    tool_result_prompt: str
-    skills: list[AiSkillDefaultResponse]
-
-
 class AiChatRequest(BaseModel):
     model_config = ConfigDict(extra="forbid")
 
@@ -384,32 +361,6 @@ def reset_openai_cost(
     return _openai_cost_response(openai_cost_tracker.snapshot())
 
 
-@router.get("/prompts/defaults", response_model=AiPromptDefaultsResponse)
-def get_ai_prompt_defaults(
-    response: Response,
-    token: Annotated[str, Depends(require_request_auth_token)],
-) -> AiPromptDefaultsResponse:
-    del token
-    response.headers["Cache-Control"] = "no-store"
-    return AiPromptDefaultsResponse(
-        system_prompt=DEFAULT_AGENT_PROMPTS.system_prompt,
-        final_response_prompt=DEFAULT_AGENT_PROMPTS.final_response_prompt,
-        tool_result_prompt=DEFAULT_AGENT_PROMPTS.tool_result_prompt,
-        skills=[
-            AiSkillDefaultResponse(
-                skill_id=skill.skill_id,
-                title=skill.title,
-                description=skill.description,
-                trigger_action=skill.trigger_action,
-                preference_key=skill.preference_key,
-                content=skill.content,
-                superseded_preference_keys=list(skill.superseded_preference_keys),
-            )
-            for skill in DEFAULT_AGENT_SKILLS.skills
-        ],
-    )
-
-
 @router.post(
     "/cloud-privacy/preview",
     response_model=CloudPrivacyPreviewResponse,
@@ -637,8 +588,8 @@ def stream_ai_chat(
         session_key=session_key,
     )
     preferences = load_client_preferences(token=token)
-    prompts = resolve_agent_prompt_set(preferences=preferences)
-    skills = resolve_agent_skill_set(preferences=preferences)
+    prompts = DEFAULT_AGENT_PROMPTS
+    skills = DEFAULT_AGENT_SKILLS
     retrieval_settings = resolve_agent_retrieval_settings(
         preferences=preferences,
         provider=payload.provider,
@@ -691,6 +642,7 @@ def stream_ai_chat(
         model=payload.model,
     )
     provider_messages = ai_chat_store.provider_messages(session_key=session_key)
+    earlier_actions = ai_chat_store.earlier_actions(session_key=session_key)
     initial_messages = agent_context_builder.build_initial_messages(
         canonical_messages=provider_messages,
         prompts=prompts,
@@ -703,7 +655,7 @@ def stream_ai_chat(
     events = _stream_runtime_events(payload=payload, openai_api_key=openai_api_key,
         session_key=session_key, provider_messages=provider_messages, prompts=prompts, skills=skills,
         retrieval_settings=retrieval_settings, frozen_scope=frozen_scope, tagging_run=tagging_run,
-        web_settings=web_settings,
+        web_settings=web_settings, earlier_actions=earlier_actions,
         initial_input_tokens=initial_approx_input_tokens)
 
     return StreamingResponse(
@@ -722,6 +674,7 @@ async def _stream_runtime_events(
     provider_messages: list[dict[str, str]], prompts: AgentPromptSet, skills: AgentSkillSet,
     retrieval_settings: AgentRetrievalSettings, frozen_scope: ScopedSearchSnapshot,
     tagging_run: TaggingRun, initial_input_tokens: int, web_settings: AgentWebSettings,
+    earlier_actions: list[dict[str, object]],
 ) -> AsyncIterator[dict[str, object]]:
     action, start_label = 'provider_runtime', 'Connecting to OpenAI API'
     ready_label = 'OpenAI API ready · 1,050,000-token context'
@@ -731,11 +684,11 @@ async def _stream_runtime_events(
                approx_input_tokens=initial_input_tokens, output_tokens_received=0, duration_ms=0.0)
     yield dict(type='action_status', action=action, status='completed', label=ready_label,
                approx_input_tokens=initial_input_tokens, output_tokens_received=0, duration_ms=0.0)
-    events = runtime.stream_scoped(session_key=session_key, base_url=base_url,
+    events = runtime.stream_agent(session_key=session_key, base_url=base_url,
             selected_model=payload.model, thinking_level=payload.thinking_level,
             canonical_messages=provider_messages, prompts=prompts, skills=skills,
             retrieval_settings=retrieval_settings, web_settings=web_settings,
-            frozen_scope=frozen_scope, tag_handler=tagging_run.stream)
+            frozen_scope=frozen_scope, tagging_run=tagging_run, earlier_actions=earlier_actions)
     try:
         async for event in events:
             yield event
@@ -771,6 +724,8 @@ class BulkAnswerRequest(BaseModel):
         "prefix_focus_both",
         "summarize_all",
         "use_prefix",
+        "yes",
+        "no",
     ]
 
