feat 007: major refactor for future maintainability

.github/agents/copilot-instructions.md

@@ -20,6 +20,8 @@
 - Browser localStorage for per-user built-in agent overrides; existing `config/agents.yaml` remains canonical shared standard profile source; backend in-memory sessions unchanged (005-baked-agent-customization)
 - TypeScript 5.9 frontend; React 19; Python 3.12.6 backend unchanged + React, Vite 8, existing frontend API client, existing FastAPI skill endpoints, Playwright for visual verification (006-admin-skills-design-match)
 - Existing filesystem-backed skills under `skills/<name>/SKILL.md`; no new storage (006-admin-skills-design-match)
+- Python 3.12.6, TypeScript 5.9, React 19 + FastAPI, agent-framework-core/openai/azure-ai-search, Azure SDKs, React, Vite, react-markdown, MSAL (007-application-refactor)
+- In-memory backend sessions, file-backed `skills/`, browser localStorage, Terraform-managed Azure App Service resources (007-application-refactor)
 
 ## Project Structure
 
@@ -52,6 +54,6 @@ uv run pytest                      # Run tests
 - Package manager: uv only (never pip)
 
 ## Recent Changes
+- 007-application-refactor: Added Python 3.12.6, TypeScript 5.9, React 19 + FastAPI, agent-framework-core/openai/azure-ai-search, Azure SDKs, React, Vite, react-markdown, MSAL
 - 006-admin-skills-design-match: Added TypeScript 5.9 frontend; React 19; Python 3.12.6 backend unchanged + React, Vite 8, existing frontend API client, existing FastAPI skill endpoints, Playwright for visual verification
 - 005-baked-agent-customization: Added Python 3.12.6 backend; TypeScript 5.9 frontend; React 19; Vite 8 + FastAPI, agent-framework-core/openai/azure-ai-search, React, MSAL, localStorage APIs
-- 004-azd-deploy-frontend: Added Python 3.12 (backend), TypeScript (frontend) + FastAPI, React/Vite, Terraform (azurerm ~> 4.0)

.gitignore

@@ -8,9 +8,12 @@ wheels/
 
 # Virtual environments
 .venv
+.venv/
+venv/
 
 .env
 .env*
+*.log
 
 .ruff*
 
@@ -34,6 +37,8 @@ terraform.rc
 
 .mypy_cache/
 .pytest_cache/
+coverage/
+frontend/coverage/
 # Certificates
 certs/
 .env.appsettings.json

agent_factory.py

@@ -3,7 +3,7 @@
 import re
 from dataclasses import dataclass
 from pathlib import Path
-from typing import Any, Sequence, overload
+from typing import Any, Sequence
 
 from agent_framework import CompactionProvider, InMemoryHistoryProvider, SkillsProvider
 from agent_framework import Agent as RuntimeAgent
@@ -19,20 +19,10 @@
 from dotenv import load_dotenv
 
 
-from pydantic import BaseModel
-
 from prompt_config import load_agent_profile
 from mcp_servers import get_search_context_provider
 
 
-class AgentBase(BaseModel):
-    name: str
-    instructions: str
-    description: str
-    token_budget: int = 16_000
-    tools: list[str] | None = None
-
-
 @dataclass(frozen=True)
 class ChatRuntime:
     agent: RuntimeAgent
@@ -151,113 +141,6 @@ def _build_context_providers(
     return providers
 
