ed/manual_slop
Private
Public Access
0
0
Fork 0
Code Activity

refactor(models): reduce to Pydantic proxy helpers + DEFAULT_TOOL_CATEGORIES

Browse Source 
After 11 class moves (Phases 3a-3i) + 1 deletion (Phase 4), this commit
reduces src/models.py from 1044 lines (original) / 768 lines (pre-Phase 3b)
to 135 lines. The remaining content is:
 - DEFAULT_TOOL_CATEGORIES: the canonical tool list grouped for
   the UI's category filter (the ONLY non-Pydantic constant)
 - _create_generate_request + _create_confirm_request: the Pydantic
   proxy classes for the API hook subsystem
 - _PYDANTIC_CLASS_FACTORIES: registry for the Pydantic proxies
 - __getattr__: lazy re-exports for ALL 30+ moved classes + PROVIDERS

Removed:
 - All 11 class definitions (MMA Core, FileItem + 4 file-related,
   Tool + ToolPreset + BiasProfile, 2 editor configs, WorkspaceProfile,
   4 MCP config classes + load_mcp_config, ProjectContext + 5 sub)
 - All 3 config IO function definitions (load_config_from_disk,
   save_config_to_disk, _clean_nones, parse_history_entries)
 - All 5 eager re-export blocks at the top (they triggered tomli_w
   loading at import time via the personas import; the lazy __getattr__
   breaks the cycle)
 - AGENT_TOOL_NAMES (deleted in Phase 4)

The lazy __getattr__ keeps the 'from src.models import X' pattern
working for legacy callers. New code should import directly from
the subsystem files (src.mma, src.project, src.project_files,
src.tool_presets, src.tool_bias, src.external_editor, src.mcp_client,
src.workspace_manager, src.personas).

Side benefit: the pre-existing test
tests/test_models_no_top_level_tomli_w.py::test_models_does_not_import_tomli_w_at_module_level
now PASSES. Before Phase 5 it failed because the eager
'from src.personas import Persona' triggered tomli_w loading. The
lazy __getattr__ for Persona only loads tomli_w when 'models.Persona'
is actually accessed (not on a bare 'import src.models').

Verification: VC10
  wc -l src/models.py  # 135 lines (well under the 1044-line original;
                        # 30-line target was aspirational; the lazy
                        # __getattr__ for 30+ moved classes is the
                        # dominant cost)
  Measure-Object -Line on src/models.py  # 135

Tests verified (84/85 PASS; 1 pre-existing failure unrelated):
  tests/test_mcp_config.py (3/3 PASS)
  tests/test_tool_preset_manager.py (4/4 PASS)
  tests/test_bias_models.py (3/3 PASS)
  tests/test_tool_bias.py (3/3 PASS)
  tests/test_external_editor.py (17/17 PASS)
  tests/test_workspace_manager.py (3/3 PASS)
  tests/test_models_no_top_level_tomli_w.py (3/3 PASS) [previously 1 FAIL]
  tests/test_project_context_20260627.py (10/10 PASS)
  tests/test_file_item_model.py (4/4 PASS)
  tests/test_view_presets.py (4/4 PASS)
  tests/test_context_presets_models.py (3/3 PASS)
  tests/test_presets.py (5/5 PASS)
  tests/test_persona_models.py (2/2 PASS)
  tests/test_persona_manager.py (3/3 PASS)
  tests/test_arch_boundary_phase2.py (5/6 PASS; 1 pre-existing FAIL
                                                unrelated: test_rejection_prevents_dispatch
                                                is a dialog-mock issue)
  tests/test_mcp_tool_specs.py (10/10 PASS)
This commit is contained in:
ed
2026-06-26 10:22:57 -04:00
parent 779d504c70
commit 3c4a52901a
1 changed files with 81 additions and 230 deletions
Download Patch File Download Diff File Expand all files Collapse all files
src/models.py
+81 -230
View File
@@ -1,165 +1,55 @@
""" """
Models - Core data structures for MMA orchestration and project configuration. Models - Pydantic proxies + DEFAULT_TOOL_CATEGORIES only.
