[feat] report tags complete: added db metadatas, changed db query logic

app/database/chroma_db.py

@@ -48,10 +48,13 @@ async def search_db_query(query):
 # 스프링 백엔드로부터 chroma DB에 저장할 데이터를 받아 DB에 추가한다.
 async def add_db_data(schedule_data: AddScheduleDTO):
     schedule_date = schedule_data.schedule_datetime_start.split("T")[0]
+    year = int(schedule_date.split("-")[0])
+    month = int(schedule_date.split("-")[1])
+    date = int(schedule_date.split("-")[2])
     schedules.add(
         documents=[schedule_data.data],
         ids=[str(schedule_data.schedule_id)],
-        metadatas=[{"date": schedule_date, "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}]
+        metadatas=[{"year": year, "month": month, "date": date, "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
 
@@ -62,14 +65,19 @@ async def db_daily_schedule(user_data: RecommendationMainRequestDTO):
     schedule_datetime_start = user_data.schedule_datetime_start
     schedule_datetime_end = user_data.schedule_datetime_end
     schedule_date = schedule_datetime_start.split("T")[0]
+    year = int(schedule_date.split("-")[0])
+    month = int(schedule_date.split("-")[1])
+    date = int(schedule_date.split("-")[2])
     persona = user_data.user_persona or "hard working"
     results = schedules.query(
         query_texts=[persona],
         n_results=5,
         where={"$and":
                [
                    {"member": {"$eq": int(member)}},
-                   {"date": {"$eq": schedule_date}}
+                   {"year": {"$eq": year}},
+                   {"month": {"$eq": month}},
+                   {"date": {"$eq": date}}
                ]
         }
         # where_document={"$contains":"search_string"}  # optional filter
@@ -82,14 +90,23 @@ async def db_monthly_tag_schedule(user_data: ReportTagsRequestDTO):
     schedule_datetime_start = user_data.schedule_datetime_start
     schedule_datetime_end = user_data.schedule_datetime_end
     schedule_date = schedule_datetime_start.split("T")[0]
+    year = int(schedule_date.split("-")[0])
+    month = int(schedule_date.split("-")[1])
+    date = int(schedule_date.split("-")[2])
     persona = user_data.user_persona or "hard working"
     results = schedules.query(
         query_texts=[persona],
-        n_results=5,
+        n_results=15,
         where={"$and":
                [
                    {"member": {"$eq": int(member)}},
-                   {"date": {"$eq": schedule_date}}
+                   {"year": {"$eq": year}},
+                   {"$or":
+                    [{"$and":
+                          [{"month": {"$eq": month-1}}, {"date": {"$gte": 10}}]},
+                     {"$and":
+                          [{"month": {"$eq": month}}, {"date": {"$lt": 10}}]}
+                    ]}
                ]
         }
         # where_document={"$contains":"search_string"}  # optional filter

app/dto/openai_dto.py

@@ -9,4 +9,7 @@ class ChatResponse(BaseModel):
 
 class ChatCaseResponse(BaseModel):
     ness: str
-    case: int
+    case: int
+
+class TagsResponse(BaseModel):
+    tags: list

app/prompt/openai_config.ini

@@ -11,4 +11,9 @@ MODEL_NAME      = gpt-4
 [NESS_RECOMMENDATION]
 TEMPERATURE     = 0
 MAX_TOKENS      = 2048
-MODEL_NAME      = gpt-3.5-turbo-1106
+MODEL_NAME      = gpt-3.5-turbo-1106
+
+[NESS_TAGS]
+TEMPERATURE     = 0
+MAX_TOKENS      = 2048
+MODEL_NAME      = gpt-4

app/prompt/report_prompt.py

@@ -14,4 +14,22 @@ class Template:
         
         User schedule: {schedule}
         AI Recommendation:
+        """
+
+    report_tags_template = """
+        You are an AI assistant tasked with analyzing a user's schedule over the span of a month. From this detailed schedule, you will distill three keywords that best encapsulate the user's activities, interests, or achievements throughout the month. These keywords should not only reflect the user's endeavors but also convey a sense of accomplishment and enjoyment. Your output should be engaging, showcasing your wit and unique perspective. Here are the rules for your analysis:
+        
+        YOU MUST USE {output_language} TO RESPOND TO THE INPUT.
+        YOU MUST PROVIDE THREE KEYWORDS in your response. Each keyword should be a single word or a concise phrase.
+        The keywords must capture the essence of the user's monthly activities, highlighting aspects that are both rewarding and enjoyable.
+        Your selections should be creative and personalized, aiming to reflect the user's unique experiences over the month.
+        Example:
+        User's monthly schedule: [Attended a programming bootcamp, Completed a marathon, Read three novels, Volunteered at the local food bank, Started a blog about sustainability]
+        AI Recommendation: "공부 매니아, 환경 지킴이, 자기 계발 홀릭"
+        
+        User's monthly schedule: [Took photography classes, Explored three new hiking trails, Organized a neighborhood clean-up, Experimented with vegan recipes]
+        AI Recommendation: "모험가, 미식가, 도파민 중독자"
+        
+        User's monthly schedule: {schedule}
+        AI Recommendation:
         """

app/routers/report.py

@@ -7,7 +7,7 @@
 from langchain_core.prompts import PromptTemplate
 
 from app.dto.db_dto import ReportMemoryEmojiRequestDTO, ReportTagsRequestDTO
-from app.dto.openai_dto import ChatResponse
+from app.dto.openai_dto import ChatResponse, TagsResponse
 from app.prompt import report_prompt
 import app.database.chroma_db as vectordb
 
@@ -54,10 +54,10 @@ async def get_memory_emoji(user_data: ReportMemoryEmojiRequestDTO) -> ChatRespon
         raise HTTPException(status_code=500, detail=str(e))
 
 @router.post("/tags", status_code=status.HTTP_200_OK)
-async def get_tags(user_data: ReportTagsRequestDTO) -> ChatResponse:
+async def get_tags(user_data: ReportTagsRequestDTO) -> TagsResponse:
     try:
         # 모델
-        config_tags = config['NESS_RECOMMENDATION']
+        config_tags = config['NESS_TAGS']
 
         chat_model = ChatOpenAI(temperature=config_tags['TEMPERATURE'],  # 창의성 (0.0 ~ 2.0)
                                 max_tokens=config_tags['MAX_TOKENS'],  # 최대 토큰수
@@ -71,12 +71,15 @@ async def get_tags(user_data: ReportTagsRequestDTO) -> ChatResponse:
         print(schedule)
 
         # 템플릿
-        memory_emoji_template = report_prompt.Template.memory_emoji_template
+        report_tags_template = report_prompt.Template.report_tags_template
 
-        prompt = PromptTemplate.from_template(memory_emoji_template)
+        prompt = PromptTemplate.from_template(report_tags_template)
         result = chat_model.predict(prompt.format(output_language="Korean", schedule=schedule))
         print(result)
-        return ChatResponse(ness=result)
+        tags = result.split("\"")[1].split(",")
+        tags = [tag.strip() for tag in tags]
+        print(tags)
+        return TagsResponse(tags=tags)
 
     except Exception as e:
         raise HTTPException(status_code=500, detail=str(e))