[feat] CI/CD 코드 완성

.github/workflows/cicd.yml

@@ -2,25 +2,33 @@ name: CI/CD
 
 on:
   push:
-    branches: [ dev ]
-  pull_request:
     branches: [ main, dev ]
+#  pull_request:
+#    branches: [ main, dev ]
 
 env:
   DOCKER_IMAGE: ghcr.io/${{ github.actor }}/ness-fastapi
   VERSION: ${{ github.sha }}
   NAME: ness-fastapi
-  # Docker image 를 ghcr.io 에 올릴 때 우리의github이름/이미지이름 으로 저장한다. 이미지이름을 정해주면 된다.
-  # Docker image 의 이름을 newproject 이라고 해놓은 것. 이름 뭐할지 정하면 된다.
 
 jobs:
   build:
     name: Build
     runs-on: ubuntu-latest
+    env:
+      OPENAI_API_KEY: ${{secrets.OPENAI_API_KEY}}
+      CHROMA_DB_IP_ADDRESS: ${{secrets.CHROMA_DB_IP_ADDRESS}}
+
     steps:
       - name: Checkout
         uses: actions/checkout@v2
 
+      - name: Create .env file
+        run: |
+          echo "CHROMA_DB_IP_ADDRESS=${{ secrets.CHROMA_DB_IP_ADDRESS }}" >> .env
+          echo "OPENAI_API_KEY=${{ secrets.OPENAI_API_KEY }}" >> .env
+        shell: bash
+
       - name: Setup docker buildx
         id: buildx
         uses: docker/setup-buildx-action@v1
@@ -47,28 +55,6 @@ jobs:
           push: true
           tags: ${{ env.DOCKER_IMAGE }}:latest
 
-#  deploy:
-#    needs: build
-#    name: Deploy
-#    runs-on: [ self-hosted, label-NESS ]
-#    # label-newproject 라는 이름으로 AWS EC2 가 Runner 를 작동시킬 때 사용했던 그 label
-#    steps:
-#      - name: Login to ghcr
-#        uses: docker/login-action@v1
-#        with:
-#          registry: ghcr.io
-#          username: ${{ github.actor }}
-#          password: ${{ secrets.GHCR_TOKEN }}
-#
-#      - name: Docker run
-#        run: |
-#          docker stop ${{ env.NAME }} && docker rm ${{ env.NAME }} && docker rmi ${{ env.DOCKER_IMAGE }}:latest
-#          docker run -d -p 3000:3000 --name ness-fastapi --restart always ${{ env.DOCKER_IMAGE }}:latest
-##          docker pull ${{ env.DOCKER_IMAGE }}:latest
-##          docker stop ${{ env.NAME }} || true
-##          docker rm ${{ env.NAME }} || true
-##          docker rmi ${{ env.DOCKER_IMAGE }}:latest || true
-##          docker run -d -p 3000:3000 --name ${{ env.NAME }} --restart always ${{ env.DOCKER_IMAGE }}:latest
 # EC2 볼륨 16기가로 늘려서 테스트 -> 파티션 용량 조절 필요
   deploy:
     needs: build
@@ -82,12 +68,20 @@ jobs:
           username: ${{ github.actor }}
           password: ${{ secrets.GHCR_TOKEN }}
 
-      - name: Remove old Docker image
-        run: docker rmi -f $(docker images -aq) || true #docker rmi -f ${{ env.DOCKER_IMAGE }}:latest || true
+      - name: Create .env file
+        run: |
+          touch .env
+          echo "$DEV_ENV" > .env
+        env:
+          DEV_ENV: ${{ secrets.DEV_ENV }}
 
       - name: Docker run
         run: |
-          docker pull ${{ env.DOCKER_IMAGE }}:latest
           docker stop ${{ env.NAME }} || true
           docker rm ${{ env.NAME }} || true
-          docker run -d -p 3000:3000 --name ${{ env.NAME }} --restart always ${{ env.DOCKER_IMAGE }}:latest
+          docker rmi -f $(docker images -aq) || true
+          docker pull ${{ env.DOCKER_IMAGE }}:latest
+          docker run -d -p 80:80 --name ${{ env.NAME }} --restart always \
+          -e CHROMA_DB_IP_ADDRESS=${{ secrets.CHROMA_DB_IP_ADDRESS }} \
+          -e OPENAI_API_KEY=${{ secrets.OPENAI_API_KEY }} \
+          ${{ env.DOCKER_IMAGE }}:latest

app/database/chroma_db.py

@@ -3,6 +3,9 @@
 from chromadb.utils import embedding_functions
 from chromadb.config import Settings
 