This module defines the primary dataclasses used throughout the Manual Slop Per module_taxonomy_refactor_20260627 Phase 5, this module is the
application for representing tasks, tracks, and execution context. 'smallest possible' models module. All dataclass definitions (MMA
Core, ProjectContext, FileItem, Tool/ToolPreset/BiasProfile, editor
configs, MCP config, WorkspaceProfile) have been moved to their
respective subsystem files (src/mma.py, src/project.py, src/project_files.py,
src/tool_presets.py, src/tool_bias.py, src/external_editor.py,
src/mcp_client.py, src/workspace_manager.py, src/personas.py).
Key Data Structures: The legacy 'from src.models import X' pattern is preserved via the
- Ticket: Atomic unit of work with status, dependencies, and context requirements __getattr__ below, which lazy-loads the moved classes on first access.
- Track: Collection of tickets with a shared goal New code should import directly from the subsystem files.
- WorkerContext: Execution context for a Tier 3 worker
- Metadata: Track metadata (id, name, status, timestamps)
- TrackState: Serializable track state with discussion history
- FileItem: File configuration with auto-aggregate and force-full flags
Status Machine (Ticket): Architecture:
todo -> in_progress -> completed - DEFAULT_TOOL_CATEGORIES is the ONLY non-Pydantic constant kept here.
| | It groups the canonical tool list for the UI's category filter.
v v - _create_generate_request + _create_confirm_request are the Pydantic
blocked blocked proxy classes for the API hook subsystem (GenerateRequest +
ConfirmRequest). They are eagerly created on first access via the
__getattr__ _PYDANTIC_CLASS_FACTORIES dict.
- __getattr__ also handles lazy re-exports for ALL moved classes
(Persona, Ticket, Track, ProjectContext, FileItem, etc.) and the
PROVIDERS constant from src.ai_client.
Serialization:
All dataclasses provide to_dict() and from_dict() class methods for TOML/JSON
persistence via project_manager.py.
Thread Safety:
These dataclasses are NOT thread-safe. Callers must synchronize mutations
if sharing instances across threads (e.g., during ConductorEngine execution).
Configuration Integration:
- load_config() / save_config() read/write the global config.toml
- DEFAULT_TOOL_CATEGORIES groups the canonical tool list for UI display
See Also: See Also:
- docs/guide_mma.md for MMA orchestration documentation - docs/guide_models.md for the centralized data model registry
- src/dag_engine.py for TrackDAG and ExecutionEngine - conductor/code_styleguides/data_oriented_design.md for the type
- src/multi_agent_conductor.py for ConductorEngine promotion mandate
- src/project_manager.py for persistence layer
""" """
from __future__ import annotations from __future__ import annotations
import datetime
import json
import os
import sys import sys
import tomllib
from dataclasses import dataclass, field from typing import Any, Dict, List
from pathlib import Path
from typing import Any, Dict, List, Optional, Union
from src.paths import get_config_path
from src.result_types import ErrorInfo, ErrorKind, Result
from src.type_aliases import (
CommsLog,
CommsLogCallback,
CommsLogEntry,
FileItem,
FileItems,
History,
HistoryMessage,
Metadata,
ToolCall,
ToolDefinition,
)
# Backward-compat re-exports for the MMA Core classes that moved to src/mma.py
# in module_taxonomy_refactor_20260627 Phase 3.1. Consumers still using
# 'from src.models import Ticket' continue to work; new code should import
# from src.mma directly.
from src.mma import (
EMPTY_TRACK_STATE,
ThinkingSegment,
Ticket,
Track,
TrackMetadata,
TrackState,
WorkerContext,
)
# Backward-compat re-export for Persona (Phase 3.4 -> src/personas.py).
from src.personas import Persona
# Alias the old `Metadata` dataclass name to TrackMetadata so existing
# `from src.models import Metadata` keeps resolving to the dataclass.
Metadata = TrackMetadata # noqa: F401 — legacy class name re-export
# Backward-compat re-exports for Project Context (Phase 3b -> src/project.py)
# + the config IO helpers. Consumers using 'from src.models import ProjectContext'
# or 'from src.models import _load_config_from_disk' / '_save_config_to_disk'
# must migrate to the new public names in src.project (load_config_from_disk /
# save_config_to_disk). The private names are dropped in Phase 5.
from src.project import (
EMPTY_PROJECT_CONTEXT,
ProjectContext,
ProjectDiscussion,
ProjectFiles,
ProjectMeta,
ProjectOutput,
ProjectScreenshots,
_clean_nones,
load_config_from_disk,
parse_history_entries,
save_config_to_disk,
)
