qtype 0.0.12__py3-none-any.whl → 0.1.7__py3-none-any.whl

qtype/interpreter/base/stream_emitter.py

@@ -0,0 +1,469 @@
+"""
+Context managers for emitting streaming events during step execution.
+
+This module provides a clean, idiomatic Python API for executors to emit
+streaming events without directly handling the StreamEvent types.
+
+Usage Example:
+    ```python
+    class MyExecutor(StepExecutor):
+        async def process_message(self, message: FlowMessage):
+            emitter = self.stream_emitter
+
+            # Status update
+            await emitter.status("Processing started...")
+
+            # Text streaming
+            async with emitter.text_stream("response-1") as streamer:
+                async for chunk in some_generator():
+                    await streamer.delta(chunk)
+
+            # Tool execution
+            async with emitter.tool_execution(
+                tool_call_id="tool-1",
+                tool_name="search",
+                tool_input={"query": "test"}
+            ) as tool_ctx:
+                result = await execute_tool()
+                await tool_ctx.complete(result)
+
+            yield message
+    ```
+"""
+
+from __future__ import annotations
+
+from typing import Any
+
+from qtype.interpreter.types import (
+    ErrorEvent,
+    ReasoningStreamDeltaEvent,
+    ReasoningStreamEndEvent,
+    ReasoningStreamStartEvent,
+    StatusEvent,
+    StepEndEvent,
+    StepStartEvent,
+    StreamingCallback,
+    TextStreamDeltaEvent,
+    TextStreamEndEvent,
+    TextStreamStartEvent,
+    ToolExecutionEndEvent,
+    ToolExecutionErrorEvent,
+    ToolExecutionStartEvent,
+)
+from qtype.semantic.model import Step
+
+
+class TextStreamContext:
+    """
+    Async context manager for text streaming.
+
+    Automatically emits TextStreamStartEvent on entry and TextStreamEndEvent
+    on exit. Provides delta() method for emitting text chunks.
+
+    Example:
+        ```python
+        async with emitter.text_stream("llm-response") as streamer:
+            async for chunk in llm_client.stream():
+                await streamer.delta(chunk.text)
+        ```
+    """
+
+    def __init__(
+        self,
+        step: Step,
+        stream_id: str,
+        on_stream_event: StreamingCallback | None,
+    ):
+        self.step = step
+        self.stream_id = stream_id
+        self.on_stream_event = on_stream_event
+
+    async def __aenter__(self) -> TextStreamContext:
+        """Emit TextStreamStartEvent when entering context."""
+        if self.on_stream_event:
+            await self.on_stream_event(
+                TextStreamStartEvent(step=self.step, stream_id=self.stream_id)
+            )
+        return self
+
+    async def __aexit__(
+        self,
+        exc_type: type[BaseException] | None,
+        exc_val: BaseException | None,
+        exc_tb: Any,
+    ) -> bool:
+        """Emit TextStreamEndEvent when exiting context."""
+        if self.on_stream_event:
+            await self.on_stream_event(
+                TextStreamEndEvent(step=self.step, stream_id=self.stream_id)
+            )
+        return False
+
+    async def delta(self, text: str) -> None:
+        """
+        Emit a text delta chunk.
+
+        Args:
+            text: The incremental text content to append to the stream
+        """
+        if self.on_stream_event:
+            await self.on_stream_event(
+                TextStreamDeltaEvent(
+                    step=self.step,
+                    stream_id=self.stream_id,
+                    delta=text,
+                )
+            )
+
+
+class ReasoningStreamContext:
+    """
+    Async context manager for reasoning streaming.
+
+    Automatically emits ReasoningStreamStartEvent on entry and
+    ReasoningStreamEndEvent on exit. Provides delta() method for emitting
+    reasoning chunks.
+
+    Example:
+        ```python
+        async with emitter.reasoning_stream("agent-reasoning") as streamer:
+            async for chunk in agent.stream_reasoning():
+                await streamer.delta(chunk.text)
+        ```
+    """
+
+    def __init__(
+        self,
+        step: Step,
+        stream_id: str,
+        on_stream_event: StreamingCallback | None,
+    ):
+        self.step = step
+        self.stream_id = stream_id
+        self.on_stream_event = on_stream_event
+
+    async def __aenter__(self) -> ReasoningStreamContext:
+        """Emit ReasoningStreamStartEvent when entering context."""