+# FAST API
+from fastapi import Depends
+
 # ETC
 import os
 import datetime
@@ -11,12 +14,6 @@
 
 load_dotenv()
 CHROMA_DB_IP_ADDRESS = os.getenv("CHROMA_DB_IP_ADDRESS")
-# # 로컬에 ChromaDB가 저장된 경우
-# DIR = os.path.dirname(os.path.abspath(__file__))
-# DB_PATH = os.path.join(DIR, 'data')
-# # 로컬 디스크에서 ChromaDB 실행: allow_reset으로 데이터베이스 초기화, anonymized_telemetry를 False로 설정하여 텔레메트리 수집 비활성화
-# chroma_client = chromadb.PersistentClient(path=DB_PATH,
-#                                           settings=Settings(allow_reset=True, anonymized_telemetry=False))
 
 # description: 원격 EC2 인스턴스에서 ChromaDB에 연결
 chroma_client = chromadb.HttpClient(host=CHROMA_DB_IP_ADDRESS, port=8000)
@@ -53,6 +50,9 @@ async def add_db_data(schedule_data: AddScheduleDTO):
     schedules.add(
         documents=[schedule_data.data],
         ids=[str(schedule_data.schedule_id)],
-        metadatas=[{"datetime": schedule_data.schedule_datetime, "member": schedule_data.member_id, "category": schedule_data.category}]
+        metadatas=[{"datetime_start": schedule_data.schedule_datetime_start, "datetime_end": schedule_data.schedule_datetime_end, "member": schedule_data.member_id, "category": schedule_data.category, "location": schedule_data.location, "person": schedule_data.person}]
     )
-    return True
+    return True
+
+def get_chroma_client():
+    return chroma_client

app/dto/db_dto.py

@@ -3,8 +3,11 @@
 
 class AddScheduleDTO(BaseModel):
     data: str
-    schedule_datetime: str  # datetime.datetime 타입을 명시적으로 사용합니다.
+    schedule_datetime_start: str  # datetime.datetime 타입을 명시적으로 사용합니다.
+    schedule_datetime_end: str
     schedule_id: int
     member_id: int
     category: str
+    location: str
+    person: str
 

app/dto/openai_dto.py

@@ -3,3 +3,6 @@
 
 class PromptRequest(BaseModel):
     prompt: str
+
+class ChatResponse(BaseModel):
+    ness: str

app/main.py

@@ -3,21 +3,7 @@
 from dotenv import load_dotenv
 
 # BACKEND
-from fastapi import FastAPI
-from fastapi import FastAPI, HTTPException
-import asyncio
-
-# VECTOR DB Module
-import app.database.chroma_db as vectordb
-
-# AI
-from langchain.chat_models import ChatOpenAI
-from langchain.prompts import PromptTemplate
-
-# DTO
-from app.dto import openai_dto
-from app.prompt import openai_prompt
-from app.dto.db_dto import AddScheduleDTO
+from fastapi import FastAPI, Depends
 
 # ETC
 import os

app/routers/chat.py

@@ -3,11 +3,11 @@
 import os
 
 from dotenv import load_dotenv
-from fastapi import APIRouter, HTTPException
+from fastapi import APIRouter, HTTPException, status
 from langchain_community.chat_models import ChatOpenAI
 from langchain_core.prompts import PromptTemplate
 
-from app.dto import openai_dto
+from app.dto.openai_dto import PromptRequest, ChatResponse
 from app.prompt import openai_prompt
 
 import app.database.chroma_db as vectordb
@@ -26,8 +26,8 @@
 config = configparser.ConfigParser()
 config.read(CONFIG_FILE_PATH)
 
-@router.post("/case")
-async def get_langchain_case(data: openai_dto.PromptRequest):
+@router.post("/case", status_code=status.HTTP_200_OK, response_model=ChatResponse)
+async def get_langchain_case(data: PromptRequest) -> ChatResponse:
     # description: use langchain
 
     config_normal = config['NESS_NORMAL']
@@ -64,7 +64,7 @@ async def get_langchain_case(data: openai_dto.PromptRequest):
 