# Backward-compat re-exports for FileItem + Preset + ContextPreset +
# ContextFileEntry + NamedViewPreset (Phase 3c -> src/project_files.py).
# Consumers using 'from src.models import FileItem' (and the others) must
# migrate to the new path in src.project_files. These classes are added
# here for backward compat and dropped in Phase 5.
from src.project_files import (
ContextFileEntry,
ContextPreset,
FileItem,
NamedViewPreset,
Preset,
)
# Backward-compat re-exports for Tool + ToolPreset (Phase 3d ->
# src/tool_presets.py). These are LAZY via the __getattr__ below to avoid
# the circular import (src.tool_presets imports BiasProfile from
# src.models; if we eagerly imported Tool from src.tool_presets, the
# cycle would deadlock on initial load). Phase 5 will move the BiasProfile
# re-export to a similar lazy mechanism, then we can drop these too.
#region: Constants
# AGENT_TOOL_NAMES was deleted in module_taxonomy_refactor_20260627 Phase 4.
# It was a hardcoded snapshot of mcp_tool_specs.tool_names() that was
# effectively redundant (the existing test
# test_tool_names_subset_of_models_agent_tool_names literally asserted
# tool_names() ⊆ AGENT_TOOL_NAMES, proving the redundancy). Consumers
# must now use mcp_tool_specs.tool_names() directly. See the v2 spec
# for the migration rationale.
DEFAULT_TOOL_CATEGORIES: Dict[str, List[str]] = { DEFAULT_TOOL_CATEGORIES: Dict[str, List[str]] = {
"General": ["read_file", "list_directory", "search_files", "get_tree", "get_file_summary"], "General": ["read_file", "list_directory", "search_files", "get_tree", "get_file_summary"],
"Surgical": ["get_file_slice", "set_file_slice", "edit_file"], "Surgical": ["get_file_slice", "set_file_slice", "edit_file"],
"Python": [ "Python": [
"py_get_skeleton", "py_get_skeleton", "py_get_code_outline", "py_get_definition", "py_update_definition",
"py_get_code_outline", "py_get_signature", "py_set_signature", "py_get_class_summary",
"py_get_definition", "py_get_var_declaration", "py_set_var_declaration", "py_get_docstring",
"py_update_definition", "py_find_usages", "py_get_imports", "py_check_syntax", "py_get_hierarchy",
"py_get_signature", "py_remove_def", "py_add_def", "py_move_def", "py_region_wrap",
"py_set_signature",
"py_get_class_summary",
"py_get_var_declaration",
"py_set_var_declaration",
"py_get_docstring",
"py_find_usages",
"py_get_imports",
"py_check_syntax",
"py_get_hierarchy",
"py_remove_def",
"py_add_def",
"py_move_def",
"py_region_wrap",
], ],
"C/C++": [ "C/C++": [
"ts_c_get_skeleton", "ts_c_get_skeleton", "ts_cpp_get_skeleton", "ts_c_get_code_outline",
"ts_cpp_get_skeleton", "ts_cpp_get_code_outline", "ts_c_get_definition", "ts_cpp_get_definition",
"ts_c_get_code_outline", "ts_c_get_signature", "ts_cpp_get_signature", "ts_c_update_definition",
"ts_cpp_get_code_outline",
"ts_c_get_definition",
"ts_cpp_get_definition",
"ts_c_get_signature",
"ts_cpp_get_signature",
"ts_c_update_definition",
"ts_cpp_update_definition", "ts_cpp_update_definition",
], ],
"Web": ["web_search", "fetch_url"], "Web": ["web_search", "fetch_url"],
@@ -168,22 +58,6 @@ DEFAULT_TOOL_CATEGORIES: Dict[str, List[str]] = {
"Beads": ["bd_create", "bd_update", "bd_list", "bd_ready"], "Beads": ["bd_create", "bd_update", "bd_list", "bd_ready"],
