Initiate Agent component (#228)

* initial agent component implementation

1. default uses langchain React Agent
2. optional: PlanExec agent provided based on Langgraph (initiate by Minmin)

Signed-off-by: Chendi Xue <[email protected]>

* Add UT and update README

Signed-off-by: Chendi.Xue <[email protected]>

* update README and test with Rag endpoint

Signed-off-by: Chendi.Xue <[email protected]>

* rename planexec to plan_execute

Signed-off-by: Chendi.Xue <[email protected]>

* provide react prompt locally

Signed-off-by: Chendi.Xue <[email protected]>

* update stream output with thoughts

Signed-off-by: Chendi.Xue <[email protected]>

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci

* Refactoring codes put strategy in seperate folders

Signed-off-by: Chendi.Xue <[email protected]>

* update huggingfacenedpoint parameter and fix agenticrag

Signed-off-by: Chendi.Xue <[email protected]>

* update agenticRag with Readme

Signed-off-by: Chendi.Xue <[email protected]>

* update interface

Signed-off-by: Chendi.Xue <[email protected]>

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci

---------

Signed-off-by: Chendi Xue <[email protected]>
Signed-off-by: Chendi.Xue <[email protected]>
Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>

comps/agent/langchain/README.md

@@ -0,0 +1,150 @@
+# langchain Agent Microservice
+
+The langchain agent model refers to a framework that integrates the reasoning capabilities of large language models (LLMs) with the ability to take actionable steps, creating a more sophisticated system that can understand and process information, evaluate situations, take appropriate actions, communicate responses, and track ongoing situations.
+
+![Architecture Overview](agent_arch.jpg)
+
+# 🚀1. Start Microservice with Python（Option 1）
+
+## 1.1 Install Requirements
+
+```bash
+cd comps/agent/langchain/
+pip install -r requirements.txt
+```
+
+## 1.2 Start Microservice with Python Script
+
+```bash
+cd comps/agent/langchain/
+python agent.py
+```
+
+# 🚀2. Start Microservice with Docker (Option 2)
+
+## Build Microservices
+
+```bash
+cd GenAIComps/ # back to GenAIComps/ folder
+docker build -t opea/comps-agent-langchain:latest -f comps/agent/langchain/docker/Dockerfile .
+```
+
+## start microservices
+
+```bash
+export ip_address=$(hostname -I | awk '{print $1}')
+export model=meta-llama/Meta-Llama-3-8B-Instruct
+export HUGGINGFACEHUB_API_TOKEN=${HUGGINGFACEHUB_API_TOKEN}
+export HF_TOKEN=${HUGGINGFACEHUB_API_TOKEN}
+
+# TGI serving
+docker run -d --runtime=habana --name "comps-tgi-gaudi-service" -p 8080:80 -v ./data:/data -e HF_TOKEN=$HF_TOKEN -e HABANA_VISIBLE_DEVICES=all -e OMPI_MCA_btl_vader_single_copy_mechanism=none --cap-add=sys_nice --ipc=host ghcr.io/huggingface/tgi-gaudi:latest --model-id $model --max-input-tokens 4096 --max-total-tokens 8092
+
+# check status
+docker logs comps-tgi-gaudi-service
+
+# Agent
+docker run -d --runtime=runc --name="comps-langchain-agent-endpoint" -v $WORKPATH/comps/agent/langchain/tools:/home/user/comps/agent/langchain/tools -p 9090:9090 --ipc=host -e HUGGINGFACEHUB_API_TOKEN=${HUGGINGFACEHUB_API_TOKEN} -e model=${model} -e ip_address=${ip_address} -e strategy=react -e llm_endpoint_url=http://${ip_address}:8080 -e llm_engine=tgi -e recursion_limit=5 -e require_human_feedback=false -e tools=/home/user/comps/agent/langchain/tools/custom_tools.yaml opea/comps-agent-langchain:latest
+
+# check status
+docker logs comps-langchain-agent-endpoint
+```
+
+> debug mode
+>
+> ```bash