 
@@ -820,6 +775,4 @@ def manage_bulk_proposals(payload: BulkManageRequest, token: Annotated[str, Depe
 @router.get("/proposals/settings")
 def get_tagging_settings(token: Annotated[str, Depends(require_request_auth_token)]):
     preferences = load_client_preferences(token=token)
-    return {"policy": preferences.get(TAGGING_POLICY_KEY, ""),
-            "prompt": preferences.get(TAGGING_PROMPT_KEY, DEFAULT_TAGGING_PROMPT),
-            "default_prompt": DEFAULT_TAGGING_PROMPT}
+    return {"policy": preferences.get(TAGGING_POLICY_KEY, "")}

{metalist-0.11.0 → metalist-0.13.0}/app/api/routes/notes.py

@@ -20,9 +20,11 @@ from app.api.note_requests import (
     SearchSuggestionInteractionRequest,
     TabSearchInteractionRequest,
     TagSuggestionsRequest,
+    DictationPasteRequest,
     CreateNoteTopRequest,
     CreateSiblingRequest,
     CreateChildRequest,
+    CreateFilePreviewChildRequest,
     UpdateNoteRequest,
     SaveNoteRequest,
     AddSelectedTextTagRequest,
@@ -61,7 +63,9 @@ from app.services.exception_capture import CapturedExceptionContext
 from app.usecases.create_note import CmdCreateNote
 from app.usecases.base import QueryCommand
 from app.usecases.create_sibling import CmdCreateSibling
+from app.services.file_registry import file_registry
 from app.usecases.create_child import CmdCreateChild
+from app.usecases.create_file_preview_child import CmdCreateFilePreviewChild
 from app.usecases.update_content import CmdUpdateContent
 from app.usecases.prefetch_link_titles import CmdPrefetchLinkTitles
 from app.usecases.add_selected_text_tag import CmdAddSelectedTextTag
@@ -127,6 +131,7 @@ from app.services.html_export import build_notes_export_filename
 from app.services.note_fullscreen import build_note_fullscreen_markup
 from app.services.link_titles import link_title_store
 from app.services.search_index import search_index
+from app.services.dictation_cleanup import clean_dictated_paste
 from app.services.tag_suggestions import suggest_tags_for_note
 from app.services.undo_state import reset_undo_stack
 from app.usecases.prioritize import list_prioritize_tag_suggestions
@@ -257,12 +262,9 @@ def view_diff(payload: ViewDiffRequest):
     if normalized_search is not None:
         if not isinstance(normalized_search, str):
             raise HTTPException(status_code=400, detail="search must be a string or null")
-        capture = CapturedExceptionContext(InputRejected, boundary='app/api/routes/notes.py:view_diff:capture')
-        with capture:
-            parse_search_query(normalized_search)
-        if capture.captured_exception is not None:
-            exc = capture.captured_exception
-            raise HTTPException(status_code=400, detail=str(exc)) from exc
+        # Validated before anything changes; a search that does not parse is
+        # answered as rejected input (main.py), shown as a polite warning.
+        parse_search_query(normalized_search)
 
     maybe_reset_on_context(client_id, undo_context)
 
@@ -285,14 +287,19 @@ def view_diff(payload: ViewDiffRequest):
     if tab_view_empty:
         warm_view = None
     client_known_note_ids: set[str] = set()
+    # The tab's previous root order: a sorted tab keeps the edited note's root there.
+    previous_root_ids: list[str] = []
     if warm_view is not None:
         client_known_note_ids = set(warm_view.hash_by_id)
+        if None in warm_view.children_by_parent:
+            previous_root_ids = warm_view.children_by_parent[None]
 
     state = build_view_state(
         editing_note_id=normalized_editing_note_id,
         search=normalized_search,
         sort_mode=sort_mode,
         client_known_note_ids=client_known_note_ids,
+        previous_root_ids=previous_root_ids,
         visible_top_root_id=visible_top_root_id,
         visible_bottom_root_id=visible_bottom_root_id,
         is_untagged_view=is_untagged_view,
@@ -549,6 +556,22 @@ def delete_tag_interactions(request: Request) -> Dict[str, object]:
     return {"deletedCount": deleted_count}
 
