rasa-pro 3.9.17__py3-none-any.whl → 3.10.3__py3-none-any.whl

rasa/core/featurizers/tracker_featurizers.py

@@ -1,9 +1,11 @@
 from __future__ import annotations
-
-import logging
-from abc import abstractmethod
-from collections import defaultdict
 from pathlib import Path
+from collections import defaultdict
+from abc import abstractmethod
+import jsonpickle
+import logging
+
+from tqdm import tqdm
 from typing import (
     Tuple,
     List,
@@ -16,30 +18,25 @@ from typing import (
     Set,
     DefaultDict,
     cast,
-    Type,
-    Callable,
-    ClassVar,
 )
-
 import numpy as np
-from tqdm import tqdm
 
+from rasa.core.featurizers.single_state_featurizer import SingleStateFeaturizer
+from rasa.core.featurizers.precomputation import MessageContainerForCoreFeaturization
+from rasa.core.exceptions import InvalidTrackerFeaturizerUsageError
 import rasa.shared.core.trackers
 import rasa.shared.utils.io
-from rasa.core.exceptions import InvalidTrackerFeaturizerUsageError
-from rasa.core.featurizers.precomputation import MessageContainerForCoreFeaturization
-from rasa.core.featurizers.single_state_featurizer import SingleStateFeaturizer
+from rasa.shared.nlu.constants import TEXT, INTENT, ENTITIES, ACTION_NAME
+from rasa.shared.nlu.training_data.features import Features
+from rasa.shared.core.trackers import DialogueStateTracker
+from rasa.shared.core.domain import State, Domain
+from rasa.shared.core.events import Event, ActionExecuted, UserUttered
 from rasa.shared.core.constants import (
     USER,
     ACTION_UNLIKELY_INTENT_NAME,
     PREVIOUS_ACTION,
 )
-from rasa.shared.core.domain import State, Domain
-from rasa.shared.core.events import Event, ActionExecuted, UserUttered
-from rasa.shared.core.trackers import DialogueStateTracker
 from rasa.shared.exceptions import RasaException
-from rasa.shared.nlu.constants import TEXT, INTENT, ENTITIES, ACTION_NAME
-from rasa.shared.nlu.training_data.features import Features
 from rasa.utils.tensorflow.constants import LABEL_PAD_ID
 from rasa.utils.tensorflow.model_data import ragged_array_to_ndarray
 
@@ -67,10 +64,6 @@ class InvalidStory(RasaException):
 class TrackerFeaturizer:
     """Base class for actual tracker featurizers."""
 
-    # Class registry to store all subclasses
-    _registry: ClassVar[Dict[str, Type["TrackerFeaturizer"]]] = {}
-    _featurizer_type: str = "TrackerFeaturizer"
-
     def __init__(
         self, state_featurizer: Optional[SingleStateFeaturizer] = None
     ) -> None:
@@ -81,36 +74,6 @@ class TrackerFeaturizer:
         """
         self.state_featurizer = state_featurizer
 
-    @classmethod
-    def register(cls, featurizer_type: str) -> Callable:
-        """Decorator to register featurizer subclasses."""
-
-        def wrapper(subclass: Type["TrackerFeaturizer"]) -> Type["TrackerFeaturizer"]:
-            cls._registry[featurizer_type] = subclass
-            # Store the type identifier in the class for serialization
-            subclass._featurizer_type = featurizer_type
-            return subclass
-
-        return wrapper
-
-    @classmethod
-    def from_dict(cls, data: Dict[str, Any]) -> "TrackerFeaturizer":
-        """Create featurizer instance from dictionary."""
-        featurizer_type = data.pop("type")
-
-        if featurizer_type not in cls._registry:
-            raise ValueError(f"Unknown featurizer type: {featurizer_type}")
-
-        # Get the correct subclass and instantiate it
-        subclass = cls._registry[featurizer_type]
-        return subclass.create_from_dict(data)
-
-    @classmethod
-    @abstractmethod
-    def create_from_dict(cls, data: Dict[str, Any]) -> "TrackerFeaturizer":
-        """Each subclass must implement its own creation from dict method."""
-        pass
-
     @staticmethod
     def _create_states(
         tracker: DialogueStateTracker,
@@ -502,7 +465,9 @@ class TrackerFeaturizer:
             self.state_featurizer.entity_tag_specs = []
 
         # noinspection PyTypeChecker
-        rasa.shared.utils.io.dump_obj_as_json_to_file(featurizer_file, self.to_dict())
+        rasa.shared.utils.io.write_text_file(
+            str(jsonpickle.encode(self)), featurizer_file
+        )
 
