Merge branch 'master' into google-provider

docs/docs/integrations/llms/tongyi.ipynb

@@ -8,6 +8,13 @@
     "Tongyi Qwen is a large-scale language model developed by Alibaba's Damo Academy. It is capable of understanding user intent through natural language understanding and semantic analysis, based on user input in natural language. It provides services and assistance to users in different domains and tasks. By providing clear and detailed instructions, you can obtain results that better align with your expectations."
    ]
   },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "## Setting up"
+   ]
+  },
   {
    "cell_type": "code",
    "execution_count": null,
@@ -25,7 +32,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 2,
+   "execution_count": 1,
    "metadata": {
     "ExecuteTime": {
      "end_time": "2023-07-10T19:55:38.553933Z",
@@ -34,10 +41,10 @@
    },
    "outputs": [
     {
-     "name": "stdout",
+     "name": "stdin",
      "output_type": "stream",
      "text": [
-      "········\n"
+      " ········\n"
      ]
     }
    ],
@@ -50,7 +57,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 3,
+   "execution_count": 2,
    "metadata": {
     "ExecuteTime": {
      "end_time": "2023-07-10T19:55:38.554152Z",
@@ -66,7 +73,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 4,
+   "execution_count": 3,
    "metadata": {
     "ExecuteTime": {
      "end_time": "2023-07-10T19:55:39.812664Z",
@@ -75,14 +82,57 @@
    },
    "outputs": [],
    "source": [
-    "from langchain.chains import LLMChain\n",
-    "from langchain.prompts import PromptTemplate\n",
     "from langchain_community.llms import Tongyi"
    ]
   },
+  {
+   "cell_type": "code",
+   "execution_count": 4,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "'Justin Bieber was born on March 1, 1994. The Super Bowl that took place in the same year was Super Bowl XXVIII, which was played on January 30, 1994. The winner of that Super Bowl was the Dallas Cowboys, who defeated the Buffalo Bills with a score of 30-13.'"
+      ]
+     },
+     "execution_count": 4,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "Tongyi().invoke(\"What NFL team won the Super Bowl in the year Justin Bieber was born?\")"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "## Using in a chain"
+   ]
+  },
   {
    "cell_type": "code",
    "execution_count": 5,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "from langchain_core.prompts import PromptTemplate"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 6,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "llm = Tongyi()"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 7,
    "metadata": {
     "ExecuteTime": {
      "end_time": "2023-07-10T19:55:39.817327Z",
@@ -100,50 +150,34 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 6,
-   "metadata": {},
-   "outputs": [],
-   "source": [
-    "llm = Tongyi()"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 7,
+   "execution_count": 8,
    "metadata": {},
    "outputs": [],
    "source": [
-    "llm_chain = LLMChain(prompt=prompt, llm=llm)"
+    "chain = prompt | llm"
    ]
   },
   {
    "cell_type": "code",
-   "execution_count": 8,
+   "execution_count": 9,
    "metadata": {},
    "outputs": [
     {
      "data": {
       "text/plain": [
-       "\"The year Justin Bieber was born was 1994. The Denver Broncos won the Super Bowl in 1997, which means they would have been the team that won the Super Bowl during Justin Bieber's birth year. So the answer is the Denver Broncos.\""
+       "'Justin Bieber was born on March 1, 1994. The Super Bowl that took place in the same calendar year was Super Bowl XXVIII, which was played on January 30, 1994. The winner of Super Bowl XXVIII was the Dallas Cowboys, who defeated the Buffalo Bills with a score of 30-13.'"
       ]
      },
-     "execution_count": 8,
+     "execution_count": 9,
      "metadata": {},
      "output_type": "execute_result"
     }
    ],
    "source": [
-    "question = \"What NFL team won the Super Bowl in the year Justin Beiber was born?\"\n",
+    "question = \"What NFL team won the Super Bowl in the year Justin Bieber was born?\"\n",
     "\n",
-    "llm_chain.run(question)"
+    "chain.invoke({\"question\": question})"
    ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": null,
-   "metadata": {},
-   "outputs": [],
-   "source": []
   }
  ],
  "metadata": {
@@ -162,9 +196,9 @@
    "name": "python",
    "nbconvert_exporter": "python",
    "pygments_lexer": "ipython3",
-   "version": "3.10.12"
+   "version": "3.11.7"
   }
  },
  "nbformat": 4,
- "nbformat_minor": 1
+ "nbformat_minor": 4
 }

docs/docs/integrations/platforms/index.mdx

@@ -11,15 +11,21 @@ LangChain integrates with many providers.
 
