From 6ce50b72ad4dfa75a68e382fef1f24219ba64033 Mon Sep 17 00:00:00 2001
From: Hossein <fazlihossein6@gmail.com>
Date: Thu, 7 Aug 2025 03:10:33 +0330
Subject: [PATCH] dodododdo

---
 schemas.py |  7 +------
 scraper.py | 22 ++++++++--------------
 2 files changed, 9 insertions(+), 20 deletions(-)

diff --git a/schemas.py b/schemas.py
index cda4f93..779e75c 100644
--- a/schemas.py
+++ b/schemas.py
@@ -10,15 +10,10 @@ class ScrapeRequest(BaseModel):
     max_count: int = Field(default=5, ge=1, le=50)
 
 
-class MediaItem(BaseModel):
-    media_type: str  # "image" or "video"
-    url: str         # ✅ only S3/MinIO URL (renamed from local_media)
-
-
 class ScrapedPost(BaseModel):
     ig_post_id: str
     media_type: str  # "image", "video", or "carousel"
     caption: Optional[str]
     thumbnail_url: Optional[str]
-    local_media: List[MediaItem]  # ✅ now only contains {"media_type", "url"}
+    local_media: List[str]  # ✅ list of S3 URLs only
     created_at: datetime
diff --git a/scraper.py b/scraper.py
index be1cc45..3254f07 100644
--- a/scraper.py
+++ b/scraper.py
@@ -40,10 +40,10 @@ def scrape_and_store(db: Session, req: ScrapeRequest) -> List[ScrapedPost]:
         if db.query(InstagramPost).filter_by(ig_post_id=post_id).first():
             continue
 
-        media_items = []
+        media_urls: List[str] = []
 
         try:
-            # Handle carousel
+            # Carousel
             if post.media_type == 8 and hasattr(post, "resources"):
                 for res in post.resources:
                     media_url = res.video_url if res.media_type == 2 else res.thumbnail_url
@@ -52,10 +52,7 @@ def scrape_and_store(db: Session, req: ScrapeRequest) -> List[ScrapedPost]:
                     try:
                         local_url = upload_media_from_url(media_url, str(req.seller_id), f"{post_id}/")
                         if local_url:
-                            media_items.append({
-                                "media_type": "video" if res.media_type == 2 else "image",
-                                "url": local_url
-                            })
+                            media_urls.append(local_url)
                     except Exception as e:
                         print(f"⚠️ Error uploading media: {e}")
                         continue
@@ -66,10 +63,7 @@ def scrape_and_store(db: Session, req: ScrapeRequest) -> List[ScrapedPost]:
                 try:
                     local_url = upload_media_from_url(media_url, str(req.seller_id), f"{post_id}/")
                     if local_url:
-                        media_items.append({
-                            "media_type": "video" if post.media_type == 2 else "image",
-                            "url": local_url
-                        })
+                        media_urls.append(local_url)
                 except Exception as e:
                     print(f"⚠️ Error uploading media: {e}")
                     continue
@@ -81,10 +75,10 @@ def scrape_and_store(db: Session, req: ScrapeRequest) -> List[ScrapedPost]:
             print(f"⚠️ Unexpected error while handling media: {e}")
             continue
 
-        if not media_items:
+        if not media_urls:
             continue
 
-        # ✅ Upload thumbnail safely
+        # ✅ Upload and override thumbnail
         final_thumbnail = post.thumbnail_url or ""
         if post.thumbnail_url:
             try:
@@ -101,7 +95,7 @@ def scrape_and_store(db: Session, req: ScrapeRequest) -> List[ScrapedPost]:
             media_type="carousel" if post.media_type == 8 else ("video" if post.media_type == 2 else "image"),
             caption=post.caption_text or "",
             thumbnail_url=final_thumbnail,
-            local_media=media_items,
+            local_media=media_urls,  # ✅ Now a list of strings (S3 URLs)
             created_at=post.taken_at or datetime.utcnow(),
         )
         db.add(post_entry)
@@ -113,7 +107,7 @@ def scrape_and_store(db: Session, req: ScrapeRequest) -> List[ScrapedPost]:
             media_type=post_entry.media_type,
             caption=post_entry.caption,
             thumbnail_url=final_thumbnail,
-            local_media=media_items,
+            local_media=media_urls,  # ✅ Again, list of strings
             created_at=post_entry.created_at
         ))