rasa-pro 3.13.12__py3-none-any.whl → 3.14.0__py3-none-any.whl

rasa/builder/copilot/copilot.py

@@ -0,0 +1,562 @@
+import asyncio
+import importlib
+import json
+from contextlib import asynccontextmanager
+from typing import Any, Dict, List, Optional, Union
+
+import openai
+import structlog
+from jinja2 import Template
+from typing_extensions import AsyncGenerator
+
+from rasa.builder import config
+from rasa.builder.config import COPILOT_DOCUMENTATION_SEARCH_QUERY_HISTORY_MESSAGES
+from rasa.builder.copilot.constants import (
+    COPILOT_LAST_USER_MESSAGE_CONTEXT_PROMPT_FILE,
+    COPILOT_PROMPTS_DIR,
+    COPILOT_PROMPTS_FILE,
+    COPILOT_TRAINING_ERROR_HANDLER_PROMPT_FILE,
+    ROLE_COPILOT,
+    ROLE_COPILOT_INTERNAL,
+    ROLE_USER,
+)
+from rasa.builder.copilot.exceptions import CopilotStreamError
+from rasa.builder.copilot.models import (
+    ChatMessage,
+    CopilotChatMessage,
+    CopilotContext,
+    CopilotGenerationContext,
+    CopilotSystemMessage,
+    EventContent,
+    FileContent,
+    InternalCopilotRequestChatMessage,
+    ResponseCategory,
+    UsageStatistics,
+    UserChatMessage,
+)
+from rasa.builder.document_retrieval.inkeep_document_retrieval import (
+    InKeepDocumentRetrieval,
+)
+from rasa.builder.document_retrieval.models import Document
+from rasa.builder.exceptions import (
+    DocumentRetrievalError,
+)
+from rasa.builder.shared.tracker_context import TrackerContext
+from rasa.shared.constants import PACKAGE_NAME
+
+structlogger = structlog.get_logger()
+
+
+class Copilot:
+    def __init__(self) -> None:
+        self._inkeep_document_retrieval = InKeepDocumentRetrieval()
+
+        # Load both prompt templates up-front
+        self._system_message_prompt_template = Template(
+            importlib.resources.read_text(
+                f"{PACKAGE_NAME}.{COPILOT_PROMPTS_DIR}",
+                COPILOT_PROMPTS_FILE,
+            )
+        )
+        self._last_user_message_context_prompt_template = Template(
+            importlib.resources.read_text(
+                f"{PACKAGE_NAME}.{COPILOT_PROMPTS_DIR}",
+                COPILOT_LAST_USER_MESSAGE_CONTEXT_PROMPT_FILE,
+            )
+        )
+        self._training_error_handler_prompt_template = Template(
+            importlib.resources.read_text(
+                f"{PACKAGE_NAME}.{COPILOT_PROMPTS_DIR}",
+                COPILOT_TRAINING_ERROR_HANDLER_PROMPT_FILE,
+            )
+        )
+
+        # The final stream chunk includes usage statistics.
+        self.usage_statistics = UsageStatistics()
+
+    @asynccontextmanager
+    async def _get_client(self) -> AsyncGenerator[openai.AsyncOpenAI, None]:
+        """Create a fresh OpenAI client, yield it, and always close it."""
+        client = openai.AsyncOpenAI(timeout=config.OPENAI_TIMEOUT)
+        try:
+            yield client
+        except Exception as e:
+            structlogger.error("copilot.llm_client_error", error=str(e))
+            raise
+        finally:
+            try:
+                await client.close()
+            except Exception as exc:
+                # Closing should not break request processing, but we log it
+                structlogger.warning(
+                    "copilot.llm_client_close_error",
+                    event_info="Failed to close OpenAI client cleanly.",
+                    error=str(exc),
+                )
+
+    async def search_rasa_documentation(
+        self,
+        context: CopilotContext,
+    ) -> List[Document]:
+        """Search Rasa documentation for relevant information.
+
+        Args:
+            context: The context of the copilot.
+
+        Returns:
+            A list of Document objects. Empty list is returned if the search fails.
+        """
+        try:
+            query = self._create_documentation_search_query(context)
+            return await self._inkeep_document_retrieval.retrieve_documents(query)
+        except DocumentRetrievalError as e:
+            structlogger.error(
+                "copilot.search_rasa_documentation.error",
+                event_info=(
+                    f"Copilot: Searching Rasa documentation for query '{query}' "
+                    f"failed with the following error: {e}. Returning empty list."
+                ),
+                query=query,
+                error=str(e),
+            )
+            return []
+
+    async def generate_response(
+        self,
+        context: CopilotContext,
+    ) -> tuple[AsyncGenerator[str, None], CopilotGenerationContext]:
+        """Generate a response from the copilot.
