fix: added missing query tool, updated readme

.env

@@ -17,7 +17,7 @@ EMBEDDING_DIM=1024
 # CONVERSATION_STARTERS=
 
 # The OpenAI API key to use.
-OPENAI_API_KEY=sk-proj-7qNCi2Btib0Y7AADy3tgT3BlbkFJ9d7IFV2lk2IfRnvDQbxp
+# OPENAI_API_KEY=
 
 # Temperature for sampling from the model.
 # LLM_TEMPERATURE=
@@ -40,18 +40,5 @@ APP_HOST=0.0.0.0
 # The port to start the backend app.
 APP_PORT=8000
 
-# MESSAGE_QUEUE_PORT=
-
-# CONTROL_PLANE_PORT=
-
-# HUMAN_CONSUMER_PORT=
-
-AGENT_QUERY_ENGINE_PORT=8003
-
-AGENT_QUERY_ENGINE_DESCRIPTION=Query information from the provided data
-
-AGENT_DUMMY_PORT=8004
-
 # The system prompt for the AI model.
 SYSTEM_PROMPT=You are a helpful assistant who helps users with their questions.
-

README.md

@@ -8,7 +8,6 @@ First, setup the environment with poetry:
 
 ```shell
 poetry install
-poetry shell
 ```
 
 Then check the parameters that have been pre-configured in the `.env` file in this directory. (E.g. you might need to configure an `OPENAI_API_KEY` if you're using OpenAI as model provider).
@@ -19,28 +18,12 @@ Second, generate the embeddings of the documents in the `./data` directory (if t
 poetry run generate
 ```
 
-Third, run all the services in one command:
+Third, run all the agents in one command:
 
 ```shell
 poetry run python main.py
 ```
 
-You can monitor and test the agent services with `llama-agents` monitor TUI:
-
-```shell
-poetry run llama-agents monitor --control-plane-url http://127.0.0.1:8001
-```
-
-## Services:
-
-- Message queue (port 8000): To exchange the message between services
-- Control plane (port 8001): A gateway to manage the tasks and services.
-- Human consumer (port 8002): To handle result when the task is completed.
-- Agent service `query_engine` (port 8003): Agent that can query information from the configured LlamaIndex index.
-- Agent service `dummy_agent` (port 8004): A dummy agent that does nothing. Good starting point to add more agents.
-
-The ports listed above are set by default, but you can change them in the `.env` file.
-
 ## Learn More
 
 To learn more about LlamaIndex, take a look at the following resources:

app/core/function_call.py

@@ -31,7 +31,7 @@ class FunctionCallingAgent(Workflow):
         tools: List[BaseTool] | None = None,
         system_prompt: str | None = None,
         verbose: bool = True,
-        timeout: float = 120.0,
+        timeout: float = 360.0,
         name: str,
         **kwargs: Any,
     ) -> None:

main.py

@@ -1,16 +1,39 @@
 # flake8: noqa: E402
 import asyncio
+import os
 from dotenv import load_dotenv
-from app.agents.researcher.agent import get_query_engine_tool
 from app.core.agent_call import AgentCallingAgent
 from app.core.function_call import FunctionCallingAgent
+from app.engine.index import get_index
 from app.settings import init_settings
-
+from llama_index.core.tools import QueryEngineTool, ToolMetadata
 
 load_dotenv()
 init_settings()
 
 
+def get_query_engine_tool() -> QueryEngineTool:
+    """
+    Provide an agent worker that can be used to query the index.
+    """
+    index = get_index()
+    if index is None:
+        raise ValueError("Index not found. Please create an index first.")
+    top_k = int(os.getenv("TOP_K", 0))
+    query_engine = index.as_query_engine(
+        **({"similarity_top_k": top_k} if top_k != 0 else {})
+    )
+    return QueryEngineTool(
+        query_engine=query_engine,
+        metadata=ToolMetadata(
+            name="query_index",
+            description="""
+                Use this tool to retrieve information about the text corpus from the index.
+            """,
+        ),
+    )
+
+
 async def main():
     researcher = FunctionCallingAgent(
         name="researcher",
@@ -24,6 +47,7 @@ async def main():
     writer = AgentCallingAgent(
         name="writer",
         agents=[researcher, reviewer],
+        verbose=False,
         system_prompt="""You are an expert in writing blog posts. You are given a task to write a blog post. Before starting to write the post, consult the researcher agent to get the information you need. Don't make up any information yourself.
         After creating a draft for the post, send it to the reviewer agent to receive some feedback and make sure to incorporate the feedback from the reviewer.
         You can consult the reviewer and researcher multiple times. Only finish the task once the reviewer is satisfied.""",