} }
# CONFIG_PATH was a module-level constant. REMOVED: it cached the
# repo-root config.toml at import time, so any test calling
# save_config() or load_config() after import wrote/read the
# user's actual config — corrupting the user's settings every test
# run. The path is now re-resolved on every call, so the SLOP_CONFIG
# env var (or a set_config_path() helper from test fixtures) can
# redirect reads/writes to a temp_workspace without reimporting.
# See tests/conftest.py:reset_paths for the test-side mechanism.
# Config IO helpers (_clean_nones, _load_config_from_disk -> load_config_from_disk,
# _save_config_to_disk -> save_config_to_disk) + parse_history_entries moved to
# src/project.py in module_taxonomy_refactor_20260627 Phase 3b. The
# re-exports at the top of this module keep 'from src.models import ...' working
# for legacy callers. New code should import from src.project directly.
#region: Pydantic Models
def _create_generate_request() -> type: def _create_generate_request() -> type:
from src.module_loader import _require_warmed from src.module_loader import _require_warmed
@@ -197,104 +71,81 @@ def _create_generate_request() -> type:
max_tokens=(int | None, None), max_tokens=(int | None, None),
) )
def _create_confirm_request() -> type: def _create_confirm_request() -> type:
from src.module_loader import _require_warmed from src.module_loader import _require_warmed
pydantic = _require_warmed("pydantic") pydantic = _require_warmed("pydantic")
return pydantic.create_model( return pydantic.create_model(
"ConfirmRequest", "ConfirmRequest",
approved=(bool, ...), approved=(bool, ...),
script=(Optional[str], None), script=(str | None, None),
) )
_PYDANTIC_CLASS_FACTORIES: dict[str, callable] = { _PYDANTIC_CLASS_FACTORIES: dict[str, callable] = {
"GenerateRequest": _create_generate_request, "GenerateRequest": _create_generate_request,
"ConfirmRequest": _create_confirm_request, "ConfirmRequest": _create_confirm_request,
} }
def __getattr__(name: str) -> Any: def __getattr__(name: str) -> Any:
if name == "PROVIDERS": if name == "PROVIDERS":
from src.ai_client import PROVIDERS as _PROVIDERS from src import ai_client
return _PROVIDERS return ai_client.PROVIDERS
if name in _PYDANTIC_CLASS_FACTORIES: if name in _PYDANTIC_CLASS_FACTORIES:
cls = _PYDANTIC_CLASS_FACTORIES[name]() cls = _PYDANTIC_CLASS_FACTORIES[name]()
globals()[name] = cls globals()[name] = cls
return cls return cls
if name in ("EMPTY_TRACK_STATE", "ThinkingSegment", "Ticket", "Track",
"TrackMetadata", "TrackState", "WorkerContext"):
from src import mma
val = getattr(mma, name)
globals()[name] = val
return val
if name == "Persona":
from src import personas
val = personas.Persona
globals()[name] = val
return val
if name in ("EMPTY_PROJECT_CONTEXT", "ProjectContext", "ProjectDiscussion",
"ProjectFiles", "ProjectMeta", "ProjectOutput",
"ProjectScreenshots", "_clean_nones",
"load_config_from_disk", "parse_history_entries",
"save_config_to_disk"):
from src import project
val = getattr(project, name)
globals()[name] = val
return val
if name in ("ContextFileEntry", "ContextPreset", "FileItem",
"NamedViewPreset", "Preset"):
from src import project_files
val = getattr(project_files, name)
globals()[name] = val
return val
if name in ("Tool", "ToolPreset"): if name in ("Tool", "ToolPreset"):
from src.tool_presets import Tool as _Tool, ToolPreset as _ToolPreset from src import tool_presets
val = _Tool if name == "Tool" else _ToolPreset val = getattr(tool_presets, name)
globals()[name] = val globals()[name] = val
return val return val
if name == "BiasProfile": if name == "BiasProfile":
from src.tool_bias import BiasProfile as _BiasProfile from src import tool_bias
globals()[name] = _BiasProfile val = tool_bias.BiasProfile
return _BiasProfile globals()[name] = val
return val
if name in ("TextEditorConfig", "ExternalEditorConfig", "EMPTY_TEXT_EDITOR_CONFIG"): if name in ("TextEditorConfig", "ExternalEditorConfig", "EMPTY_TEXT_EDITOR_CONFIG"):