 # case 1 : normal
 #@router.post("/case/normal") # 테스트용 엔드포인트
-async def get_langchain_normal(data: openai_dto.PromptRequest): # case 1 : normal
+async def get_langchain_normal(data: PromptRequest) -> ChatResponse: # case 1 : normal
     print("running case 1")
     # description: use langchain
     chat_model = ChatOpenAI(temperature=0,  # 창의성 (0.0 ~ 2.0)
@@ -82,11 +82,11 @@ async def get_langchain_normal(data: openai_dto.PromptRequest): # case 1 : norma
     prompt = PromptTemplate.from_template(my_template)
     response = chat_model.predict(prompt.format(output_language="Korean", question=question))
     print(response)
-    return response
+    return ChatResponse(ness=response)
 
 # case 2 : 일정 생성
 #@router.post("/case/make_schedule") # 테스트용 엔드포인트
-async def get_langchain_schedule(data: openai_dto.PromptRequest):
+async def get_langchain_schedule(data: PromptRequest) -> ChatResponse:
     print("running case 2")
     # description: use langchain
     chat_model = ChatOpenAI(temperature=0,  # 창의성 (0.0 ~ 2.0)
@@ -100,11 +100,11 @@ async def get_langchain_schedule(data: openai_dto.PromptRequest):
     prompt = PromptTemplate.from_template(case2_template)
     response = chat_model.predict(prompt.format(output_language="Korean", question=question))
     print(response)
-    return response
+    return ChatResponse(ness=response)
 
 # case 3 : rag
 #@router.post("/case/rag") # 테스트용 엔드포인트
-async def get_langchain_rag(data: openai_dto.PromptRequest):
+async def get_langchain_rag(data: PromptRequest) -> ChatResponse:
     print("running case 3")
     # description: use langchain
     chat_model = ChatOpenAI(temperature=0,  # 창의성 (0.0 ~ 2.0)
@@ -121,9 +121,6 @@ async def get_langchain_rag(data: openai_dto.PromptRequest):
     case3_template = openai_prompt.Template.case3_template
 
     prompt = PromptTemplate.from_template(case3_template)
-    # 여기서는 chat_model.predict가 비동기 함수인지, 동기 함수인지에 따라 처리가 달라질 수 있습니다.
-    # 만약 비동기 함수라면 await를 사용해야 합니다. 아래 코드는 동기 함수를 가정하고 작성되었습니다.
-    # 비동기 함수라면, 예: response = await chat_model.predict(...) 형태로 수정해야 합니다.
     response = chat_model.predict(prompt.format(output_language="Korean", question=question, schedule=schedule))
     print(response)
-    return response
+    return ChatResponse(ness=response)

app/routers/chromadb.py

@@ -2,9 +2,10 @@
 import os
 
 from dotenv import load_dotenv
-from fastapi import APIRouter, HTTPException
+from fastapi import APIRouter, HTTPException, Depends, status
 
 from app.dto.db_dto import AddScheduleDTO
+from app.database.chroma_db import add_db_data, get_chroma_client
 
 router = APIRouter(
     prefix="/chromadb",
@@ -21,12 +22,11 @@
 config.read(CONFIG_FILE_PATH)
 
 
-@router.post("/add_schedule")
-async def add_schedule_endpoint(schedule_data: AddScheduleDTO, vectordb=None):
+@router.post("/add_schedule", status_code=status.HTTP_201_CREATED)
+async def add_schedule_endpoint(schedule_data: AddScheduleDTO, chroma_client=Depends(get_chroma_client)):
     try:
-        # vectordb.add_db_data 함수를 비동기적으로 호출합니다.
-        await vectordb.add_db_data(schedule_data)
+        # 직접 `add_db_data` 함수를 비동기적으로 호출합니다.
+        await add_db_data(schedule_data)
         return {"message": "Schedule added successfully"}
     except Exception as e:
-        # 에러 처리: 에러가 발생하면 HTTP 500 응답을 반환합니다.
-        raise HTTPException(status_code=500, detail=str(e))
+        raise HTTPException(status_code=500, detail=str(e))

app/routers/recommendation.py

@@ -23,7 +23,7 @@
 config.read(CONFIG_FILE_PATH)
 
 @router.get("/main")
-def get_recommendation():
+async def get_recommendation():
 
     # 모델
     chat_model = ChatOpenAI(temperature=0,  # 창의성 (0.0 ~ 2.0)

requirements.txt

@@ -1,10 +1,9 @@
 chromadb==0.4.22
 fastapi==0.109.0
 uvicorn==0.26.0
-langchain==0.1.1
 langchain-community==0.0.13
 openai==1.8.0
 python-dotenv==1.0.0
 starlette==0.35.1
 pydantic==2.5.3
-sentence-transformers==2.5.1
+sentence-transformers==2.5.1