rasa-pro 3.13.0.dev8__py3-none-any.whl → 3.13.0.dev10__py3-none-any.whl

rasa/core/channels/socketio.py

@@ -191,6 +191,60 @@ class SocketIOInput(InputChannel):
             return None
         return SocketIOOutput(self.sio, self.bot_message_evt)
 
+    async def handle_session_request(
+        self, sid: Text, data: Optional[Dict] = None
+    ) -> None:
+        """Handles session requests from the client."""
+        if data is None:
+            data = {}
+        if "session_id" not in data or data["session_id"] is None:
+            data["session_id"] = uuid.uuid4().hex
+        if self.session_persistence:
+            if inspect.iscoroutinefunction(self.sio.enter_room):  # type: ignore[union-attr]
+                await self.sio.enter_room(sid, data["session_id"])  # type: ignore[union-attr]
+            else:
+                # for backwards compatibility with python-socketio < 5.10.
+                # previously, this function was NOT async.
+                self.sio.enter_room(sid, data["session_id"])  # type: ignore[union-attr]
+        await self.sio.emit("session_confirm", data["session_id"], room=sid)  # type: ignore[union-attr]
+        logger.debug(f"User {sid} connected to socketIO endpoint.")
+
+    async def handle_user_message(
+        self,
+        sid: Text,
+        data: Dict,
+        on_new_message: Callable[[UserMessage], Awaitable[Any]],
+    ) -> None:
+        """Handles user messages received from the client."""
+        output_channel = SocketIOOutput(self.sio, self.bot_message_evt)
+
+        if self.session_persistence:
+            if not data.get("session_id"):
+                rasa.shared.utils.io.raise_warning(
+                    "A message without a valid session_id "
+                    "was received. This message will be "
+                    "ignored. Make sure to set a proper "
+                    "session id using the "
+                    "`session_request` socketIO event."
+                )
+                return
+            sender_id = data["session_id"]
+        else:
+            sender_id = sid
+
+        metadata = data.get(self.metadata_key, {})
+        if isinstance(metadata, Text):
+            metadata = json.loads(metadata)
+
+        message = UserMessage(
+            data.get("message", ""),
+            output_channel,
+            sender_id,
+            input_channel=self.name(),
+            metadata=metadata,
+        )
+        await on_new_message(message)
+
     def blueprint(
         self, on_new_message: Callable[[UserMessage], Awaitable[Any]]
     ) -> SocketBlueprint:
@@ -233,49 +287,10 @@ class SocketIOInput(InputChannel):
 
         @sio.on("session_request", namespace=self.namespace)
         async def session_request(sid: Text, data: Optional[Dict]) -> None:
-            if data is None:
-                data = {}
-            if "session_id" not in data or data["session_id"] is None:
-                data["session_id"] = uuid.uuid4().hex
-            if self.session_persistence:
-                if inspect.iscoroutinefunction(sio.enter_room):
-                    await sio.enter_room(sid, data["session_id"])
-                else:
-                    # for backwards compatibility with python-socketio < 5.10.
-                    # previously, this function was NOT async.
-                    sio.enter_room(sid, data["session_id"])
-            await sio.emit("session_confirm", data["session_id"], room=sid)
-            logger.debug(f"User {sid} connected to socketIO endpoint.")
+            await self.handle_session_request(sid, data)
 
         @sio.on(self.user_message_evt, namespace=self.namespace)
         async def handle_message(sid: Text, data: Dict) -> None:
-            output_channel = SocketIOOutput(sio, self.bot_message_evt)
-
-            if self.session_persistence:
-                if not data.get("session_id"):
-                    rasa.shared.utils.io.raise_warning(
-                        "A message without a valid session_id "
-                        "was received. This message will be "
-                        "ignored. Make sure to set a proper "
-                        "session id using the "
-                        "`session_request` socketIO event."
-                    )
-                    return
-                sender_id = data["session_id"]
-            else:
-                sender_id = sid
-
-            metadata = data.get(self.metadata_key, {})
-            if isinstance(metadata, Text):
-                metadata = json.loads(metadata)
-
-            message = UserMessage(
-                data.get("message", ""),
-                output_channel,
-                sender_id,
-                input_channel=self.name(),
-                metadata=metadata,
-            )
-            await on_new_message(message)
+            await self.handle_user_message(sid, data, on_new_message)
 