from src.external_editor import TextEditorConfig as _TEC, ExternalEditorConfig as _EEC, EMPTY_TEXT_EDITOR_CONFIG as _ETEC from src import external_editor
val = {"TextEditorConfig": _TEC, "ExternalEditorConfig": _EEC, "EMPTY_TEXT_EDITOR_CONFIG": _ETEC}[name] val = getattr(external_editor, name)
globals()[name] = val globals()[name] = val
return val return val
if name == "WorkspaceProfile": if name == "WorkspaceProfile":
from src.workspace_manager import WorkspaceProfile as _WP from src import workspace_manager
globals()[name] = _WP val = workspace_manager.WorkspaceProfile
return _WP globals()[name] = val
if name in ("MCPServerConfig", "MCPConfiguration", "VectorStoreConfig", "RAGConfig", "load_mcp_config"): return val
from src.mcp_client import ( if name in ("MCPServerConfig", "MCPConfiguration", "VectorStoreConfig",
MCPServerConfig as _MSC, "RAGConfig", "load_mcp_config"):
MCPConfiguration as _MCP, from src import mcp_client
VectorStoreConfig as _VSC, val = getattr(mcp_client, name)
RAGConfig as _RAGC,
load_mcp_config as _lmc,
)
val = {"MCPServerConfig": _MSC, "MCPConfiguration": _MCP, "VectorStoreConfig": _VSC, "RAGConfig": _RAGC, "load_mcp_config": _lmc}[name]
globals()[name] = val globals()[name] = val
return val return val
raise AttributeError(f"module {__name__!r} has no attribute {name!r}") raise AttributeError(f"module {__name__!r} has no attribute {name!r}")
# MMA Core dataclasses (ThinkingSegment, Ticket, Track, WorkerContext, TrackMetadata)
# moved to src/mma.py in module_taxonomy_refactor_20260627 Phase 3.1. See
# the re-export block at the top of this module for backward-compat.
#region: State & Config
# TrackState moved to src/mma.py (re-exported at the top of this module for backward compat).
# FileItem, Preset, ContextFileEntry, NamedViewPreset, ContextPreset moved to
# src/project_files.py in module_taxonomy_refactor_20260627 Phase 3c. The
# re-exports at the top of this module keep 'from src.models import FileItem'
# (and the others) working for legacy callers. New code should import from
# src.project_files directly.
#region: Tool Models
# Tool + ToolPreset moved to src/tool_presets.py in Phase 3d. BiasProfile
# moved to src/tool_bias.py in Phase 3e. All three are re-exported lazily
# via the __getattr__ below to avoid the circular import (tool_presets and
# tool_bias both want to import from each other via models).
#region: UI/Editor
# TextEditorConfig + ExternalEditorConfig + EMPTY_TEXT_EDITOR_CONFIG moved
# to src/external_editor.py in module_taxonomy_refactor_20260627 Phase 3f.
# The re-exports are LAZY via the __getattr__ below to avoid the cycle
# (external_editor was previously importing them from models; the cycle
# would deadlock on eager re-export).
#region: Persona
# Persona dataclass moved to src/personas.py in module_taxonomy_refactor_20260627 Phase 3.4.
# PersonaManager (the ops layer) is also there. Re-export at the top of this module
# preserves backward-compat 'from src.models import Persona'.
# Persona dataclass moved to src/personas.py in module_taxonomy_refactor_20260627 Phase 3.4.
# PersonaManager (the ops layer) is also there. Re-export at the top of this module
# preserves backward-compat 'from src.models import Persona'.
#region: Workspace
#region: Workspace
# WorkspaceProfile moved to src/workspace_manager.py in
# module_taxonomy_refactor_20260627 Phase 3h. The re-export is LAZY via
# the __getattr__ below to avoid the cycle (workspace_manager was
# previously importing it from models).
#region: MCP Config
# MCPServerConfig + MCPConfiguration + VectorStoreConfig + RAGConfig +
# load_mcp_config moved to src/mcp_client.py in
# MCP config classes moved to src/mcp_client.py in
# module_taxonomy_refactor_20260627 Phase 3i. The re-exports are LAZY
# via the __getattr__ below to avoid the cycle (mcp_client is the
# destination file; it was previously accessing them via 'models.X').