     @staticmethod
     def load(path: Union[Text, Path]) -> Optional[TrackerFeaturizer]:
@@ -516,17 +481,7 @@ class TrackerFeaturizer:
         """
         featurizer_file = Path(path) / FEATURIZER_FILE
         if featurizer_file.is_file():
-            data = rasa.shared.utils.io.read_json_file(featurizer_file)
-
-            if "type" not in data:
-                logger.error(
-                    f"Couldn't load featurizer for policy. "
-                    f"File '{featurizer_file}' does not contain all "
-                    f"necessary information. 'type' is missing."
-                )
-                return None
-
-            return TrackerFeaturizer.from_dict(data)
+            return jsonpickle.decode(rasa.shared.utils.io.read_file(featurizer_file))
 
         logger.error(
             f"Couldn't load featurizer for policy. "
@@ -553,16 +508,7 @@ class TrackerFeaturizer:
             )
         ]
 
-    def to_dict(self) -> Dict[str, Any]:
-        return {
-            "type": self.__class__._featurizer_type,
-            "state_featurizer": (
-                self.state_featurizer.to_dict() if self.state_featurizer else None
-            ),
-        }
-
 
-@TrackerFeaturizer.register("FullDialogueTrackerFeaturizer")
 class FullDialogueTrackerFeaturizer(TrackerFeaturizer):
     """Creates full dialogue training data for time distributed architectures.
 
@@ -700,20 +646,7 @@ class FullDialogueTrackerFeaturizer(TrackerFeaturizer):
 
         return trackers_as_states
 
-    def to_dict(self) -> Dict[str, Any]:
-        return super().to_dict()
 
-    @classmethod
-    def create_from_dict(cls, data: Dict[str, Any]) -> "FullDialogueTrackerFeaturizer":
-        state_featurizer = SingleStateFeaturizer.create_from_dict(
-            data["state_featurizer"]
-        )
-        return cls(
-            state_featurizer,
-        )
-
-
-@TrackerFeaturizer.register("MaxHistoryTrackerFeaturizer")
 class MaxHistoryTrackerFeaturizer(TrackerFeaturizer):
     """Truncates the tracker history into `max_history` long sequences.
 
@@ -951,25 +884,7 @@ class MaxHistoryTrackerFeaturizer(TrackerFeaturizer):
 
         return trackers_as_states
 
-    def to_dict(self) -> Dict[str, Any]:
-        data = super().to_dict()
-        data.update(
-            {
-                "remove_duplicates": self.remove_duplicates,
-                "max_history": self.max_history,
-            }
-        )
-        return data
-
-    @classmethod
-    def create_from_dict(cls, data: Dict[str, Any]) -> "MaxHistoryTrackerFeaturizer":
-        state_featurizer = SingleStateFeaturizer.create_from_dict(
-            data["state_featurizer"]
-        )
-        return cls(state_featurizer, data["max_history"], data["remove_duplicates"])
 
-
-@TrackerFeaturizer.register("IntentMaxHistoryTrackerFeaturizer")
 class IntentMaxHistoryTrackerFeaturizer(MaxHistoryTrackerFeaturizer):
     """Truncates the tracker history into `max_history` long sequences.
 
@@ -1244,18 +1159,6 @@ class IntentMaxHistoryTrackerFeaturizer(MaxHistoryTrackerFeaturizer):
 
         return trackers_as_states
 
-    def to_dict(self) -> Dict[str, Any]:
-        return super().to_dict()
-
-    @classmethod
-    def create_from_dict(
-        cls, data: Dict[str, Any]
-    ) -> "IntentMaxHistoryTrackerFeaturizer":
-        state_featurizer = SingleStateFeaturizer.create_from_dict(
-            data["state_featurizer"]
-        )
-        return cls(state_featurizer, data["max_history"], data["remove_duplicates"])
-
 
 def _is_prev_action_unlikely_intent_in_state(state: State) -> bool:
     prev_action_name = state.get(PREVIOUS_ACTION, {}).get(ACTION_NAME)

rasa/core/information_retrieval/faiss.py

@@ -2,9 +2,10 @@ from pathlib import Path
 from typing import TYPE_CHECKING, List, Optional, Text, Any, Dict
 
 import structlog
-from langchain.document_loaders import DirectoryLoader, TextLoader
 from langchain.text_splitter import RecursiveCharacterTextSplitter
-from langchain.vectorstores.faiss import FAISS
+from langchain_community.document_loaders.text import TextLoader
+from langchain_community.document_loaders.directory import DirectoryLoader
+from langchain_community.vectorstores.faiss import FAISS
 from rasa.utils.endpoints import EndpointConfig
 
 from rasa.core.information_retrieval import (
@@ -46,7 +47,9 @@ class FAISS_Store(InformationRetrieval):
             logger.info(
                 "information_retrieval.faiss_store.load_index", path=path.absolute()
             )
-            self.index = FAISS.load_local(str(path), embeddings)
+            self.index = FAISS.load_local(
+                str(path), embeddings, allow_dangerous_deserialization=True
+            )
 