         return socketio_webhook

rasa/core/channels/studio_chat.py

@@ -1,5 +1,8 @@
 import asyncio
+import audioop
+import base64
 import json
+import uuid
 from functools import partial
 from typing import (
     TYPE_CHECKING,
@@ -10,12 +13,25 @@ from typing import (
     List,
     Optional,
     Text,
+    Tuple,
 )
 
 import structlog
-from sanic import Sanic
 
+from rasa.core.channels import UserMessage
 from rasa.core.channels.socketio import SocketBlueprint, SocketIOInput
+from rasa.core.channels.voice_ready.utils import CallParameters
+from rasa.core.channels.voice_stream.audio_bytes import RasaAudioBytes
+from rasa.core.channels.voice_stream.call_state import call_state
+from rasa.core.channels.voice_stream.tts import TTSEngine
+from rasa.core.channels.voice_stream.voice_channel import (
+    ContinueConversationAction,
+    EndConversationAction,
+    NewAudioAction,
+    VoiceChannelAction,
+    VoiceInputChannel,
+    VoiceOutputChannel,
+)
 from rasa.hooks import hookimpl
 from rasa.plugin import plugin_manager
 from rasa.shared.core.constants import ACTION_LISTEN_NAME
@@ -23,7 +39,10 @@ from rasa.shared.core.events import ActionExecuted
 from rasa.shared.core.trackers import EventVerbosity
 
 if TYPE_CHECKING:
-    from rasa.core.channels.channel import UserMessage
+    from sanic import Sanic, Websocket  # type: ignore[attr-defined]
+    from socketio import AsyncServer
+
+    from rasa.core.channels.channel import InputChannel, UserMessage
     from rasa.shared.core.trackers import DialogueStateTracker
 
 
@@ -102,26 +121,84 @@ class StudioTrackerUpdatePlugin:
         self._cancel_tasks()
 
 
-class StudioChatInput(SocketIOInput):
+class StudioChatInput(SocketIOInput, VoiceInputChannel):
     """Input channel for the communication between Rasa Studio and Rasa Pro."""
 
+    requires_voice_license = False
+
     @classmethod
     def name(cls) -> Text:
         return "studio_chat"
 
     def __init__(
         self,
-        *args: Any,
-        **kwargs: Any,
+        server_url: str,
+        asr_config: Dict,
+        tts_config: Dict,
+        user_message_evt: Text = "user_uttered",
+        bot_message_evt: Text = "bot_uttered",
+        namespace: Optional[Text] = None,
+        session_persistence: bool = False,
+        socketio_path: Optional[Text] = "/socket.io",
+        jwt_key: Optional[Text] = None,
+        jwt_method: Optional[Text] = "HS256",
+        metadata_key: Optional[Text] = "metadata",
     ) -> None:
-        """Creates a ``SocketIOInput`` object."""
+        """Creates a `StudioChatInput` object."""
         from rasa.core.agent import Agent
 
-        super().__init__(*args, **kwargs)
         self.agent: Optional[Agent] = None
 
+        # Initialize the SocketIO input channel
+        SocketIOInput.__init__(
+            self,
+            user_message_evt=user_message_evt,
+            bot_message_evt=bot_message_evt,
+            namespace=namespace,
+            session_persistence=session_persistence,
+            socketio_path=socketio_path,
+            jwt_key=jwt_key,
+            jwt_method=jwt_method,
+            metadata_key=metadata_key,
+        )
+
+        # Initialize the Voice Input Channel
+        VoiceInputChannel.__init__(
+            self,
+            server_url=server_url,
+            asr_config=asr_config,
+            tts_config=tts_config,
+        )
+
+        # Dictionaries to manage active connections and background tasks
+        # `active_connections` holds the active voice sessions
+        # `background_tasks` holds the asyncio tasks for voice streaming
+        self.active_connections: Dict[str, SocketIOVoiceWebsocketAdapter] = {}
+        self.background_tasks: Dict[str, asyncio.Task] = {}
+
         self._register_tracker_update_hook()
 
