fix: openapi schema

api/main.py

@@ -9,7 +9,7 @@
     query_media,
     query_mediabiasfactcheck,
 )
-from api.tools.youtube import Video, youtube_search, youtube_transcripts
+from api.youtube import Video, VideoTranscript, youtube_search, youtube_transcripts
 from lib.auth import verify_apikey
 
 app = FastAPI()
@@ -59,6 +59,9 @@ async def get_youtube_search(
     get_transcripts: bool = False,
     _: None = Depends(verify_apikey),
 ) -> List[Video]:
+    """
+    Get the details of matching videos by either providing Youtube channels, a query, or both
+    """
     if not (channels or query):
         raise HTTPException(
             status_code=400, detail="No query given when no channels are provided!"
@@ -74,11 +77,14 @@ async def get_youtube_search(
     )
 
 
-@app.get("/youtube-transcripts", response_model=Dict[str, str])
+@app.get("/youtube-transcripts", response_model=List[VideoTranscript])
 async def get_youtube_transcripts(
     ids: str,
     _: None = Depends(verify_apikey),
 ) -> Dict[str, str]:
+    """
+    Extract transcripts from a list of Youtube video ids
+    """
     return youtube_transcripts(ids)
 
 

api/tools/youtube.py → api/youtube.py

@@ -34,6 +34,13 @@ class Transcript(BaseModel):
     duration: int
 
 
+class VideoTranscript(BaseModel):
+    """Video transcript model"""
+
+    id: str
+    text: str
+
+
 class Video(BaseModel):
     """Video model"""
 
@@ -136,9 +143,6 @@ async def youtube_search(
     get_descriptions: bool = False,
     get_transcripts: bool = False,
 ) -> List[Video]:
-    """
-    Get the details of matching videos by either providing Youtube channels, a query, or both
-    """
     if channels:
         channels_arr = channels.lower().split(",")
         media = [
@@ -184,14 +188,14 @@ async def youtube_search(
 @cache(ttl=3600)
 def youtube_transcripts(
     ids: str,
-) -> Dict[str, str]:
+) -> List[VideoTranscript]:
     """
     Extract transcripts from a list of Youtube video ids
     """
-    results: Dict[str, str] = {}
+    results: List[VideoTranscript] = []
     for video_id in ids.split(","):
         transcript = _get_video_transcript(video_id, strip_timestamps=True)
-        results[video_id] = transcript
+        results.append(VideoTranscript(id=video_id, text=transcript))
     return results
 
 

openapi.yaml

@@ -145,6 +145,8 @@ paths:
   /media-videos:
     get:
       summary: Get Youtube Search
+      description: Get the details of matching videos by either providing Youtube
+        channels, a query, or both
       operationId: get_youtube_search_media_videos_get
       security:
       - APIKeyQuery: []
@@ -217,6 +219,7 @@ paths:
   /youtube-transcripts:
     get:
       summary: Get Youtube Transcripts
+      description: Extract transcripts from a list of Youtube video ids
       operationId: get_youtube_transcripts_youtube_transcripts_get
       security:
       - APIKeyQuery: []
@@ -235,9 +238,9 @@ paths:
           content:
             application/json:
               schema:
-                type: object
-                additionalProperties:
-                  type: string
+                type: array
+                items:
+                  $ref: '#/components/schemas/VideoTranscript'
                 title: Response Get Youtube Transcripts Youtube Transcripts Get
         '422':
           description: Validation Error
@@ -411,6 +414,20 @@ components:
       - url_suffix
       title: Video
       description: Video model
+    VideoTranscript:
+      properties:
+        id:
+          type: string
+          title: Id
+        text:
+          type: string
+          title: Text
+      type: object
+      required:
+      - id
+      - text
+      title: VideoTranscript
+      description: Video transcript model
   securitySchemes:
     APIKeyQuery:
       type: apiKey

pages/2-Youtube.py

@@ -3,8 +3,7 @@
 import streamlit as st
 import streamlit.components.v1 as components
 
-from api.main import youtube_search
-from api.tools.youtube import youtube_search
+from api.youtube import youtube_search
 
 with open("index.html", "r") as f:
     html_code = f.read()