 
-def _create_agent(
-    *,
-    name: str,
-    instructions: str,
-    description: str,
-    token_budget: int,
-    temperature: float,
-    tools: Sequence[Any] | None = None,
-    client: OpenAIChatClient | None = None,
-    summarizer_client: OpenAIChatClient | None = None,
-    logical_profile: str | None = None,
-    enable_search_context: bool = False,
-    skill_names: list[str] | None = None,
-) -> RuntimeAgent:
-    runtime_client = client or OpenAIChatClient()
-
-    return runtime_client.as_agent(
-        name=_sanitize_agent_name(name),
-        instructions=instructions,
-        description=description,
-        tools=tools,
-        default_options={"temperature": temperature},
-        context_providers=_build_context_providers(
-            token_budget=token_budget,
-            summarizer_client=summarizer_client,
-            logical_profile=logical_profile,
-            enable_search_context=enable_search_context,
-            skill_names=skill_names,
-        ),
-    )
-
-
-@overload
-def spawn_agent(agent: AgentBase, /) -> RuntimeAgent: ...
-
-
-@overload
-def spawn_agent(
-    *,
-    name: str,
-    instructions: str,
-    description: str,
-    token_budget: int = 16_000,
-    temperature: float | None = None,
-    tools: Sequence[Any] | None = None,
-    client: OpenAIChatClient | None = None,
-    summarizer_client: OpenAIChatClient | None = None,
-    logical_profile: str | None = None,
-    enable_search_context: bool = False,
-    skill_names: list[str] | None = None,
-) -> RuntimeAgent: ...
-
-
-def spawn_agent(
-    agent: AgentBase | None = None,
-    /,
-    *,
-    name: str | None = None,
-    instructions: str | None = None,
-    description: str | None = None,
-    token_budget: int = 16_000,
-    temperature: float | None = None,
-    tools: Sequence[Any] | None = None,
-    client: OpenAIChatClient | None = None,
-    summarizer_client: OpenAIChatClient | None = None,
-    logical_profile: str | None = None,
-    enable_search_context: bool = False,
-    skill_names: list[str] | None = None,
-) -> RuntimeAgent:
-    if agent is not None:
-        if any(value is not None for value in (name, instructions, description)):
-            raise TypeError("Pass either an agent model or expanded agent fields, not both.")
-
-        return _create_agent(
-            name=agent.name,
-            instructions=agent.instructions,
-            description=agent.description,
-            token_budget=agent.token_budget,
-            temperature=temperature if temperature is not None else _get_default_temperature(),
-            tools=tools,
-            client=client,
-            summarizer_client=summarizer_client,
-            logical_profile=logical_profile,
-            enable_search_context=enable_search_context,
-            skill_names=skill_names,
-        )
-
-    if name is None or instructions is None or description is None:
-        raise TypeError(
-            "spawn_agent() requires either an agent model or name, instructions, and description keywords."
-        )
-
-    return _create_agent(
-        name=name,
-        instructions=instructions,
-        description=description,
-        token_budget=token_budget,
-        temperature=temperature if temperature is not None else _get_default_temperature(),
-        tools=tools,
-        client=client,
-        summarizer_client=summarizer_client,
-        logical_profile=logical_profile,
-        enable_search_context=enable_search_context,
-        skill_names=skill_names,
-    )
-
-
 def _build_openai_clients() -> tuple[OpenAIChatClient, OpenAIChatClient]:
     """Create primary and summarizer OpenAI clients from environment variables.
 
@@ -344,18 +227,19 @@ def create_chat_runtime(
 
     resolved_temperature = temperature if temperature is not None else _get_default_temperature()
 
-    agent = spawn_agent(
-        client=primary_client,
-        name=agent_name,
+    agent = primary_client.as_agent(
+        name=_sanitize_agent_name(agent_name),
         instructions=runtime_instructions,
         description=description,
-        token_budget=token_budget,
-        temperature=resolved_temperature,
         tools=all_tools,
-        summarizer_client=summarizer_client,
-        logical_profile=logical_profile,
-        enable_search_context=enable_search_context,
-        skill_names=skill_names,
+        default_options={"temperature": resolved_temperature},
+        context_providers=_build_context_providers(
+            token_budget=token_budget,
+            summarizer_client=summarizer_client,
+            logical_profile=logical_profile,
+            enable_search_context=enable_search_context,
+            skill_names=skill_names,
+        ),
     )
 
     logger.info(

frontend/package.json

@@ -5,6 +5,7 @@
   "type": "module",
   "scripts": {
     "dev": "vite",
+    "test": "tsc -b --pretty false",
     "build": "tsc -b && vite build",
     "lint": "eslint .",
     "preview": "vite preview"