     @staticmethod
     def load_documents(docs_folder: str) -> List["Document"]:
@@ -114,7 +117,7 @@ class FAISS_Store(InformationRetrieval):
     ) -> SearchResultList:
         logger.debug("information_retrieval.faiss_store.search", query=query)
         try:
-            documents = await self.index.as_retriever().aget_relevant_documents(query)
+            documents = await self.index.as_retriever().ainvoke(query)
         except Exception as exc:
             raise InformationRetrievalException from exc
 

rasa/core/information_retrieval/information_retrieval.py

@@ -19,6 +19,14 @@ logger = structlog.get_logger()
 
 @dataclass
 class SearchResult:
+    """A search result object.
+
+    Attributes:
+        text: The text content of the retrieved document result.
+        metadata: The metadata associated with the document result.
+        score: The score of the search result.
+    """
+
     text: str
     metadata: dict
     score: Optional[float] = None

rasa/core/information_retrieval/milvus.py

@@ -1,7 +1,7 @@
 from typing import Text, Any, Dict
 
 import structlog
-from langchain.vectorstores.milvus import Milvus
+from langchain_community.vectorstores.milvus import Milvus
 from rasa.utils.endpoints import EndpointConfig
 
 from rasa.core.information_retrieval import (
@@ -48,5 +48,12 @@ class Milvus_Store(InformationRetrieval):
         except Exception as exc:
             raise InformationRetrievalException from exc
 
-        filtered_hits = [doc for doc, score in hits if score >= threshold]
+        scores = [score for _, score in hits]
+        logger.debug(
+            "information_retrieval.milvus_store.search_results_before_threshold",
+            scores=scores,
+        )
+        # Milvus uses Euclidean distance metric by default
+        # so the lower the score, the better the match.
+        filtered_hits = [doc for doc, score in hits if score <= threshold]
         return SearchResultList.from_document_list(filtered_hits)

rasa/core/information_retrieval/qdrant.py

@@ -1,7 +1,7 @@
 from typing import Text, Any, Dict
 
 import structlog
-from langchain.vectorstores.qdrant import Qdrant
+from langchain_community.vectorstores.qdrant import Qdrant
 from pydantic import ValidationError
 from qdrant_client import QdrantClient
 from rasa.utils.endpoints import EndpointConfig

rasa/core/nlg/contextual_response_rephraser.py

@@ -5,6 +5,15 @@ from jinja2 import Template
 
 from rasa import telemetry
 from rasa.core.nlg.response import TemplatedNaturalLanguageGenerator
+from rasa.shared.constants import (
+    LLM_CONFIG_KEY,
+    MODEL_CONFIG_KEY,
+    MODEL_NAME_CONFIG_KEY,
+    PROMPT_CONFIG_KEY,
+    PROVIDER_CONFIG_KEY,
+    OPENAI_PROVIDER,
+    TIMEOUT_CONFIG_KEY,
+)
 from rasa.shared.core.domain import KEY_RESPONSES_TEXT, Domain
 from rasa.shared.core.events import BotUttered, UserUttered
 from rasa.shared.core.trackers import DialogueStateTracker
@@ -15,6 +24,7 @@ from rasa.shared.utils.llm import (
     combine_custom_and_default_config,
     get_prompt_template,
     llm_factory,
+    try_instantiate_llm_client,
 )
 from rasa.utils.endpoints import EndpointConfig
 
@@ -31,11 +41,11 @@ RESPONSE_REPHRASING_TEMPLATE_KEY = "rephrase_prompt"
 DEFAULT_REPHRASE_ALL = False
 
 DEFAULT_LLM_CONFIG = {
-    "_type": "openai",
-    "request_timeout": 5,
+    PROVIDER_CONFIG_KEY: OPENAI_PROVIDER,
+    MODEL_CONFIG_KEY: DEFAULT_OPENAI_GENERATE_MODEL_NAME,
     "temperature": 0.3,
-    "model_name": DEFAULT_OPENAI_GENERATE_MODEL_NAME,
     "max_tokens": DEFAULT_OPENAI_MAX_GENERATED_TOKENS,
+    TIMEOUT_CONFIG_KEY: 5,
 }
 
