General anthropic functions, steps towards experimental integration tests (#11727)

To match change in js here
https://github.com/langchain-ai/langchainjs/pull/2892

Some integration tests need a bit more work in experimental:
![Screenshot 2023-10-12 at 12 02 49
PM](https://github.com/langchain-ai/langchain/assets/9557659/262d7d22-c405-40e9-afef-669e8d585307)

Pretty sure the sqldatabase ones are an actual regression or change in
interface because it's returning a placeholder.

---------

Co-authored-by: Bagatur <baskaryan@gmail.com>

libs/experimental/Makefile

@@ -18,6 +18,9 @@ test_watch:
 extended_tests:
 	poetry run pytest --only-extended tests/unit_tests
 
+integration_tests:
+	poetry run pytest tests/integration_tests
+
 
 ######################
 # LINTING AND FORMATTING

libs/experimental/langchain_experimental/llms/anthropic_functions.py

@@ -124,11 +124,17 @@ def _destrip(tool_input: Any) -> Any:
 
 
 class AnthropicFunctions(BaseChatModel):
-    model: ChatAnthropic
+    llm: BaseChatModel
 
     @root_validator(pre=True)
     def validate_environment(cls, values: Dict) -> Dict:
-        return {"model": ChatAnthropic(**values)}
+        values["llm"] = values.get("llm") or ChatAnthropic(**values)
+        return values
+
+    @property
+    def model(self) -> BaseChatModel:
+        """For backwards compatibility."""
+        return self.llm
 
     def _generate(
         self,

libs/experimental/pyproject.toml

@@ -84,5 +84,6 @@ addopts = "--strict-markers --strict-config --durations=5"
 # Registering custom markers.
 # https://docs.pytest.org/en/7.1.x/example/markers.html#registering-markers
 markers = [
-  "requires: mark tests as requiring a specific library"
+  "requires: mark tests as requiring a specific library",
+  "asyncio: mark tests as requiring asyncio"
 ]

libs/experimental/tests/integration_tests/chains/test_cpal.py

@@ -39,7 +39,7 @@ from langchain_experimental.cpal.templates.univariate.narrative import (
 from langchain_experimental.cpal.templates.univariate.query import (
     template as query_template,
 )
-from tests.unit_tests.llms.fake_llm import FakeLLM
+from tests.unit_tests.fake_llm import FakeLLM
 
 
 class TestUnitCPALChain_MathWordProblems(unittest.TestCase):

libs/experimental/tests/integration_tests/chains/test_pal.py

@@ -1,8 +1,9 @@
 """Test PAL chain."""
 
-from langchain.chains.pal.base import PALChain
 from langchain.llms import OpenAI
 
+from langchain_experimental.pal_chain.base import PALChain
+
 
 def test_math_prompt() -> None:
     """Test math prompt."""

libs/experimental/tests/integration_tests/chains/test_sql_database.py

@@ -1,11 +1,12 @@
 """Test SQL Database Chain."""
 from langchain.llms.openai import OpenAI
 from langchain.utilities.sql_database import SQLDatabase
-from libs.experimental.langchain_experimental.sql.base import (
+from sqlalchemy import Column, Integer, MetaData, String, Table, create_engine, insert
+
+from langchain_experimental.sql.base import (
     SQLDatabaseChain,
     SQLDatabaseSequentialChain,
 )
-from sqlalchemy import Column, Integer, MetaData, String, Table, create_engine, insert
 
 metadata_obj = MetaData()
 

libs/experimental/tests/integration_tests/llms/test_anthropic_functions.py

@@ -0,0 +1,109 @@
+"""Test AnthropicFunctions"""
+
+import unittest
+
+from langchain.chat_models.anthropic import ChatAnthropic
+from langchain.chat_models.bedrock import BedrockChat
+
+from langchain_experimental.llms.anthropic_functions import AnthropicFunctions
+
+
+class TestAnthropicFunctions(unittest.TestCase):
+    """
+    Test AnthropicFunctions with default llm (ChatAnthropic) as well as a passed-in llm
+    """
+
+    def test_default_chat_anthropic(self) -> None:
+        base_model = AnthropicFunctions(model="claude-2")
+        self.assertIsInstance(base_model.model, ChatAnthropic)
+
+        # bind functions
+        model = base_model.bind(
+            functions=[
+                {
+                    "name": "get_current_weather",
+                    "description": "Get the current weather in a given location",
+                    "parameters": {
+                        "type": "object",
+                        "properties": {
+                            "location": {
+                                "type": "string",
+                                "description": "The city and state, "
+                                "e.g. San Francisco, CA",
+                            },
+                            "unit": {
+                                "type": "string",
+                                "enum": ["celsius", "fahrenheit"],
+                            },
+                        },
+                        "required": ["location"],
+                    },
+                }
+            ],
+            function_call={"name": "get_current_weather"},
+        )
+
+        res = model.invoke("What's the weather in San Francisco?")
+
+        function_call = res.additional_kwargs.get("function_call")
+        assert function_call
+        self.assertEqual(function_call.get("name"), "get_current_weather")
+        self.assertEqual(
+            function_call.get("arguments"),
+            '{"location": "San Francisco, CA", "unit": "fahrenheit"}',
+        )
+
+    def test_bedrock_chat_anthropic(self) -> None:
+        """
+              const chatBedrock = new ChatBedrock({
+          region: process.env.BEDROCK_AWS_REGION ?? "us-east-1",
+          model: "anthropic.claude-v2",
+          temperature: 0.1,
+          credentials: {
+            secretAccessKey: process.env.BEDROCK_AWS_SECRET_ACCESS_KEY!,
+            accessKeyId: process.env.BEDROCK_AWS_ACCESS_KEY_ID!,
+          },
+        });"""
+        llm = BedrockChat(
+            model_id="anthropic.claude-v2",
+            model_kwargs={"temperature": 0.1},
+            region_name="us-east-1",
+        )
+        base_model = AnthropicFunctions(llm=llm)
+        assert isinstance(base_model.model, BedrockChat)
+
+        # bind functions
+        model = base_model.bind(
+            functions=[
+                {
+                    "name": "get_current_weather",
+                    "description": "Get the current weather in a given location",
+                    "parameters": {
+                        "type": "object",
+                        "properties": {
+                            "location": {
+                                "type": "string",
+                                "description": "The city and state, "
+                                "e.g. San Francisco, CA",
+                            },
+                            "unit": {
+                                "type": "string",
+                                "enum": ["celsius", "fahrenheit"],
+                            },
+                        },
+                        "required": ["location"],
+                    },
+                }
+            ],
+            function_call={"name": "get_current_weather"},
+        )
+
+        res = model.invoke("What's the weather in San Francisco?")
+
+        function_call = res.additional_kwargs.get("function_call")
+        assert function_call
+        self.assertEqual(function_call.get("name"), "get_current_weather")
+        self.assertEqual(
+            function_call.get("arguments"),
+            '{"location": "San Francisco, CA", "unit": "fahrenheit"}',
+        )