[agent][property type] Change allowed_tools to Set as Duplicate doesn…

…’t make sense (langchain-ai#3840)

- ActionAgent has a property called, `allowed_tools`, which is declared
as `List`. It stores all provided tools which is available to use during
agent action.
- This collection shouldn’t allow duplicates. The original datatype List
doesn’t make sense. Each tool should be unique. Even when there are
variants (assuming in the future), it would be named differently in
load_tools.


Test:
- confirm the functionality in an example by initializing an agent with
a list of 2 tools and confirm everything works.
```python3
def test_agent_chain_chat_bot():
	from langchain.agents import load_tools
	from langchain.agents import initialize_agent
	from langchain.agents import AgentType
	from langchain.chat_models import ChatOpenAI
	from langchain.llms import OpenAI
	from langchain.utilities.duckduckgo_search import DuckDuckGoSearchAPIWrapper

	chat = ChatOpenAI(temperature=0)
	llm = OpenAI(temperature=0)
	tools = load_tools(["ddg-search", "llm-math"], llm=llm)

	agent = initialize_agent(tools, chat, agent=AgentType.CHAT_ZERO_SHOT_REACT_DESCRIPTION, verbose=True)
	agent.run("Who is Olivia Wilde's boyfriend? What is his current age raised to the 0.23 power?")
test_agent_chain_chat_bot()
```
Result:
<img width="863" alt="Screenshot 2023-05-01 at 7 58 11 PM"
src="https://user-images.githubusercontent.com/62768671/235572157-0937594c-ddfb-4760-acb2-aea4cacacd89.png">

langchain/agents/agent.py

@@ -7,7 +7,7 @@
 import time
 from abc import abstractmethod
 from pathlib import Path
-from typing import Any, Dict, List, Optional, Sequence, Tuple, Union
+from typing import Any, Dict, List, Optional, Sequence, Set, Tuple, Union
 
 import yaml
 from pydantic import BaseModel, root_validator
@@ -46,8 +46,8 @@ def return_values(self) -> List[str]:
         """Return values of the agent."""
         return ["output"]
 
-    def get_allowed_tools(self) -> Optional[List[str]]:
-        return None
+    def get_allowed_tools(self) -> Set[str]:
+        return set()
 
     @abstractmethod
     def plan(
@@ -178,8 +178,8 @@ def return_values(self) -> List[str]:
         """Return values of the agent."""
         return ["output"]
 
-    def get_allowed_tools(self) -> Optional[List[str]]:
-        return None
+    def get_allowed_tools(self) -> Set[str]:
+        return set()
 
     @abstractmethod
     def plan(
@@ -372,9 +372,9 @@ class Agent(BaseSingleActionAgent):
 
     llm_chain: LLMChain
     output_parser: AgentOutputParser
-    allowed_tools: Optional[List[str]] = None
+    allowed_tools: Set[str] = set()
 
-    def get_allowed_tools(self) -> Optional[List[str]]:
+    def get_allowed_tools(self) -> Set[str]:
         return self.allowed_tools
 
     @property
@@ -607,12 +607,11 @@ def validate_tools(cls, values: Dict) -> Dict:
         agent = values["agent"]
         tools = values["tools"]
         allowed_tools = agent.get_allowed_tools()
-        if allowed_tools is not None:
-            if set(allowed_tools) != set([tool.name for tool in tools]):
-                raise ValueError(
-                    f"Allowed tools ({allowed_tools}) different than "
-                    f"provided tools ({[tool.name for tool in tools]})"
-                )
+        if allowed_tools != set([tool.name for tool in tools]):
+            raise ValueError(
+                f"Allowed tools ({allowed_tools}) different than "
+                f"provided tools ({[tool.name for tool in tools]})"
+            )
         return values
 
     @root_validator()

tests/integration_tests/agent/__init__.py

@@ -0,0 +1 @@
+"""All integration tests for agent."""

tests/integration_tests/agent/test_agent.py

@@ -0,0 +1,16 @@
+from langchain.agents.chat.base import ChatAgent
+from langchain.llms.openai import OpenAI
+from langchain.tools.ddg_search.tool import DuckDuckGoSearchRun
+
+
+class TestAgent:
+    def test_agent_generation(self) -> None:
+        web_search = DuckDuckGoSearchRun()
+        tools = [web_search]
+        agent = ChatAgent.from_llm_and_tools(
+            ai_name="Tom",
+            ai_role="Assistant",
+            tools=tools,
+            llm=OpenAI(maxTokens=10),
+        )
+        assert agent.allowed_tools == set([web_search.name])