 
+@router.post("/notes/dictation-paste")
+@transactional_route
+def dictation_paste(payload: DictationPasteRequest) -> Dict[str, object]:
+    """Read-only: the cleaned-up text to insert for a paste (docs/ui/dictation-paste.md)."""
+    text = clean_dictated_paste(
+        text=payload["text"],
+        target=payload["target"],
+        existing_tag_frequencies=search_index.list_explicit_tag_frequencies(),
+        current_value=payload["current_value"],
+        quote_open=payload["quote_open"],
+        quote_close=payload["quote_close"],
+        negate_phrase=payload["negate_phrase"],
+    )
+    return {"text": text}
+
+
 @router.post("/notes/tag-suggestions")
 @transactional_route
 def tag_suggestions(payload: TagSuggestionsRequest) -> Dict[str, object]:
@@ -748,6 +771,27 @@ def create_child(request: Request, note_id: str, body: CreateChildRequest):
     return cmd.execute()
 
 
+@router.post("/notes/file-preview-child/{note_id}")
+@transactional_route
+def create_file_preview_child(request: Request, note_id: str, body: CreateFilePreviewChildRequest):
+    token = _require_bearer_token(request)
+    viewport = _require_viewport(body)
+    _require_note_present(note_id, context="notes.file-preview-child")
+    file_id = body["fileId"]
+    if not file_registry.has_file(file_id):
+        raise HTTPException(status_code=404, detail=f"File not found: {file_id}")
+    cmd = CmdCreateFilePreviewChild(
+        parent_note_id=note_id,
+        file_id=file_id,
+        search_query=body["search_query"],
+        token=token,
+        client_id=body["clientId"],
+        undo_context=body["undoContext"],
+        viewport=viewport,
+    )
+    return cmd.execute()
+
+
 @router.put("/notes/{note_id}")
 @transactional_route
 def update_note(request: Request, note_id: str, body: UpdateNoteRequest):

{metalist-0.11.0 → metalist-0.13.0}/app/api/transactions.py

@@ -125,6 +125,8 @@ def assert_mutation_routes_wrapped(app: FastAPI) -> None:
 def _check_bulk_guard(endpoint_name: str) -> None:
     if endpoint_name in _STORAGE_TRANSITIONS and bulk_operation_guard.active_mutations > 1:
         raise HTTPException(status_code=409, detail="Wait for the current request to finish before changing namespace storage.")
-    permitted = {"answer_bulk_question", "preview_cloud_privacy", "put_ai_debug_details"}
+    # Preferences never touch notes, and every operation reads them once when it
+    # starts, so saving them (e.g. the chat panel width) cannot disturb it.
+    permitted = {"answer_bulk_question", "preview_cloud_privacy", "put_ai_debug_details", "put_client_preferences"}
     if bulk_operation_guard.operation_id and endpoint_name not in permitted:
         raise HTTPException(status_code=409, detail="A bulk operation is running. Wait or cancel it before changing the app.")

{metalist-0.11.0 → metalist-0.13.0}/app/exception_boundaries.py

@@ -14,7 +14,9 @@ CAPTURE_BOUNDARIES = {
     'app/services/agent/web_fetch.py:_extract_pdf_content:pdf_capture': ('PyPdfError',),
     'app/services/agent/web_fetch.py:_decode_web_text:encoding_capture': ('LookupError',),
     'app/services/agent/web_fetch.py:handle_starttag:join_capture': ('ValueError',),
-    'app/services/agent/runtime.py:_open_web_action:evidence_capacity_capture': ('WebEvidenceCapacityError',),
+    'app/services/agent/web_actions.py:open_web_pages:evidence_capacity_capture': ('WebEvidenceCapacityError',),
+    # A model's tool call is untrusted input; rejected arguments are explained back to the model.
+    'app/services/agent/agent_tools.py:parse_tool_call:arguments_capture': ('ValidationError',),
     # In-app updates translate only external transport, installation eligibility and job lookup errors.
     'app/services/app_updates.py:check_for_update:eligibility': ('AppUpdateRejected',),
     'app/services/app_updates.py:job_path:validation': ('ValueError',),

{metalist-0.11.0 → metalist-0.13.0}/app/main.py

@@ -86,7 +86,7 @@ from urllib.parse import parse_qsl
 from urllib.parse import urlencode
 from urllib.parse import urlsplit
 from urllib.parse import urlunsplit
-from app.services.input_errors import NamespaceInputRejected
+from app.services.input_errors import InputRejected, NamespaceInputRejected
 
 logger.remove()
 logger.add(
@@ -158,6 +158,13 @@ async def request_validation_error(request: Request, exc: RequestValidationError
     return JSONResponse(status_code=422, content={"detail": summarize_validation_errors(exc.errors())})
 
