Merge pull request #44 from cuappdev/claire/youtubeVideoDuration

Add duration field to YoutubeVideo model and related mutations

Author: claiireyu

src/models/youtube_video.py

@@ -12,10 +12,11 @@ class YoutubeVideo:
         - `thumbnail`       The thumbnail of the video, as a URL string pointing to a `.jpg` file.
         - `url`             The URL of the video.
         - `published_at`    The date and time the video was published.
+        - `duration`        The duration of the video.
     """
 
     def __init__(
-        self, title, description, thumbnail, b64_thumbnail, url, published_at, id=None
+        self, title, description, thumbnail, b64_thumbnail, url, published_at, duration=None, id=None
     ):
         self.id = id if id else str(ObjectId())
         self.title = title
@@ -24,6 +25,7 @@ def __init__(
         self.b64_thumbnail = b64_thumbnail
         self.url = url
         self.published_at = published_at
+        self.duration = duration
 
     def to_dict(self):
         """
@@ -37,6 +39,7 @@ def to_dict(self):
             "b64_thumbnail": self.b64_thumbnail,
             "url": self.url,
             "published_at": self.published_at,
+            "duration": self.duration,
         }
 
     @staticmethod
@@ -52,4 +55,5 @@ def from_dict(data):
             b64_thumbnail=data.get("b64_thumbnail"),
             url=data.get("url"),
             published_at=data.get("published_at"),
+            duration=data.get("duration"),
         )

src/mutations/create_youtube_video.py

@@ -11,10 +11,11 @@ class Arguments:
         b64_thumbnail = String(required=True)
         url = String(required=True)
         published_at = String(required=True)
+        duration = String(required=True)
 
     youtube_video = Field(lambda: YoutubeVideoType)
 
-    def mutate(self, info, id, title, description, thumbnail, url, published_at):
+    def mutate(self, info, id, title, description, thumbnail, b64_thumbnail, url, published_at, duration):
         video_data = {
             "id": id,
             "title": title,
@@ -23,6 +24,7 @@ def mutate(self, info, id, title, description, thumbnail, url, published_at):
             "b64_thumbnail": b64_thumbnail,
             "url": url,
             "published_at": published_at,
+            "duration": duration,
         }
         new_video = YoutubeVideoService.create_video(video_data)
         return CreateYoutubeVideo(youtube_video=new_video)

src/scrapers/youtube_stats.py

@@ -6,6 +6,7 @@
 import base64
 import os
 import html
+from bs4 import BeautifulSoup
 
 load_dotenv()
 YOUTUBE_API_KEY = os.getenv("YOUTUBE_API_KEY")
@@ -25,6 +26,54 @@ def fetch_videos():
         process_video_item(item)
 
 
+def get_video_duration(video_id):
+    """
+    Gets video duration using YouTube API
+    """
+    try:
+        url = f"https://www.googleapis.com/youtube/v3/videos?key={YOUTUBE_API_KEY}&id={video_id}&part=contentDetails"
+        response = requests.get(url)
+        response.raise_for_status()
+        data = response.json()
+        
+        if data.get("items"):
+            duration_iso = data["items"][0]["contentDetails"]["duration"]
+            return convert_iso_duration(duration_iso)
+        return None
+    except Exception as e:
+        print(f"Error getting video duration: {e}")
+        return None
+
+
+def convert_iso_duration(iso_duration):
+    """
+    Converts ISO 8601 duration (PT2M5S) to readable format (2:05)
+    Examples:
+    - PT2M5S -> 2:05
+    - PT1H23M45S -> 1:23:45
+    - PT30S -> 0:30
+    """
+    import re
+    
+    # Remove PT prefix
+    duration = iso_duration.replace('PT', '')
+    
+    # Extract hours, minutes, seconds
+    hours = re.search(r'(\d+)H', duration)
+    minutes = re.search(r'(\d+)M', duration)
+    seconds = re.search(r'(\d+)S', duration)
+    
+    h = int(hours.group(1)) if hours else 0
+    m = int(minutes.group(1)) if minutes else 0
+    s = int(seconds.group(1)) if seconds else 0
+    
+    # Format as MM:SS or HH:MM:SS
+    if h > 0:
+        return f"{h}:{m:02d}:{s:02d}"
+    else:
+        return f"{m}:{s:02d}"
+
+
 def process_video_item(item):
     """
     Extracts the required data from a video item and
@@ -55,14 +104,17 @@ def process_video_item(item):
     published_at = snippet.get("publishedAt")
     video_url = f"https://www.youtube.com/watch?v={video_id}"
 
+    duration = get_video_duration(video_id)
+
     video_data = {
-        "id": video_id,  # use video id for easy retrieval
+        "id": video_id,  
         "title": title,
         "description": description,
         "thumbnail": thumbnail,
         "b64_thumbnail": encoded_thumbnail,
         "url": video_url,
         "published_at": published_at,
+        "duration": duration,
     }
     process_video_data(video_data)
 

src/services/youtube_video_service.py

@@ -30,6 +30,7 @@ def create_video(data):
             b64_thumbnail=data.get("b64_thumbnail"),
             url=data.get("url"),
             published_at=data.get("published_at"),
+            duration=data.get("duration"),
         )
         YoutubeVideoRepository.insert(video)
         return video

src/types.py

@@ -155,6 +155,7 @@ class YoutubeVideoType(ObjectType):
         - thumbnail: The URL of the video's thumbnail.
         - url: The URL to the video.
         - published_at: The date and time the video was published.
+        - duration: The duration of the video (optional).
     """
     id = String(required=False)
     title = String(required=True)
@@ -163,6 +164,7 @@ class YoutubeVideoType(ObjectType):
     b64_thumbnail = String(required=True)
     url = String(required=True)
     published_at = String(required=True)
+    duration = String(required=False)
 
     def __init__(self, **kwargs):
         for key, value in kwargs.items():