 These providers have standalone `langchain-{provider}` packages for improved versioning, dependency management and testing.
 
-- [OpenAI](/docs/integrations/platforms/openai)
+- [AI21](/docs/integrations/providers/ai21)
 - [Anthropic](/docs/integrations/platforms/anthropic)
-- [Google](/docs/integrations/platforms/google)
-- [MistralAI](/docs/integrations/providers/mistralai)
-- [NVIDIA AI](/docs/integrations/providers/nvidia)
-- [Together AI](/docs/integrations/providers/together)
-- [Robocorp](/docs/integrations/providers/robocorp)
+- [Astra DB](/docs/integrations/providers/astradb)
 - [Exa Search](/docs/integrations/providers/exa_search)
+- [Google Generative AI](/docs/integrations/platforms/google)
+- [Google Vertex AI](/docs/integrations/platforms/google)
+- [IBM](/docs/integrations/providers/ibm)
+- [MistralAI](/docs/integrations/providers/mistralai)
 - [Nomic](/docs/integrations/providers/nomic)
+- [Nvidia AI Endpoints](/docs/integrations/providers/nvidia)
+- [Nvidia AI](/docs/integrations/providers/nvidia)
+- [OpenAI](/docs/integrations/platforms/openai)
+- [Pinecone](/docs/integrations/providers/pinecone)
+- [Robocorp](/docs/integrations/providers/robocorp)
+- [Together AI](/docs/integrations/providers/together)
 
 
 ## Featured Community Providers

docs/docusaurus.config.js

@@ -13,7 +13,7 @@ const baseDarkCodeBlockTheme = require("prism-react-renderer/themes/vsDark");
 const config = {
   title: "🦜️🔗 Langchain",
   tagline: "LangChain Python Docs",
-  favicon: "img/favicon.ico",
+  favicon: "img/brand/favicon.png",
   // Set the production url of your site here
   url: "https://python.langchain.com",
   // Set the /<baseUrl>/ pathname under which your site is served
@@ -144,9 +144,9 @@ const config = {
           },
         },
       },
-      image: "img/parrot-chainlink-icon.png",
+      image: "img/brand/theme-image.png",
       navbar: {
-        title: "🦜️🔗 LangChain",
+        logo: {src: "img/brand/wordmark.png", srcDark: "img/brand/wordmark-dark.png"},
         items: [
           {
             to: "/docs/get_started/introduction",

libs/community/langchain_community/vectorstores/thirdai_neuraldb.py

@@ -26,7 +26,7 @@ class NeuralDBVectorStore(VectorStore):
             vectorstore = NeuralDBVectorStore(db=db)
     """
 
-    def __init__(self, db):
+    def __init__(self, db: Any) -> None:
         self.db = db
 
     db: Any = None  #: :meta private:

libs/core/langchain_core/language_models/base.py

@@ -5,24 +5,27 @@
 from typing import (
     TYPE_CHECKING,
     Any,
+    Dict,
     List,
     Optional,
     Sequence,
     Set,
+    Type,
     TypeVar,
     Union,
 )
 
 from typing_extensions import TypeAlias
 
-from langchain_core._api import deprecated
+from langchain_core._api import beta, deprecated
 from langchain_core.messages import (
     AnyMessage,
     BaseMessage,
     MessageLikeRepresentation,
     get_buffer_string,
 )
 from langchain_core.prompt_values import PromptValue
+from langchain_core.pydantic_v1 import BaseModel
 from langchain_core.runnables import Runnable, RunnableSerializable
 from langchain_core.utils import get_pydantic_field_names
 
@@ -155,6 +158,13 @@ async def agenerate_prompt(
                 prompt and additional model provider-specific output.
         """
 
+    @beta()
+    def with_structured_output(
+        self, schema: Union[Dict, Type[BaseModel]], **kwargs: Any
+    ) -> Runnable[LanguageModelInput, Union[Dict, BaseModel]]:
+        """Implement this if there is a way of steering the model to generate responses that match a given schema."""  # noqa: E501
+        raise NotImplementedError()
+
     @deprecated("0.1.7", alternative="invoke", removal="0.2.0")
     @abstractmethod
     def predict(