+        if self.on_stream_event:
+            await self.on_stream_event(
+                ReasoningStreamStartEvent(
+                    step=self.step, stream_id=self.stream_id
+                )
+            )
+        return self
+
+    async def __aexit__(
+        self,
+        exc_type: type[BaseException] | None,
+        exc_val: BaseException | None,
+        exc_tb: Any,
+    ) -> bool:
+        """Emit ReasoningStreamEndEvent when exiting context."""
+        if self.on_stream_event:
+            await self.on_stream_event(
+                ReasoningStreamEndEvent(
+                    step=self.step, stream_id=self.stream_id
+                )
+            )
+        return False
+
+    async def delta(self, text: str) -> None:
+        """
+        Emit a reasoning delta chunk.
+
+        Args:
+            text: The incremental reasoning content to append to the stream
+        """
+        if self.on_stream_event:
+            await self.on_stream_event(
+                ReasoningStreamDeltaEvent(
+                    step=self.step,
+                    stream_id=self.stream_id,
+                    delta=text,
+                )
+            )
+
+
+class StepBoundaryContext:
+    """
+    Async context manager for step boundaries.
+
+    Automatically emits StepStartEvent on entry and StepEndEvent on exit.
+    Use this to group related events together visually in the UI.
+
+    Example:
+        ```python
+        async with emitter.step_boundary():
+            await emitter.status("Step 1: Loading data...")
+            # ... do work ...
+            await emitter.status("Step 1: Complete")
+        ```
+    """
+
+    def __init__(
+        self,
+        step: Step,
+        on_stream_event: StreamingCallback | None,
+    ):
+        self.step = step
+        self.on_stream_event = on_stream_event
+
+    async def __aenter__(self) -> StepBoundaryContext:
+        """Emit StepStartEvent when entering context."""
+        if self.on_stream_event:
+            await self.on_stream_event(StepStartEvent(step=self.step))
+        return self
+
+    async def __aexit__(
+        self,
+        exc_type: type[BaseException] | None,
+        exc_val: BaseException | None,
+        exc_tb: Any,
+    ) -> bool:
+        """Emit StepEndEvent when exiting context."""
+        if self.on_stream_event:
+            await self.on_stream_event(StepEndEvent(step=self.step))
+        return False
+
+
+class ToolExecutionContext:
+    """
+    Async context manager for tool execution.
+
+    Automatically emits ToolExecutionStartEvent on entry. On exit, if an
+    exception occurred, emits ToolExecutionErrorEvent. Otherwise, you must
+    call complete() or error() explicitly.
+
+    Example:
+        ```python
+        async with emitter.tool_execution(
+            tool_call_id="tool-1",
+            tool_name="search",
+            tool_input={"query": "test"}
+        ) as tool_ctx:
+            result = await execute_tool()
+            await tool_ctx.complete(result)
+        ```
+    """
+
+    def __init__(
+        self,
+        step: Step,
+        tool_call_id: str,
+        tool_name: str,
+        tool_input: dict[str, Any],
+        on_stream_event: StreamingCallback | None,
+    ):
+        self.step = step
+        self.tool_call_id = tool_call_id
+        self.tool_name = tool_name
+        self.tool_input = tool_input
+        self.on_stream_event = on_stream_event
+        self._completed = False
+
+    async def __aenter__(self) -> ToolExecutionContext:
+        """Emit ToolExecutionStartEvent when entering context."""
+        if self.on_stream_event:
+            await self.on_stream_event(
+                ToolExecutionStartEvent(
+                    step=self.step,
+                    tool_call_id=self.tool_call_id,
+                    tool_name=self.tool_name,
+                    tool_input=self.tool_input,
+                )
+            )
+        return self
+
+    async def __aexit__(
+        self,
+        exc_type: type[BaseException] | None,
+        exc_val: BaseException | None,
+        exc_tb: Any,
+    ) -> bool:
+        """
+        Emit ToolExecutionErrorEvent if exception occurred.
+
+        If no exception and complete()/error() wasn't called, this is a
+        programming error but we don't raise to avoid masking other issues.
+        """
+        if exc_type is not None and self.on_stream_event:
+            await self.on_stream_event(
+                ToolExecutionErrorEvent(
+                    step=self.step,
+                    tool_call_id=self.tool_call_id,
+                    error_message=str(exc_val),
+                )
+            )
+            self._completed = True
+        return False
+
+    async def complete(self, output: Any) -> None:
+        """
+        Mark tool execution as complete with successful output.