+    @classmethod
+    def from_credentials(cls, credentials: Optional[Dict[Text, Any]]) -> "InputChannel":
+        """Creates a StudioChatInput channel from credentials."""
+        credentials = credentials or {}
+
+        return cls(
+            # Voice specific parameters
+            server_url=credentials.get("server_url", ""),
+            asr_config=credentials.get("asr", {}),
+            tts_config=credentials.get("tts", {}),
+            # SocketIO parameters
+            user_message_evt=credentials.get("user_message_evt", "user_uttered"),
+            bot_message_evt=credentials.get("bot_message_evt", "bot_uttered"),
+            namespace=credentials.get("namespace"),
+            session_persistence=credentials.get("session_persistence", False),
+            socketio_path=credentials.get("socketio_path", "/socket.io"),
+            jwt_key=credentials.get("jwt_key"),
+            jwt_method=credentials.get("jwt_method", "HS256"),
+            metadata_key=credentials.get("metadata_key", "metadata"),
+        )
+
     def _register_tracker_update_hook(self) -> None:
         plugin_manager().register(StudioTrackerUpdatePlugin(self))
 
@@ -191,6 +268,117 @@ class StudioChatInput(SocketIOInput):
 
         await self.on_tracker_updated(tracker)
 
+    def channel_bytes_to_rasa_audio_bytes(self, input_bytes: bytes) -> RasaAudioBytes:
+        """Voice method to convert channel bytes to RasaAudioBytes."""
+        return RasaAudioBytes(audioop.lin2ulaw(input_bytes, 4))
+
+    async def collect_call_parameters(
+        self, channel_websocket: "Websocket"
+    ) -> Optional[CallParameters]:
+        """Voice method to collect call parameters"""
+        session_id = channel_websocket.session_id
+        return CallParameters(session_id, "local", "local", stream_id=session_id)
+
+    def map_input_message(
+        self,
+        message: Any,
+        ws: "Websocket",
+    ) -> VoiceChannelAction:
+        """Voice method to map websocket messages to actions."""
+        if "audio" in message:
+            channel_bytes = base64.b64decode(message["audio"])
+            audio_bytes = self.channel_bytes_to_rasa_audio_bytes(channel_bytes)
+            return NewAudioAction(audio_bytes)
+        elif "marker" in message:
+            if message["marker"] == call_state.latest_bot_audio_id:
+                # Just finished streaming last audio bytes
+                call_state.is_bot_speaking = False  # type: ignore[attr-defined]
+                if call_state.should_hangup:
+                    structlogger.debug(
+                        "studio_chat.hangup", marker=call_state.latest_bot_audio_id
+                    )
+                    return EndConversationAction()
+            else:
+                call_state.is_bot_speaking = True  # type: ignore[attr-defined]
+        return ContinueConversationAction()
+
+    def create_output_channel(
+        self, voice_websocket: "Websocket", tts_engine: TTSEngine
+    ) -> VoiceOutputChannel:
+        """Create a voice output channel"""
+        return StudioVoiceOutputChannel(
+            voice_websocket,
+            tts_engine,
+            self.tts_cache,
+        )
+
+    def _start_voice_session(
+        self,
+        session_id: str,
+        sid: str,
+        on_new_message: Callable[[UserMessage], Awaitable[Any]],
+    ) -> None:
+        """Create SocketIO WebSocket Adaptor & start async task for voice streaming."""
+        if sid in self.active_connections:
+            structlogger.warning(
+                "studio_chat._start_voice_session.session_already_active",
+                session_id=sid,
+            )
+            return
+
+        structlogger.info(
+            "studio_chat._start_voice_session.starting_session", session_id=sid
+        )
+
+        # Create a websocket adapter for this connection
+        ws_adapter = SocketIOVoiceWebsocketAdapter(
+            sio=self.sio,
+            session_id=session_id,
+            sid=sid,
+            bot_message_evt=self.bot_message_evt,
+        )
+        self.active_connections[sid] = ws_adapter
+
+        # Start voice streaming in an async task
+        task = asyncio.create_task(
+            self._handle_voice_streaming(on_new_message, ws_adapter, sid)
+        )
+        self.background_tasks[sid] = task
+        task.add_done_callback(lambda _: self._cleanup_tasks_for_sid(sid))
+
+    async def _handle_voice_streaming(
+        self,
+        on_new_message: Callable[[UserMessage], Awaitable[Any]],
+        ws_adapter: "Websocket",
+        sid: str,
+    ) -> None:
+        """Handle voice streaming for a Socket.IO connection."""
+        try:
+            await self.run_audio_streaming(on_new_message, ws_adapter)
+        except Exception as e:
+            structlogger.exception(
+                "studio_voice.voice_streaming.error",
+                error=str(e),
+                sid=sid,
+            )
+            if sid in self.active_connections:
+                del self.active_connections[sid]
+
+    def _cleanup_tasks_for_sid(self, sid: str) -> None:
+        if sid in self.background_tasks:
+            task = self.background_tasks.pop(sid)
+            task.cancel()
+        if sid in self.active_connections:
+            del self.active_connections[sid]
+
+    @hookimpl  # type: ignore[misc]
+    def after_server_stop(self) -> None:
+        """Cleanup background tasks and active connections when the server stops."""
+        structlogger.info("studio_chat.after_server_stop.cleanup")
+        self.active_connections.clear()
+        for task in self.background_tasks.values():
+            task.cancel()
+
     def blueprint(
         self, on_new_message: Callable[["UserMessage"], Awaitable[Any]]
     ) -> SocketBlueprint:
@@ -203,11 +391,126 @@ class StudioChatInput(SocketIOInput):
             return socket_blueprint
 