+
+        This method performs document retrieval and response generation as a single
+        atomic operation. The returned documents are the supporting evidence used
+        to generate the response, ensuring consistency between the response content
+        and its sources.
+
+        Args:
+            context: The context of the copilot.
+
+        Returns:
+            A tuple containing the async response stream and a
+            CopilotGenerationContext object with relevant documents, and all the
+            messages used to generate the response.
+
+        Raises:
+            CopilotStreamError: If the stream fails.
+            Exception: If an unexpected error occurs.
+        """
+        relevant_documents = await self.search_rasa_documentation(context)
+        messages = await self._build_messages(context, relevant_documents)
+        tracker_event_attachments = self._extract_tracker_event_attachments(
+            context.copilot_chat_history[-1]
+        )
+
+        support_evidence = CopilotGenerationContext(
+            relevant_documents=relevant_documents,
+            system_message=messages[0],
+            chat_history=messages[1:-1],
+            last_user_message=messages[-1],
+            tracker_event_attachments=tracker_event_attachments,
+        )
+
+        return (
+            self._stream_response(messages),
+            support_evidence,
+        )
+
+    async def _stream_response(
+        self, messages: List[Dict[str, Any]]
+    ) -> AsyncGenerator[str, None]:
+        """Stream markdown chunks one by one."""
+        self.usage_statistics.reset()
+
+        try:
+            async with self._get_client() as client:
+                stream = await client.chat.completions.create(
+                    model=config.OPENAI_MODEL,
+                    messages=messages,  # type: ignore
+                    temperature=config.OPENAI_TEMPERATURE,
+                    stream=True,
+                    stream_options={"include_usage": True},
+                )
+                async for chunk in stream:
+                    # The final chunk, which contains the usage statistics,
+                    # arrives with an empty `choices` list.
+                    if not chunk.choices:
+                        self.usage_statistics.update_from_stream_chunk(chunk)
+                        # Nothing to yield – continue to the next chunk.
+                        continue
+
+                    delta = chunk.choices[0].delta
+                    if delta and delta.content:
+                        yield delta.content
+        except openai.OpenAIError as e:
+            structlogger.exception("copilot.stream_response.api_error", error=str(e))
+            raise CopilotStreamError(
+                "Failed to stream response from OpenAI API."
+            ) from e
+        except asyncio.TimeoutError as e:
+            structlogger.exception(
+                "copilot.stream_response.timeout_error", error=str(e)
+            )
+            raise CopilotStreamError("Request to OpenAI API timed out.") from e
+        except Exception as e:
+            structlogger.exception(
+                "copilot.stream_response.unexpected_error", error=str(e)
+            )
+            raise
+
+    async def _build_messages(
+        self,
+        context: CopilotContext,
+        relevant_documents: List[Document],
+    ) -> List[Dict[str, Any]]:
+        """Build the complete message list for the OpenAI API.
+
+        Args:
+            context: The context of the copilot.
+            relevant_documents: The relevant documents to use in the context.
+
+        Returns:
+            A list of messages in OpenAI format.
+        """
+        if not context.copilot_chat_history:
+            return []
+
+        past_messages = self._create_chat_history_messages(
+            context.copilot_chat_history[:-1]
+        )
+
+        latest_message = self._process_latest_message(
+            context.copilot_chat_history[-1], context, relevant_documents
+        )
+        system_message = self._create_system_message()
+
+        return [system_message, *past_messages, latest_message]
+
+    def _create_chat_history_messages(
+        self, chat_history: List[Union[UserChatMessage, CopilotChatMessage]]
+    ) -> List[Dict[str, Any]]:
+        """Filter and convert past messages to OpenAI format.
+
+        Excludes guardrails policy violations and non-user/copilot messages.
+
+        Args:
+            chat_history: List of chat messages to filter and convert.
+
+        Returns:
+            List of messages in OpenAI format
+        """
+        filtered_messages = []
+
+        for message in chat_history:
+            if (
+                message.response_category
+                != ResponseCategory.GUARDRAILS_POLICY_VIOLATION
+                and message.role in [ROLE_USER, ROLE_COPILOT]
+            ):
+                filtered_messages.append(message)
+
+        return [message.build_openai_message() for message in filtered_messages]
+
+    def _process_latest_message(
+        self,
+        latest_message: Any,
+        context: CopilotContext,
+        relevant_documents: List[Document],
+    ) -> Dict[str, Any]:
+        """Process the latest message and convert it to OpenAI format.
+
+        Args:
+            latest_message: The most recent message from the chat history.
+            context: The copilot context containing conversation state.
+            relevant_documents: List of relevant documents for context.
+
+        Returns:
+            Message in OpenAI format.
+
+        Raises:
+            ValueError: If the message type is not supported.