 
+@app.exception_handler(InputRejected)
+async def handle_input_rejected(request: Request, exc: InputRejected):
+    # Expected, validated user-input failures (e.g. a search that does not
+    # parse): a polite 400 the browser shows as a warning, never a 500.
+    return JSONResponse(status_code=400, content={"detail": str(exc), "inputRejected": True})
+
+
 @app.exception_handler(OntologyParseError)
 async def handle_ontology_parse_error(request: Request, exc: OntologyParseError):
     return JSONResponse(status_code=400, content={"detail": str(exc)})

{metalist-0.11.0 → metalist-0.13.0}/app/models/utils.py

@@ -17,6 +17,7 @@ from ..services.content_cache import (
 )
 from ..services.note_store import store as note_store
 from app.presentation.render.note_renderer import render_read_only_mode
+from app.services.markdown_rendering import strip_latex_source_attributes
 from app.db.notes_sql import (
     fetch_children_ordered,
     fetch_note,
@@ -652,7 +653,9 @@ def render_note_data_read_only(note_data: Dict[str, Any]) -> Dict[str, Any]:
             raise TypeError('note_data.children must be a list')
         child_payloads = [{} for _ in children]
         rendered.update(node)
-        rendered['content'] = render_read_only_mode(SimpleNamespace(content=content, tags=tags))
+        rendered['content'] = strip_latex_source_attributes(
+            render_read_only_mode(SimpleNamespace(content=content, tags=tags))
+        )
         rendered['children'] = child_payloads
         pending.extend(zip(children, child_payloads))
     return root

metalist-0.13.0/app/services/agent/action_log.py

@@ -0,0 +1,59 @@
+"""The agent's own record of what it did for an answer.
+
+Each tool call becomes one short entry: what was asked and what came of it,
+never the content a tool returned (page text and notes can be huge, and the
+model can read them again). The entries are kept with the answer for the rest of
+the chat and given to the model on later turns, so a question such as "did you
+search the web for that?" is answered from what actually happened. Only the
+model sees them.
+"""
+
+from __future__ import annotations
+
+import json
+
+
+_NOTE_READ_TOOLS = frozenset({"read_view_notes", "search_view_notes"})
+_OPERATION_TOOLS = frozenset({"summarize_view", "propose_tag_generation", "propose_tag_review"})
+_KNOWN_TOOLS = _NOTE_READ_TOOLS | _OPERATION_TOOLS | {
+    "lookup_metalist_help", "view_overview", "open_web_pages", "open_menu",
+}
+
+
+def describe_action(*, name: str, arguments: str, result_content: str, is_error: bool) -> dict[str, object]:
+    """One log entry for a tool call, from its JSON arguments and JSON result."""
+    # A rejected call may name a tool that does not exist: it is still recorded.
+    if is_error:
+        return {"tool": name, "rejected": True}
+    if name not in _KNOWN_TOOLS:
+        raise ValueError(f"No action log entry for unknown tool: {name!r}")
+    parsed_arguments = json.loads(arguments)
+    assert isinstance(parsed_arguments, dict), "Tool arguments are a JSON object"
+    if name in _OPERATION_TOOLS:
+        return {"tool": name, "arguments": parsed_arguments}
+    result = json.loads(result_content)
+    assert isinstance(result, dict), "Tool results are a JSON object"
+    if name == "lookup_metalist_help":
+        return {"tool": name, "topics": list(parsed_arguments["topics"])}
+    if name == "view_overview":
+        return {"tool": name, "notes_in_view": result["note_count"], "trees_in_view": result["tree_count"]}
+    if name == "open_menu":
+        return {"tool": name, "menu": result["label"], "status": result["status"]}
+    if name == "open_web_pages":
+        return {"tool": name, "pages": [_page_entry(page) for page in result["pages"]]}
+    entry: dict[str, object] = {"tool": name}
+    if name == "search_view_notes":
+        entry["query"] = parsed_arguments["query"]
+        entry["matching_notes"] = result["matching_note_count"]
+    entry["trees_read"] = len(result["trees"])
+    entry["trees_left_unread"] = len(result["unread_root_ids"])
+    entry["trees_too_large"] = len(result["too_large_root_ids"])
+    return entry
+
+
+def _page_entry(page: dict[str, object]) -> dict[str, object]:
+    """The address as asked, whether it opened, and its title when it did."""
+    title = ""
+    if page["status"] == "ok":
+        title = page["title"]
+    return {"url": page["requested_url"], "status": page["status"], "title": title}