         @socket_blueprint.listener("after_server_start")  # type: ignore[misc]
-        async def after_server_start(app: Sanic, _: asyncio.AbstractEventLoop) -> None:
+        async def after_server_start(
+            app: "Sanic", _: asyncio.AbstractEventLoop
+        ) -> None:
             self.agent = app.ctx.agent
 
+        @self.sio.on("disconnect", namespace=self.namespace)
+        async def disconnect(sid: Text) -> None:
+            structlogger.debug("studio_chat.sio.disconnect", sid=sid)
+            self._cleanup_tasks_for_sid(sid)
+
+        @self.sio.on("session_request", namespace=self.namespace)
+        async def session_request(sid: Text, data: Optional[Dict]) -> None:
+            """Overrides the base SocketIOInput session_request handler.
+
+            Args:
+              sid: ID of the session (from SocketIO).
+              data:
+                - session_id: Studio Chat channel is used with a Bridge Architecture
+                  (Model Service's Socket Bridge), so we use session_id to remain
+                  consistent across the bridge. Session ID becomes the sender_id
+                  for the UserMessage.
+                - is_voice: Boolean indicating if its a voice session.
+            """
+            # Call parent session_request handler first
+            await self.handle_session_request(sid, data)
+
+            # start a voice session if requested
+            if data and data.get("is_voice", False):
+                self._start_voice_session(data["session_id"], sid, on_new_message)
+
+        @self.sio.on(self.user_message_evt, namespace=self.namespace)
+        async def handle_message(sid: Text, data: Dict) -> None:
+            """Overrides the base SocketIOInput handle_message handler."""
+            # Handle voice messages
+            if "audio" in data or "marker" in data:
+                if sid in self.active_connections:
+                    # Route audio messages to the voice adapter queue
+                    ws = self.active_connections[sid]
+                    ws.put_message(data)
+                return
+
+            # Handle text messages
+            await self.handle_user_message(sid, data, on_new_message)
+
         @self.sio.on("update_tracker", namespace=self.namespace)
         async def on_update_tracker(sid: Text, data: Dict) -> None:
             await self.handle_tracker_update(sid, data)
 
         return socket_blueprint
+
+
+class StudioVoiceOutputChannel(VoiceOutputChannel):
+    @classmethod
+    def name(cls) -> str:
+        return "studio_chat"
+
+    def rasa_audio_bytes_to_channel_bytes(
+        self, rasa_audio_bytes: RasaAudioBytes
+    ) -> bytes:
+        return audioop.ulaw2lin(rasa_audio_bytes, 4)
+
+    def channel_bytes_to_message(self, recipient_id: str, channel_bytes: bytes) -> str:
+        return json.dumps({"audio": base64.b64encode(channel_bytes).decode("utf-8")})
+
+    def create_marker_message(self, recipient_id: str) -> Tuple[str, str]:
+        message_id = uuid.uuid4().hex
+        return json.dumps({"marker": message_id}), message_id
+
+
+class SocketIOVoiceWebsocketAdapter:
+    """Adapter to make Socket.IO work like a Sanic WebSocket for voice channels."""
+
+    def __init__(
+        self, sio: "AsyncServer", session_id: str, sid: str, bot_message_evt: str
+    ) -> None:
+        self.sio = sio
+        self.bot_message_evt = bot_message_evt
+        self._closed = False
+        self._receive_queue: asyncio.Queue[Any] = asyncio.Queue()
+
+        # the messages need to be emitted on room=sid
+        self.sid = sid
+
+        # used by collect_call_parameters
+        # ultimately, this becomes the sender_id
+        self.session_id = session_id
+
+    @property
+    def closed(self) -> bool:
+        return self._closed
+
+    async def send(self, data: Any) -> None:
+        """Send data to the client."""
+        if not self.closed:
+            await self.sio.emit(self.bot_message_evt, data, room=self.sid)
+
+    async def recv(self) -> Any:
+        """Receive data from the client."""
+        if self.closed:
+            raise ConnectionError("WebSocket is closed")
+        return await self._receive_queue.get()
+
+    def put_message(self, message: Any) -> None:
+        """Put message in the internal receive queue."""
+        self._receive_queue.put_nowait(message)
+
+    async def close(self, code: int = 1000, reason: str = "") -> None:
+        """Close the connection."""
+        self._closed = True
+        # at this point, the client should have disconnected
+
+    def __aiter__(self) -> "SocketIOVoiceWebsocketAdapter":
+        """Allow the adapter to be used in an async for loop."""
+        return self
+
+    async def __anext__(self) -> Any:
+        if self.closed:
+            raise StopAsyncIteration
+        try:
+            message = await self.recv()
+            return message
+        except Exception:
+            raise StopAsyncIteration