+        """
+        if isinstance(latest_message, UserChatMessage):
+            tracker_event_attachments = latest_message.get_content_blocks_by_type(
+                EventContent
+            )
+            rendered_prompt = self._render_last_user_message_context_prompt(
+                context, relevant_documents, tracker_event_attachments
+            )
+            return latest_message.build_openai_message(prompt=rendered_prompt)
+
+        elif isinstance(latest_message, InternalCopilotRequestChatMessage):
+            rendered_prompt = self._render_training_error_handler_prompt(
+                latest_message, relevant_documents
+            )
+            return latest_message.build_openai_message(prompt=rendered_prompt)
+
+        else:
+            raise ValueError(f"Unexpected message type: {type(latest_message)}")
+
+    def _create_system_message(self) -> Dict[str, Any]:
+        """Create the system message for the conversation.
+
+        Returns:
+            System message in OpenAI format with rendered prompt template.
+        """
+        rendered_prompt = self._system_message_prompt_template.render()
+        return CopilotSystemMessage().build_openai_message(prompt=rendered_prompt)
+
+    def _render_last_user_message_context_prompt(
+        self,
+        context: CopilotContext,
+        relevant_documents: List[Document],
+        tracker_event_attachments: List[EventContent],
+    ) -> str:
+        # Format relevant documentation
+        documents = [doc.model_dump() for doc in relevant_documents]
+        # Format conversation history
+        conversation = self._format_conversation_history(context.tracker_context)
+        # Format current state
+        current_state = self._format_current_state(context.tracker_context)
+        # Format tracker events
+        attachments = self._format_tracker_event_attachments(tracker_event_attachments)
+
+        rendered_prompt = self._last_user_message_context_prompt_template.render(
+            current_conversation=conversation,
+            current_state=current_state,
+            assistant_logs=context.assistant_logs,
+            assistant_files=context.assistant_files,
+            documentation_results=documents,
+            attachments=attachments,
+        )
+        return rendered_prompt
+
+    def _render_training_error_handler_prompt(
+        self,
+        internal_request_message: InternalCopilotRequestChatMessage,
+        relevant_documents: List[Document],
+    ) -> str:
+        """Render the training error handler prompt with documentation and context.
+
+        Args:
+            internal_request_message: Internal request message.
+            context: The copilot context.
+            relevant_documents: List of relevant documents for context.
+
+        Returns:
+            Rendered prompt string for training error analysis.
+        """
+        modified_files_dicts: Dict[str, str] = {
+            file.file_path: file.file_content
+            for file in internal_request_message.get_content_blocks_by_type(FileContent)
+        }
+        rendered_prompt = self._training_error_handler_prompt_template.render(
+            logs=internal_request_message.get_flattened_log_content(),
+            modified_files=modified_files_dicts,
+            documentation_results=self._format_documents(relevant_documents),
+        )
+
+        return rendered_prompt
+
+    @staticmethod
+    def _create_documentation_search_query(context: CopilotContext) -> str:
+        """Format chat messages between user and copilot for documentation search.
+
+        Filters out guardrails policy violations and only includes messages with
+        USER or COPILOT roles, then takes the last N relevant messages.
+        """
+        role_to_prefix = {
+            ROLE_USER: "User",
+            ROLE_COPILOT: "Assistant",
+            ROLE_COPILOT_INTERNAL: "User",
+        }
+        allowed_message_types = (
+            UserChatMessage,
+            InternalCopilotRequestChatMessage,
+            CopilotChatMessage,
+        )
+
+        query_chat_history: List[str] = []
+
+        for message in reversed(context.copilot_chat_history):
+            if (
+                message.response_category
+                == ResponseCategory.GUARDRAILS_POLICY_VIOLATION
+                or not isinstance(message, allowed_message_types)
+            ):
+                continue
+
+            if (
+                len(query_chat_history)
+                >= COPILOT_DOCUMENTATION_SEARCH_QUERY_HISTORY_MESSAGES
+            ):
+                break
+
+            prefix = role_to_prefix[message.role]
+            text = (
+                Copilot._format_internal_message_for_query_chat_history(message)
+                if isinstance(message, InternalCopilotRequestChatMessage)
+                else Copilot._format_normal_message_for_query_chat_history(message)
+            )
+            query_chat_history.insert(0, f"{prefix}: {text}")
+
+        return "\n".join(query_chat_history)
+
+    @staticmethod
+    def _format_documents(results: List[Document]) -> Optional[str]:
+        """Format documentation search results as JSON dump to be used in the prompt."""
+        # We want the special message that indicates no relevant documentation source
+        # found if there are no results.
+        if not results:
+            return None
+
+        formatted_results: Dict[str, Any] = {
+            "sources": [
+                {
+                    # Start the reference from 1, not 0.