 DEFAULT_RESPONSE_VARIATION_PROMPT_TEMPLATE = """The following is a conversation with
@@ -78,7 +88,7 @@ class ContextualResponseRephraser(TemplatedNaturalLanguageGenerator):
 
         self.nlg_endpoint = endpoint_config
         self.prompt_template = get_prompt_template(
-            self.nlg_endpoint.kwargs.get("prompt"),
+            self.nlg_endpoint.kwargs.get(PROMPT_CONFIG_KEY),
             DEFAULT_RESPONSE_VARIATION_PROMPT_TEMPLATE,
         )
         self.rephrase_all = self.nlg_endpoint.kwargs.get(
@@ -87,6 +97,12 @@ class ContextualResponseRephraser(TemplatedNaturalLanguageGenerator):
         self.trace_prompt_tokens = self.nlg_endpoint.kwargs.get(
             "trace_prompt_tokens", False
         )
+        try_instantiate_llm_client(
+            self.nlg_endpoint.kwargs.get(LLM_CONFIG_KEY),
+            DEFAULT_LLM_CONFIG,
+            "contextual_response_rephraser.init",
+            "ContextualResponseRephraser",
+        )
 
     def _last_message_if_human(self, tracker: DialogueStateTracker) -> Optional[str]:
         """Returns the latest message from the tracker.
@@ -115,10 +131,13 @@ class ContextualResponseRephraser(TemplatedNaturalLanguageGenerator):
         Returns:
             generated text
         """
-        llm = llm_factory(self.nlg_endpoint.kwargs.get("llm"), DEFAULT_LLM_CONFIG)
+        llm = llm_factory(
+            self.nlg_endpoint.kwargs.get(LLM_CONFIG_KEY), DEFAULT_LLM_CONFIG
+        )
 
         try:
-            return await llm.apredict(prompt)
+            llm_response = await llm.acompletion(prompt)
+            return llm_response.choices[0]
         except Exception as e:
             # unfortunately, langchain does not wrap LLM exceptions which means
             # we have to catch all exceptions here
@@ -128,7 +147,7 @@ class ContextualResponseRephraser(TemplatedNaturalLanguageGenerator):
     def llm_property(self, prop: str) -> Optional[str]:
         """Returns a property of the LLM provider."""
         return combine_custom_and_default_config(
-            self.nlg_endpoint.kwargs.get("llm"), DEFAULT_LLM_CONFIG
+            self.nlg_endpoint.kwargs.get(LLM_CONFIG_KEY), DEFAULT_LLM_CONFIG
         ).get(prop)
 
     def custom_prompt_template(self, prompt_template: str) -> Optional[str]:
@@ -161,7 +180,9 @@ class ContextualResponseRephraser(TemplatedNaturalLanguageGenerator):
         Returns:
         The history for the prompt.
         """
-        llm = llm_factory(self.nlg_endpoint.kwargs.get("llm"), DEFAULT_LLM_CONFIG)
+        llm = llm_factory(
+            self.nlg_endpoint.kwargs.get(LLM_CONFIG_KEY), DEFAULT_LLM_CONFIG
+        )
         return await summarize_conversation(tracker, llm, max_turns=5)
 
     async def rephrase(
@@ -202,8 +223,9 @@ class ContextualResponseRephraser(TemplatedNaturalLanguageGenerator):
         telemetry.track_response_rephrase(
             rephrase_all=self.rephrase_all,
             custom_prompt_template=self.custom_prompt_template(prompt_template_text),
-            llm_type=self.llm_property("_type"),
-            llm_model=self.llm_property("model") or self.llm_property("model_name"),
+            llm_type=self.llm_property(PROVIDER_CONFIG_KEY),
+            llm_model=self.llm_property(MODEL_CONFIG_KEY)
+            or self.llm_property(MODEL_NAME_CONFIG_KEY),
         )
         if not (updated_text := await self._generate_llm_response(prompt)):
             # If the LLM fails to generate a response, we

rasa/core/nlg/summarize.py

@@ -2,8 +2,8 @@ from typing import Optional
 
 import structlog
 from jinja2 import Template
-from langchain.llms.base import BaseLLM
 from rasa.core.tracker_store import DialogueStateTracker
+from rasa.shared.providers.llm.llm_client import LLMClient
 from rasa.shared.utils.llm import (
     tracker_as_readable_transcript,
 )
@@ -43,7 +43,7 @@ def _create_summarization_prompt(
 
 async def summarize_conversation(
     tracker: DialogueStateTracker,
-    llm: BaseLLM,
+    llm: LLMClient,
     max_turns: Optional[int] = MAX_TURNS_DEFAULT,
 ) -> str:
     """Summarizes the dialogue using the LLM.
@@ -58,7 +58,8 @@ async def summarize_conversation(
     """
     prompt = _create_summarization_prompt(tracker, max_turns)
     try:
-        summarization = (await llm.apredict(prompt)).strip()
+        llm_response = await llm.acompletion(prompt)
+        summarization = llm_response.choices[0].strip()
         structlogger.debug(
             "summarization.success", summarization=summarization, prompt=prompt
         )