rasa/core/channels/voice_ready/audiocodes.py

@@ -177,7 +177,7 @@ class Conversation:
             else:
                 structlogger.warning(
                     "audiocodes.handle.activities.unknown_activity_type",
-                    activity=activity,
+                    activity_type=activity["type"],
                 )
                 continue
 

rasa/core/channels/voice_stream/audiocodes.py

@@ -295,7 +295,7 @@ class AudiocodesVoiceInputChannel(VoiceInputChannel):
     def blueprint(
         self, on_new_message: Callable[[UserMessage], Awaitable[Any]]
     ) -> Blueprint:
-        """Defines a Sanic bluelogger.debug."""
+        """Defines a Sanic blueprint"""
         blueprint = Blueprint("audiocodes_stream", __name__)
 
         @blueprint.route("/", methods=["GET"])

rasa/core/channels/voice_stream/browser_audio.py

@@ -97,7 +97,7 @@ class BrowserAudioInputChannel(VoiceInputChannel):
     def blueprint(
         self, on_new_message: Callable[[UserMessage], Awaitable[Any]]
     ) -> Blueprint:
-        """Defines a Sanic bluelogger.debug."""
+        """Defines a Sanic blueprint"""
         blueprint = Blueprint("browser_audio", __name__)
 
         @blueprint.route("/", methods=["GET"])

rasa/core/channels/voice_stream/tts/__init__.py

@@ -0,0 +1,8 @@
+from rasa.core.channels.voice_stream.tts.tts_cache import TTSCache
+from rasa.core.channels.voice_stream.tts.tts_engine import (
+    TTSEngine,
+    TTSEngineConfig,
+    TTSError,
+)
+
+__all__ = ["TTSEngine", "TTSEngineConfig", "TTSError", "TTSCache"]

rasa/core/channels/voice_stream/voice_channel.py

@@ -286,13 +286,18 @@ class VoiceOutputChannel(OutputChannel):
 
 
 class VoiceInputChannel(InputChannel):
+    # All children of this class require a voice license to be used.
+    requires_voice_license = True
+
     def __init__(
         self,
         server_url: str,
         asr_config: Dict,
         tts_config: Dict,
     ):
-        validate_voice_license_scope()
+        if self.requires_voice_license:
+            validate_voice_license_scope()
+
         self.server_url = server_url
         self.asr_config = asr_config
         self.tts_config = tts_config

rasa/core/exporter.py

@@ -16,6 +16,11 @@ from rasa.exceptions import (
     NoEventsToMigrateError,
     PublishingError,
 )
+from rasa.shared.core.events import (
+    BotUttered,
+    SlotSet,
+    UserUttered,
+)
 from rasa.shared.core.trackers import EventVerbosity
 
 logger = logging.getLogger(__name__)
@@ -43,6 +48,7 @@ class Exporter:
         tracker_store: TrackerStore,
         event_broker: EventBroker,
         endpoints_path: Text,
+        is_pii_enabled: bool = False,
         requested_conversation_ids: Optional[Text] = None,
         minimum_timestamp: Optional[float] = None,
         maximum_timestamp: Optional[float] = None,
@@ -52,6 +58,7 @@ class Exporter:
         self.tracker_store = tracker_store
 
