feat(tutor): respond with 204 when wrongly formatted(#85)

* feat(tutor): respond with 204 when wrongly formatted

* fix: prompt

* adds summaries model

Author: sandragjacinto

src/app/api/api_v1/endpoints/tutor.py

@@ -1,6 +1,6 @@
 from typing import Annotated
 
-from fastapi import APIRouter, File, HTTPException, Response, UploadFile
+from fastapi import APIRouter, File, Response, UploadFile
 
 from src.app.api.dependencies import get_settings
 from src.app.models.search import EnhancedSearchQuery
@@ -12,6 +12,7 @@
 from src.app.services.tutor.agents import TEMPLATES
 from src.app.services.tutor.models import (
     ExtractorOutputList,
+    SummariesList,
     SyllabusFeedback,
     SyllabusResponse,
     SyllabusResponseAgent,
@@ -46,7 +47,7 @@
 
 @router.post("/files/content")
 async def extract_files_content(
-    files: Annotated[list[UploadFile], File()],
+    files: Annotated[list[UploadFile], File()], response: Response
 ) -> ExtractorOutputList | None:
     files_content = await get_files_content(files)
     files_content_str = ("__DOCUMENT_SEPARATOR__").join(files_content)
@@ -72,18 +73,19 @@ async def extract_files_content(
 
     except Exception as e:
         logger.error(f"Error in extractor schema: {e}")
-        HTTPException(status_code=400, detail="error in response schema")
+        response.status_code = 204
+        return None
 
 
 @router.post("/search_extracts")
 async def tutor_search_extract(
-    summaries: list[str],
+    summaries: SummariesList,
     response: Response,
 ):
 
     try:
         qp = EnhancedSearchQuery(
-            query=summaries,
+            query=summaries.summaries,
             nb_results=10,
             sdg_filter=None,
             corpora=None,
@@ -154,11 +156,12 @@ async def tutor_search(
         else:
             raise ValueError("Unexpected response format")
 
+        print(jsn)
         themes_extracted = ExtractorOutputList(**jsn)
 
     except Exception as e:
         logger.error(f"Error in chat schema: {e}")
-        # todo: handle error
+        response.status_code = 204
         return TutorSearchResponse(
             extracts=[],
             nb_results=0,

src/app/services/tutor/models.py

@@ -5,6 +5,10 @@
 from qdrant_client.models import ScoredPoint
 
 
+class SummariesList(BaseModel):
+    summaries: list[str]
+
+
 class ExtractorOutput(BaseModel):
     summary: str
     themes: list[str]

src/app/services/tutor/prompts.py

@@ -2,7 +2,7 @@
 role="An assistant to summarize a text and extract the main themes from it",
 backstory="You are specialised in analysing documents, summarizing them and extracting the main themes. You value precision and clarity.",
 goal="Analyse each document, summarize it and extract the main themes, explaining why each theme was identified.",
-expected_output="You must follow the following JSON schema: {extracts: [{'original_document': 'Document', 'summary': 'Summary', 'themes': ['Theme 1', 'Theme 2', ...]}, {'original_document': 'Document', 'summary': 'Sumamry', 'themes': ['Theme 1', 'Theme 2', ...]}, ...]} an entry by document",
+expected_output="You must follow the following JSON schema: {extracts: [{'original_document': 'Document', 'summary': 'Summary', 'themes': ['Theme 1', 'Theme 2', ...]}, {'original_document': 'Document', 'summary': 'Sumamry', 'themes': ['Theme 1', 'Theme 2', ...]}, ...]} an entry by document make sure to use double quotes when formating the JSON",
 """
 
 extractor_user_prompt = """

src/app/tests/api/api_v1/test_tutor.py

@@ -39,4 +39,4 @@ def test_tutor_file(self, *mocks):
             files={"files": ("test.txt", file)},
             headers={"x-API-Key": "test"},
         )
-        assert reponse.status_code == 200
+        assert reponse.status_code == 204