libs/core/langchain_core/output_parsers/__init__.py

@@ -24,6 +24,7 @@
     MarkdownListOutputParser,
     NumberedListOutputParser,
 )
+from langchain_core.output_parsers.pydantic import PydanticOutputParser
 from langchain_core.output_parsers.string import StrOutputParser
 from langchain_core.output_parsers.transform import (
     BaseCumulativeTransformOutputParser,
@@ -45,4 +46,5 @@
     "SimpleJsonOutputParser",
     "XMLOutputParser",
     "JsonOutputParser",
+    "PydanticOutputParser",
 ]

libs/core/langchain_core/output_parsers/base.py

@@ -15,15 +15,17 @@
 
 from typing_extensions import get_args
 
+from langchain_core.language_models import LanguageModelOutput
 from langchain_core.messages import AnyMessage, BaseMessage
 from langchain_core.outputs import ChatGeneration, Generation
-from langchain_core.runnables import RunnableConfig, RunnableSerializable
+from langchain_core.runnables import Runnable, RunnableConfig, RunnableSerializable
 from langchain_core.runnables.config import run_in_executor
 
 if TYPE_CHECKING:
     from langchain_core.prompt_values import PromptValue
 
 T = TypeVar("T")
+OutputParserLike = Runnable[LanguageModelOutput, T]
 
 
 class BaseLLMOutputParser(Generic[T], ABC):
@@ -57,7 +59,7 @@ async def aparse_result(
 
 
 class BaseGenerationOutputParser(
-    BaseLLMOutputParser, RunnableSerializable[Union[str, BaseMessage], T]
+    BaseLLMOutputParser, RunnableSerializable[LanguageModelOutput, T]
 ):
     """Base class to parse the output of an LLM call."""
 
@@ -116,7 +118,7 @@ async def ainvoke(
 
 
 class BaseOutputParser(
-    BaseLLMOutputParser, RunnableSerializable[Union[str, BaseMessage], T]
+    BaseLLMOutputParser, RunnableSerializable[LanguageModelOutput, T]
 ):
     """Base class to parse the output of an LLM call.
 

libs/core/langchain_core/output_parsers/pydantic.py

@@ -0,0 +1,62 @@
+import json
+from typing import Any, List, Type
+
+from langchain_core.exceptions import OutputParserException
+from langchain_core.output_parsers import JsonOutputParser
+from langchain_core.outputs import Generation
+from langchain_core.pydantic_v1 import BaseModel, ValidationError
+
+
+class PydanticOutputParser(JsonOutputParser):
+    """Parse an output using a pydantic model."""
+
+    pydantic_object: Type[BaseModel]
+    """The pydantic model to parse.
+    
+    Attention: To avoid potential compatibility issues, it's recommended to use
+        pydantic <2 or leverage the v1 namespace in pydantic >= 2.
+    """
+
+    def parse_result(self, result: List[Generation], *, partial: bool = False) -> Any:
+        json_object = super().parse_result(result)
+        try:
+            return self.pydantic_object.parse_obj(json_object)
+        except ValidationError as e:
+            name = self.pydantic_object.__name__
+            msg = f"Failed to parse {name} from completion {json_object}. Got: {e}"
+            raise OutputParserException(msg, llm_output=json_object)
+
+    def get_format_instructions(self) -> str:
+        # Copy schema to avoid altering original Pydantic schema.
+        schema = {k: v for k, v in self.pydantic_object.schema().items()}
+
+        # Remove extraneous fields.
+        reduced_schema = schema
+        if "title" in reduced_schema:
+            del reduced_schema["title"]
+        if "type" in reduced_schema:
+            del reduced_schema["type"]
+        # Ensure json in context is well-formed with double quotes.
+        schema_str = json.dumps(reduced_schema)
+
+        return _PYDANTIC_FORMAT_INSTRUCTIONS.format(schema=schema_str)
+
+    @property
+    def _type(self) -> str:
+        return "pydantic"
+
+    @property
+    def OutputType(self) -> Type[BaseModel]:
+        """Return the pydantic model."""
+        return self.pydantic_object
+
+
+_PYDANTIC_FORMAT_INSTRUCTIONS = """The output should be formatted as a JSON instance that conforms to the JSON schema below.
+
+As an example, for the schema {{"properties": {{"foo": {{"title": "Foo", "description": "a list of strings", "type": "array", "items": {{"type": "string"}}}}}}, "required": ["foo"]}}
+the object {{"foo": ["bar", "baz"]}} is a well-formatted instance of the schema. The object {{"properties": {{"foo": ["bar", "baz"]}}}} is not well-formatted.
+
+Here is the output schema:
+```
+{schema}
+```"""  # noqa: E501