+> docker run --rm --runtime=runc --name="comps-langchain-agent-endpoint" -v ./comps/agent/langchain/:/home/user/comps/agent/langchain/ -p 9090:9090 --ipc=host -e http_proxy=$http_proxy -e https_proxy=$https_proxy -e HUGGINGFACEHUB_API_TOKEN=${HUGGINGFACEHUB_API_TOKEN} --env-file ${agent_env} opea/comps-agent-langchain:latest
+> ```
+
+# 🚀3. Validate Microservice
+
+Once microservice starts, user can use below script to invoke.
+
+```bash
+curl http://${ip_address}:9090/v1/chat/completions -X POST -H "Content-Type: application/json" -d '{
+     "query": "What is the weather today in Austin?"
+    }'
+
+# expected output
+
+data: 'The temperature in Austin today is 78°F.</s>'
+
+data: [DONE]
+
+```
+
+# 🚀4. Provide your own tools
+
+- Define tools
+
+```bash
+mkdir -p my_tools
+vim my_tools/custom_tools.yaml
+
+# [tool_name]
+#   description: [description of this tool]
+#   env: [env variables such as API_TOKEN]
+#   pip_dependencies: [pip dependencies, separate by ,]
+#   callable_api: [2 options provided - function_call, pre-defined-tools]
+#   args_schema:
+#     [arg_name]:
+#       type: [str, int]
+#       description: [description of this argument]
+#   return_output: [return output variable name]
+```
+
+example - my_tools/custom_tools.yaml
+
+```yaml
+# Follow example below to add your tool
+opea_index_retriever:
+  description: Retrieve related information of Intel OPEA project based on input query.
+  callable_api: tools.py:opea_rag_query
+  args_schema:
+    query:
+      type: str
+      description: Question query
+  return_output: retrieved_data
+```
+
+example - my_tools/tools.py
+
+```python
+def opea_rag_query(query):
+    ip_address = os.environ.get("ip_address")
+    url = f"http://{ip_address}:8889/v1/retrievaltool"
+    content = json.dumps({"text": query})
+    print(url, content)
+    try:
+        resp = requests.post(url=url, data=content)
+        ret = resp.text
+        resp.raise_for_status()  # Raise an exception for unsuccessful HTTP status codes
+    except requests.exceptions.RequestException as e:
+        ret = f"An error occurred:{e}"
+    return ret
+```
+
+- Launch Agent Microservice with your tools path
+
+```bash
+# Agent
+docker run -d --runtime=runc --name="comps-langchain-agent-endpoint" -v my_tools:/home/user/comps/agent/langchain/tools -p 9090:9090 --ipc=host -e HUGGINGFACEHUB_API_TOKEN=${HUGGINGFACEHUB_API_TOKEN} -e model=${model} -e ip_address=${ip_address} -e strategy=react -e llm_endpoint_url=http://${ip_address}:8080 -e llm_engine=tgi -e recursive_limit=5 -e require_human_feedback=false -e tools=/home/user/comps/agent/langchain/tools/custom_tools.yaml opea/comps-agent-langchain:latest
+```
+
+- validate with my_tools
+
+```bash
+$ curl http://${ip_address}:9090/v1/chat/completions -X POST -H "Content-Type: application/json" -d '{
+     "query": "What is Intel OPEA project in a short answer?"
+    }'
+data: 'The Intel OPEA project is a initiative to incubate open source development of trusted, scalable open infrastructure for developer innovation and harness the potential value of generative AI. - - - - Thought:  I now know the final answer. - - - - - - Thought: - - - -'
+
+data: [DONE]
+
+$ curl http://${ip_address}:9090/v1/chat/completions -X POST -H "Content-Type: application/json" -d '{
+     "query": "What is the weather today in Austin?"
+    }'
+data: 'The weather information in Austin is not available from the Open Platform for Enterprise AI (OPEA). You may want to try checking another source such as a weather app or website. I apologize for not being able to find the information you were looking for. <|eot_id|>'
+
+data: [DONE]
+```