         self.event_broker = event_broker
+        self.is_pii_enabled = is_pii_enabled
         self.requested_conversation_ids = requested_conversation_ids
         self.minimum_timestamp = minimum_timestamp
         self.maximum_timestamp = maximum_timestamp
@@ -72,10 +79,12 @@ class Exporter:
         current_timestamp = None
 
         headers = self._get_message_headers()
+        warned_sender_ids: Set[Text] = set()
 
         async for event in self._fetch_events_within_time_range():
             # noinspection PyBroadException
             try:
+                self._check_anonymization_status(event, warned_sender_ids)
                 self._publish_with_message_headers(event, headers)
                 published_events += 1
                 current_timestamp = event["timestamp"]
@@ -282,3 +291,30 @@ class Exporter:
             events_with_conversation_id.append(event)
 
         return events_with_conversation_id
+
+    def _check_anonymization_status(
+        self, event: Dict[Text, Any], warned_sender_ids: Set[Text]
+    ) -> None:
+        """Check if the tracker store contains unanonymized events.
+
+        If it does, print a warning that these events will be published as is.
+
+        Args:
+            event: The event to check for anonymization status
+            warned_sender_ids: Set of sender IDs that have already been warned about
+        """
+        sender_id = event["sender_id"]
+        if (
+            self.is_pii_enabled
+            and sender_id not in warned_sender_ids
+            and event["event"]
+            in (UserUttered.type_name, BotUttered.type_name, SlotSet.type_name)
+            and not event.get("anonymized_at", None)
+        ):
+            rasa.shared.utils.cli.print_warning(
+                f"Retrieved un-anonymized event for sender_id {sender_id}. "
+                f"All events after this timestamp {event['timestamp']} "
+                "are not anonymized for this tracker. Proceeding with "
+                "publishing plaintext values for all events following this.",
+            )
+            warned_sender_ids.add(sender_id)

rasa/core/information_retrieval/faiss.py

@@ -12,6 +12,7 @@ from rasa.core.information_retrieval import (
     InformationRetrievalException,
     SearchResultList,
 )
+from rasa.core.information_retrieval.ingestion.faq_parser import _format_faq_documents
 from rasa.utils.endpoints import EndpointConfig
 from rasa.utils.ml_utils import persist_faiss_vector_store
 
@@ -31,10 +32,12 @@ class FAISS_Store(InformationRetrieval):
         index_path: str,
         docs_folder: Optional[str],
         create_index: Optional[bool] = False,
+        parse_as_faq_pairs: Optional[bool] = False,
     ):
         """Initializes the FAISS Store."""
         self.chunk_size = 1000
         self.chunk_overlap = 20
+        self.parse_as_faq_pairs = parse_as_faq_pairs
 
         path = Path(index_path) / "documents_faiss"
         if create_index:
@@ -86,21 +89,25 @@ class FAISS_Store(InformationRetrieval):
         if not docs_folder:
             raise ValueError("parameter `docs_folder` needs to be specified")
 
-        docs = self.load_documents(docs_folder)
-        splitter = RecursiveCharacterTextSplitter(
-            chunk_size=self.chunk_size,
-            chunk_overlap=self.chunk_overlap,
-            length_function=len,
-        )
-        doc_chunks = splitter.split_documents(docs)
+        documents = self.load_documents(docs_folder)
+
+        if not self.parse_as_faq_pairs:
+            splitter = RecursiveCharacterTextSplitter(
+                chunk_size=self.chunk_size,
+                chunk_overlap=self.chunk_overlap,
+                length_function=len,
+            )
+            parsed_documents = splitter.split_documents(documents)
+        else:
+            parsed_documents = _format_faq_documents(documents)
 
         logger.info(
             "information_retrieval.faiss_store._create_document_index",
-            len_chunks=len(doc_chunks),
+            len_chunks=len(parsed_documents),
         )
-        if doc_chunks:
-            texts = [chunk.page_content for chunk in doc_chunks]
-            metadatas = [chunk.metadata for chunk in doc_chunks]
+        if parsed_documents:
+            texts = [document.page_content for document in parsed_documents]
+            metadatas = [document.metadata for document in parsed_documents]
             return FAISS.from_texts(texts, embedding, metadatas=metadatas, ids=None)
         else:
             raise ValueError(f"No documents found at '{docs_folder}'.")