+
+        Args:
+            output: The result returned by the tool
+        """
+        if self._completed:
+            return
+        if self.on_stream_event:
+            await self.on_stream_event(
+                ToolExecutionEndEvent(
+                    step=self.step,
+                    tool_call_id=self.tool_call_id,
+                    tool_output=output,
+                )
+            )
+        self._completed = True
+
+    async def error(self, error_message: str) -> None:
+        """
+        Mark tool execution as failed.
+
+        Args:
+            error_message: Description of the error that occurred
+        """
+        if self._completed:
+            return
+        if self.on_stream_event:
+            await self.on_stream_event(
+                ToolExecutionErrorEvent(
+                    step=self.step,
+                    tool_call_id=self.tool_call_id,
+                    error_message=error_message,
+                )
+            )
+        self._completed = True
+
+
+class StreamEmitter:
+    """
+    Factory for creating streaming context managers.
+
+    This class is instantiated once per StepExecutor and provides factory
+    methods for creating context managers and convenience methods for
+    one-shot events.
+
+    The executor can access this via self.stream_emitter.
+
+    Example:
+        ```python
+        class MyExecutor(StepExecutor):
+            async def process_message(self, message: FlowMessage):
+                # One-shot status
+                await self.stream_emitter.status("Processing...")
+
+                # Text streaming
+                async with self.stream_emitter.text_stream("id") as s:
+                    await s.delta("Hello")
+
+                # Step boundary
+                async with self.stream_emitter.step_boundary():
+                    await self.stream_emitter.status("Step content")
+
+                # Tool execution
+                async with self.stream_emitter.tool_execution(
+                    "tool-1", "search", {"q": "test"}
+                ) as tool:
+                    result = await run_tool()
+                    await tool.complete(result)
+
+                yield message
+        ```
+    """
+
+    def __init__(
+        self,
+        step: Step,
+        on_stream_event: StreamingCallback | None,
+    ):
+        self.step = step
+        self.on_stream_event = on_stream_event
+
+    def text_stream(self, stream_id: str) -> TextStreamContext:
+        """
+        Create a context manager for text streaming.
+
+        Args:
+            stream_id: Unique identifier for this text stream
+
+        Returns:
+            Context manager that emits start/delta/end events
+        """
+        return TextStreamContext(self.step, stream_id, self.on_stream_event)
+
+    def reasoning_stream(self, stream_id: str) -> ReasoningStreamContext:
+        """
+        Create a context manager for reasoning streaming.
+
+        Args:
+            stream_id: Unique identifier for this reasoning stream
+
+        Returns:
+            Context manager that emits start/delta/end events for reasoning
+        """
+        return ReasoningStreamContext(
+            self.step, stream_id, self.on_stream_event
+        )
+
+    def step_boundary(self) -> StepBoundaryContext:
+        """
+        Create a context manager for step boundaries.
+
+        Returns:
+            Context manager that emits step start/end events
+        """
+        return StepBoundaryContext(self.step, self.on_stream_event)
+
+    def tool_execution(
+        self,
+        tool_call_id: str,
+        tool_name: str,
+        tool_input: dict[str, Any],
+    ) -> ToolExecutionContext:
+        """
+        Create a context manager for tool execution.
+
+        Args:
+            tool_call_id: Unique identifier for this tool call
+            tool_name: Name of the tool being executed
+            tool_input: Input parameters for the tool
+
+        Returns:
+            Context manager that emits tool execution events
+        """
+        return ToolExecutionContext(
+            self.step,
+            tool_call_id,
+            tool_name,
+            tool_input,
+            self.on_stream_event,
+        )
+
+    async def status(self, message: str) -> None:
+        """
+        Emit a complete status message.
+
+        This is a convenience method for simple status updates that don't
+        require streaming.
+
+        Args:
+            message: The status message to display
+        """
+        if self.on_stream_event:
+            await self.on_stream_event(
+                StatusEvent(step=self.step, message=message)
+            )
+
+    async def error(self, error_message: str) -> None:
+        """
+        Emit an error event.
+
+        Args:
+            error_message: Description of the error
+        """
+        if self.on_stream_event:
+            await self.on_stream_event(
+                ErrorEvent(step=self.step, error_message=error_message)
+            )