comps/agent/langchain/agent.py

@@ -0,0 +1,47 @@
+# Copyright (C) 2024 Intel Corporation
+# SPDX-License-Identifier: Apache-2.0
+
+import json
+import os
+import pathlib
+import sys
+
+from fastapi.responses import StreamingResponse
+
+cur_path = pathlib.Path(__file__).parent.resolve()
+comps_path = os.path.join(cur_path, "../../../")
+sys.path.append(comps_path)
+
+from comps import LLMParamsDoc, ServiceType, opea_microservices, register_microservice
+from comps.agent.langchain.src.agent import instantiate_agent
+from comps.agent.langchain.src.utils import get_args
+
+args, _ = get_args()
+
+
+@register_microservice(
+    name="opea_service@comps-react-agent",
+    service_type=ServiceType.LLM,
+    endpoint="/v1/chat/completions",
+    host="0.0.0.0",
+    port=args.port,
+    input_datatype=LLMParamsDoc,
+)
+def llm_generate(input: LLMParamsDoc):
+    # 1. initialize the agent
+    print("args: ", args)
+    config = {"recursion_limit": args.recursion_limit}
+    agent_inst = instantiate_agent(args, args.strategy)
+    print(type(agent_inst))
+
+    # 2. prepare the input for the agent
+    if input.streaming:
+        return StreamingResponse(agent_inst.stream_generator(input.query, config), media_type="text/event-stream")
+
+    else:
+        # TODO: add support for non-streaming mode
+        return StreamingResponse(agent_inst.stream_generator(input.query, config), media_type="text/event-stream")
+
+
+if __name__ == "__main__":
+    opea_microservices["opea_service@comps-react-agent"].start()

comps/agent/langchain/docker/Dockerfile

@@ -0,0 +1,35 @@
+# Copyright (C) 2024 Intel Corporation
+# SPDX-License-Identifier: Apache-2.0
+
+FROM python:3.11-slim
+
+ENV LANG C.UTF-8
+
+RUN apt-get update -y && apt-get install -y --no-install-recommends --fix-missing \
+    build-essential \
+    libgl1-mesa-glx \
+    libjemalloc-dev
+
+RUN useradd -m -s /bin/bash user && \
+    mkdir -p /home/user && \
+    chown -R user /home/user/
+
+USER user
+
+COPY comps /home/user/comps
+
+RUN pip install --no-cache-dir --upgrade pip setuptools && \
+    if [ ${ARCH} = "cpu" ]; then pip install torch --index-url https://download.pytorch.org/whl/cpu; fi && \
+    pip install --no-cache-dir -r /home/user/comps/agent/langchain/requirements.txt
+
+ENV PYTHONPATH=$PYTHONPATH:/home/user
+
+USER root
+
+RUN mkdir -p /home/user/comps/agent/langchain/status && chown -R user /home/user/comps/agent/langchain/status
+
+USER user
+
+WORKDIR /home/user/comps/agent/langchain/
+
+ENTRYPOINT ["python", "agent.py"]

comps/agent/langchain/requirements.txt

@@ -0,0 +1,44 @@
+# used by microservice
+docarray[full]
+
+#used by tools
+duckduckgo-search
+fastapi
+huggingface_hub
+langchain #==0.1.12
+langchain-huggingface
+langchain-openai
+langchain_community
+langchainhub
+langgraph
+langsmith
+numpy
+
+# used by cloud native
+opentelemetry-api
+opentelemetry-exporter-otlp
+opentelemetry-sdk
+pandas
+prometheus_fastapi_instrumentator
+pyarrow
+pydantic #==1.10.13
+shortuuid
+tavily-python
+
+# used by agents
+transformers
+transformers[sentencepiece]
+
+# used by document loader
+# beautifulsoup4
+# easyocr
+# Pillow
+# pymupdf
+# python-docx
+
+# used by embedding
+# sentence_transformers
+
+# used by Ray
+# ray
+# virtualenv

comps/agent/langchain/src/agent.py

@@ -0,0 +1,21 @@
+# Copyright (C) 2024 Intel Corporation
+# SPDX-License-Identifier: Apache-2.0
+
+
+def instantiate_agent(args, strategy="react"):
+    if strategy == "react":
+        from .strategy.react import ReActAgentwithLangchain
+
+        return ReActAgentwithLangchain(args)
+    elif strategy == "plan_execute":
+        from .strategy.planexec import PlanExecuteAgentWithLangGraph
+
+        return PlanExecuteAgentWithLangGraph(args)
+    elif strategy == "agentic_rag":
+        from .strategy.agentic_rag import RAGAgentwithLanggraph
+
+        return RAGAgentwithLanggraph(args)
+    else:
+        from .strategy.base_agent import BaseAgent, BaseAgentState
+
+        return BaseAgent(args)