+                    "idx": idx + 1,
+                    "title": result.title,
+                    "url": result.url,
+                    "content": result.content,
+                }
+                for idx, result in enumerate(results)
+            ]
+        }
+        return json.dumps(formatted_results, ensure_ascii=False, indent=2)
+
+    @staticmethod
+    def _format_conversation_history(tracker_context: Optional[TrackerContext]) -> str:
+        """Format conversation history from TrackerContext using nested turn structure.
+
+        Args:
+            tracker_context: The TrackerContext containing conversation data.
+
+        Returns:
+            A JSON string with turns containing user_input, assistant_response,
+            and context.
+
+        Example:
+            ```json
+            {
+                "conversation_history": [
+                    {
+                        "turn_id": 1,
+                        "USER": {
+                            "text": "I want to transfer money",
+                            "predicted_commands": ["start flow", "set slot", ...]
+                        },
+                        "BOT": [
+                            {"text": "How much would you like to transfer?"}
+                        ],
+                        "other_tracker_events": [
+                            {
+                                "event": "action_executed",
+                                "data": {"action_name": "action_ask_amount"}
+                            },
+                            {
+                                "event": "slot_set",
+                                "data": {
+                                    "slot_name": "amount_of_money",
+                                    "slot_value": 100,
+                                },
+                            }
+                        ]
+                    }
+                ]
+            }
+            ```
+        """
+        conversation_history: Dict[str, Any] = {
+            "conversation_history": [],
+        }
+
+        if not tracker_context or not tracker_context.conversation_turns:
+            return json.dumps(conversation_history, ensure_ascii=False, indent=2)
+
+        conversation_turns: List[Dict[str, Any]] = []
+        user_prefix = "USER"
+        assistant_prefix = "BOT"
+
+        for turn_idx, turn in enumerate(tracker_context.conversation_turns, 1):
+            turn_data: Dict[str, Any] = {"turn_id": turn_idx}
+
+            # Add user if present
+            if turn.user_message:
+                turn_data[user_prefix] = {
+                    "text": turn.user_message.text,
+                    "predicted_commands": turn.user_message.predicted_commands,
+                }
+
+            # Add assistant messages if present
+            if turn.assistant_messages:
+                turn_data[assistant_prefix] = [
+                    {"text": assistant_message.text}
+                    for assistant_message in turn.assistant_messages
+                ]
+
+            # Add other tracker events
+            if turn.context_events:
+                other_events = [event.model_dump() for event in turn.context_events]
+                turn_data["other_tracker_events"] = other_events
+
+            conversation_turns.append(turn_data)
+
+        conversation_history["conversation_history"] = conversation_turns
+        return json.dumps(conversation_history, ensure_ascii=False, indent=2)
+
+    @staticmethod
+    def _format_current_state(tracker_context: Optional[TrackerContext]) -> str:
+        """Format current state from TrackerContext for LLM consumption.
+
+        Args:
+            tracker_context: The TrackerContext containing current state data.
+
+        Returns:
+            A JSON string containing the current state information.
+        """
+        if not tracker_context or not tracker_context.current_state:
+            return json.dumps({}, ensure_ascii=False, indent=2)
+        current_state = tracker_context.current_state.model_dump()
+        return json.dumps(current_state, ensure_ascii=False, indent=2)
+
+    @staticmethod
+    def _format_normal_message_for_query_chat_history(
+        message: Union[UserChatMessage, CopilotChatMessage],
+    ) -> str:
+        """Format normal message for query chat history."""
+        return f"{message.get_flattened_text_content()}"
+
+    @staticmethod
+    def _format_internal_message_for_query_chat_history(
+        message: InternalCopilotRequestChatMessage,
+    ) -> str:
+        """Format internal copilot request message for query chat history."""
+        text_content = message.get_flattened_text_content()
+        log_content = message.get_flattened_log_content()
+        if text_content and log_content:
+            return f"{text_content}\nLogs: {log_content}"
+        elif text_content:
+            return text_content
+        elif log_content:
+            return f"Logs: {log_content}"
+        else:
+            return ""
+
+    @staticmethod
+    def _format_tracker_event_attachments(events: List[EventContent]) -> Optional[str]:
+        """Format tracker events as JSON dump to be used in the prompt."""
+        # We don't want to display the attachment sectin in the last user message
+        # context prompt if there are no attachments.
+        if not events:
+            return None
+        # If there are attachments, return the formatted JSON dump.
+        return json.dumps(
+            [event_content.model_dump() for event_content in events],
+            ensure_ascii=False,
+            indent=2,
+        )
+
+    @staticmethod
+    def _extract_tracker_event_attachments(message: ChatMessage) -> List[EventContent]:
+        """Extract the tracker event attachments from the message."""
+        if not isinstance(message, UserChatMessage):
+            return []
+        return message.get_content_blocks_by_type(EventContent)