Replace pyright with ty, add more Ruff rules, fix issues

.github/workflows/ci.yml

@@ -20,8 +20,8 @@ jobs:
 
       - run: uv run ruff format --check src tests examples
       - run: uv run ruff check src tests examples
-      - run: uv run mypy src tests
-      - run: uv run pyright src tests
+      - run: uv run mypy
+      - run: uv run ty check
 
       - run: uv run pytest
 

CLAUDE.md

@@ -6,8 +6,8 @@
 2. after making changes run format, lint, and typecheck like ci:
     - uv run ruff format --check src tests examples
     - uv run ruff check src tests examples
-    - uv run mypy src tests
-    - uv run pyright src tests
+    - uv run mypy
+    - uv run ty check
 3. imports:
    - import by module, using the shortest unambiguous relative path. `from ..core import helpers`, `from . import streaming`
    - UNLESS it's `typing` — then `from typing import Foo` (there are too many of them).
@@ -37,4 +37,3 @@ ensure state is easy to serialize and deserialize, modify, and compose at any le
 move normalization and translation complexity inside the framework and keep the public data model minimal.
 
 - *example*: public data model consists of a single unified `Message` type. the framework does not expose events and other intermediate steps unless the user is writing a custom adapter.
-

examples/fastapi-vite/backend/agent.py

@@ -32,14 +32,18 @@
 async def get_weather(city: str) -> str:
     """Get current weather for a city."""
     await asyncio.sleep(2)
-    return f"Sunny, 72F in {city}" if city == "Tokyo" else f"Cloudy, 55F in {city}"
+    return (
+        f"Sunny, 72F in {city}" if city == "Tokyo" else f"Cloudy, 55F in {city}"
+    )
 
 
 @ai.tool
 async def get_population(city: str) -> int:
     """Get population of a city."""
     await asyncio.sleep(1)
-    return {"new york": 8_336_817, "tokyo": 13_960_000}.get(city.lower(), 1_000_000)
+    return {"new york": 8_336_817, "tokyo": 13_960_000}.get(
+        city.lower(), 1_000_000
+    )
 
 
 @ai.tool(require_approval=True)

examples/fastapi-vite/backend/main.py

@@ -3,7 +3,7 @@
 from __future__ import annotations
 
 import sys
-from collections.abc import AsyncGenerator
+from typing import TYPE_CHECKING
 
 import agent as agent_
 import fastapi
@@ -14,6 +14,9 @@
 
 import ai
 
+if TYPE_CHECKING:
+    from collections.abc import AsyncGenerator
+
 app = fastapi.FastAPI(
     title="py-ai-fastapi-chat",
     description="Chat demo using Python Vercel AI SDK",
@@ -38,7 +41,9 @@ async def log_validation_errors(
         file=sys.stderr,
         flush=True,
     )
-    return fastapi.responses.JSONResponse({"detail": exc.errors()}, status_code=422)
+    return fastapi.responses.JSONResponse(
+        {"detail": exc.errors()}, status_code=422
+    )
 
 
 @app.get("/health")

examples/fastapi-vite/backend/storage.py

@@ -1,5 +1,4 @@
-"""
-Pluggable storage for checkpoints and session data.
+"""Pluggable storage for checkpoints and session data.
 
 Provides a minimal Storage protocol and a FileStorage implementation
 that persists data as JSON files on disk. Swap in any backend that
@@ -25,8 +24,7 @@ async def delete(self, key: str) -> None: ...
 
 
 class FileStorage:
-    """
-    JSON-file-per-key storage backend.
+    """JSON-file-per-key storage backend.
 
     Each key is stored as ``{directory}/{key}.json``. Good enough for
     local development; replace with a real database for production.

examples/multiagent-textual/client.py

@@ -12,12 +12,13 @@
 import asyncio
 import json
 import os
-from typing import Any
+from typing import Any, ClassVar
 
 import pydantic
 import rich.text
 import textual
 import textual.app
+import textual.binding
 import textual.containers
 import textual.widgets
 import textual.worker
@@ -107,11 +108,15 @@ class MultiAgentApp(textual.app.App[None]):
     }
     """
 
-    BINDINGS = [("q", "quit", "quit")]
+    BINDINGS: ClassVar[
+        list[textual.binding.Binding | tuple[str, str] | tuple[str, str, str]]
+    ] = [("q", "quit", "quit")]
 
     def __init__(self) -> None:
         super().__init__()
-        self._hook_queue: asyncio.Queue[ai.messages.HookPart[Any]] = asyncio.Queue()
+        self._hook_queue: asyncio.Queue[ai.messages.HookPart[Any]] = (
+            asyncio.Queue()
+        )
         self._current_hook: ai.messages.HookPart[Any] | None = None
         self._ws: websockets.ClientConnection | None = None
         self._event_adapter: pydantic.TypeAdapter[ai.events.AgentEvent] = (
@@ -186,9 +191,13 @@ def _render(self, label: str, event: ai.events.AgentEvent) -> None:
             if panel is not None:
                 for part in event.message.parts:
                     match part:
-                        case ai.messages.ToolCallPart(tool_name=name, tool_args=args):
+                        case ai.messages.ToolCallPart(
+                            tool_name=name, tool_args=args
+                        ):
                             panel.append_line(f"> {name}({args})")
-                        case ai.messages.ToolResultPart(tool_name=name, result=result):
+                        case ai.messages.ToolResultPart(
+                            tool_name=name, result=result
+                        ):
                             panel.append_line(f"< {name} = {result}")
             return
 