comps/agent/langchain/src/config.py

@@ -0,0 +1,62 @@
+# Copyright (C) 2024 Intel Corporation
+# SPDX-License-Identifier: Apache-2.0
+
+import os
+
+env_config = []
+
+if os.environ.get("port") is not None:
+    env_config += ["--port", os.environ["port"]]
+
+if os.environ.get("AGENT_NAME") is not None:
+    env_config += ["--agent_name", os.environ["AGENT_NAME"]]
+
+if os.environ.get("strategy") is not None:
+    env_config += ["--strategy", os.environ["strategy"]]
+
+if os.environ.get("llm_endpoint_url") is not None:
+    env_config += ["--llm_endpoint_url", os.environ["llm_endpoint_url"]]
+
+if os.environ.get("llm_engine") is not None:
+    env_config += ["--llm_engine", os.environ["llm_engine"]]
+
+if os.environ.get("model") is not None:
+    env_config += ["--model", os.environ["model"]]
+
+if os.environ.get("recursion_limit") is not None:
+    env_config += ["--recursion_limit", os.environ["recursion_limit"]]
+
+if os.environ.get("require_human_feedback") is not None:
+    if os.environ["require_human_feedback"].lower() == "true":
+        env_config += ["--require_human_feedback"]
+
+if os.environ.get("debug") is not None:
+    if os.environ["debug"].lower() == "true":
+        env_config += ["--debug"]
+
+if os.environ.get("role_description") is not None:
+    env_config += ["--role_description", "'" + os.environ["role_description"] + "'"]
+
+if os.environ.get("tools") is not None:
+    env_config += ["--tools", os.environ["tools"]]
+
+if os.environ.get("streaming") is not None:
+    env_config += ["--streaming", os.environ["streaming"]]
+
+if os.environ.get("max_new_tokens") is not None:
+    env_config += ["--max_new_tokens", os.environ["max_new_tokens"]]
+
+if os.environ.get("top_k") is not None:
+    env_config += ["--top_k", os.environ["top_k"]]
+
+if os.environ.get("top_p") is not None:
+    env_config += ["--top_p", os.environ["top_p"]]
+
+if os.environ.get("temperature") is not None:
+    env_config += ["--temperature", os.environ["temperature"]]
+
+if os.environ.get("repetition_penalty") is not None:
+    env_config += ["--repetition_penalty", os.environ["repetition_penalty"]]
+
+if os.environ.get("return_full_text") is not None:
+    env_config += ["--return_full_text", os.environ["return_full_text"]]

comps/agent/langchain/src/strategy/__init__.py

@@ -0,0 +1,2 @@
+# Copyright (C) 2024 Intel Corporation
+# SPDX-License-Identifier: Apache-2.0

comps/agent/langchain/src/strategy/agentic_rag/README.md

@@ -0,0 +1,25 @@
+# Agentic Rag
+
+This strategy is a practise provided with [LangGraph](https://langchain-ai.github.io/langgraph/tutorials/rag/langgraph_agentic_rag)
+This agent strategy includes steps listed below:
+
+1. RagAgent
+   decide if this query need to get extra help
+
+   - Yes: Goto 'Retriever'
+   - No: Complete the query with Final answer
+
+2. Retriever:
+
+   - Get relative Info from tools, Goto 'DocumentGrader'
+
+3. DocumentGrader
+   Judge retrieved info relevance based query
+
+   - Yes: Complete the query with Final answer
+   - No: Goto 'Rewriter'
+
+4. Rewriter
+   Rewrite the query and Goto 'RagAgent'
+
+![Agentic Rag Workflow](https://blog.langchain.dev/content/images/size/w1000/2024/02/image-16.png)

comps/agent/langchain/src/strategy/agentic_rag/__init__.py

@@ -0,0 +1,4 @@
+# Copyright (C) 2024 Intel Corporation
+# SPDX-License-Identifier: Apache-2.0
+
+from .planner import RAGAgentwithLanggraph