@@ -221,7 +230,9 @@ def _on_hook_pending(self, hook_part: ai.messages.HookPart[Any]) -> None:
 
         panel = self._get_panel(branch)
         if panel:
-            panel.append_line(f"!! approval required: {tool}", style="dim yellow")
+            panel.append_line(
+                f"!! approval required: {tool}", style="dim yellow"
+            )
             panel.status = "awaiting approval"
 
         self._hook_queue.put_nowait(hook_part)
@@ -270,7 +281,9 @@ def _maybe_activate_next_hook(self) -> None:
         inp.placeholder = f"approve {branch}/{tool}? [y/n]"
         inp.focus()
 
-    async def on_input_submitted(self, event: textual.widgets.Input.Submitted) -> None:
+    async def on_input_submitted(
+        self, event: textual.widgets.Input.Submitted
+    ) -> None:
         if self._current_hook is None:
             event.input.clear()
             return

examples/multiagent-textual/server.py

@@ -14,14 +14,16 @@
 import contextlib
 import json
 import warnings
-from collections.abc import AsyncGenerator
-from typing import Any
+from typing import TYPE_CHECKING, Any
 
 import fastapi
 import pydantic
 
 import ai
 
+if TYPE_CHECKING:
+    from collections.abc import AsyncGenerator
+
 # ToolResultPart.result is typed as dict but tools can return plain strings.
 warnings.filterwarnings("ignore", category=UserWarning, module="pydantic")
 
@@ -149,7 +151,9 @@ async def loop(
 class Orchestrator(ai.Agent):
     """Run two gated agents in parallel, then summarise their results."""
 
-    async def loop(self, context: ai.Context) -> AsyncGenerator[ai.events.AgentEvent]:
+    async def loop(
+        self, context: ai.Context
+    ) -> AsyncGenerator[ai.events.AgentEvent]:
         query = context.messages[-1].text
 
         # Fan out: both branches stream concurrently via yield_from.
@@ -219,9 +223,11 @@ async def loop(self, context: ai.Context) -> AsyncGenerator[ai.events.AgentEvent
 
 
 def _normalise_message(data: dict[str, Any]) -> dict[str, Any]:
-    """Ensure ToolResultPart.result is always a dict for safe deserialisation."""
+    """Ensure ToolResultPart.result is always safe to deserialize."""
     for part in data.get("parts", []):
-        if part.get("kind") == "tool_result" and isinstance(part.get("result"), str):
+        if part.get("kind") == "tool_result" and isinstance(
+            part.get("result"), str
+        ):
             part["result"] = {"value": part["result"]}
     return data
 

examples/multiagent-textual/test-e2e.py

@@ -19,12 +19,14 @@
 import tempfile
 import time
 import urllib.request
-from http.client import HTTPResponse
 from pathlib import Path
-from typing import cast
+from typing import TYPE_CHECKING, cast
 
 import ai
 
+if TYPE_CHECKING:
+    from http.client import HTTPResponse
+
 HERE = Path(__file__).resolve().parent
 SESSION = f"multiagent-e2e-{os.getpid()}"
 SERVER_PORT = os.environ.get("SERVER_PORT", "8000")
@@ -34,7 +36,7 @@
 def _check_health() -> bool:
     try:
         with urllib.request.urlopen(f"{SERVER_URL}/api/health", timeout=1) as r:
-            return cast(HTTPResponse, r).status == 200
+            return cast("HTTPResponse", r).status == 200
     except Exception:
         return False
 

examples/run-examples.py

@@ -161,7 +161,9 @@ def _sample_path(name: str) -> Path:
     return SAMPLES / path
 
 
-def _select_sample(name: str, known_samples: dict[str, Sample]) -> Sample | None:
+def _select_sample(
+    name: str, known_samples: dict[str, Sample]
+) -> Sample | None:
     sample = known_samples.get(name)
     if sample is not None:
         return sample
@@ -178,7 +180,9 @@ def _select_sample(name: str, known_samples: dict[str, Sample]) -> Sample | None
     return None
 
 
-def _sample_cmd(sample: Sample, model: str | None, protocol: str | None) -> list[str]:
+def _sample_cmd(
+    sample: Sample, model: str | None, protocol: str | None
+) -> list[str]:
     if sample.cmd is not None:
         return sample.cmd
     base = [
@@ -258,11 +262,21 @@ def run_sample_quiet(
 
 def main() -> None:
     parser = argparse.ArgumentParser(description="Run example samples.")
-    parser.add_argument("--text", action="store_true", help="include text samples")
-    parser.add_argument("--image", action="store_true", help="include image samples")
-    parser.add_argument("--video", action="store_true", help="include video samples")
-    parser.add_argument("--broken", action="store_true", help="include broken samples")
-    parser.add_argument("--e2e", action="store_true", help="include e2e test scripts")
+    parser.add_argument(
+        "--text", action="store_true", help="include text samples"
+    )
+    parser.add_argument(
+        "--image", action="store_true", help="include image samples"
+    )
+    parser.add_argument(
+        "--video", action="store_true", help="include video samples"
+    )
+    parser.add_argument(
+        "--broken", action="store_true", help="include broken samples"
+    )
+    parser.add_argument(
+        "--e2e", action="store_true", help="include e2e test scripts"
+    )
     parser.add_argument("--all", action="store_true", help="run all samples")
     parser.add_argument(
         "--parallel", action="store_true", help="run samples in parallel"
@@ -287,11 +301,16 @@ def main() -> None:
         "examples",
         nargs="*",
         metavar="example",
-        help="example file(s) to run, e.g. stream.py or examples/samples/stream.py",
+        help=(
+            "example file(s) to run, e.g. stream.py or "
+            "examples/samples/stream.py"
+        ),
     )
     args = parser.parse_args()
 
-    has_category = args.text or args.image or args.video or args.broken or args.e2e
+    has_category = (
+        args.text or args.image or args.video or args.broken or args.e2e
+    )
 
     samples: list[Sample] = []
     if args.examples: