feat: fetch a video from oEmbed

Author: eggplants

scrapbox/client.py

@@ -5,7 +5,7 @@
 
 import httpx
 
-from .models import GyazoOEmbedResponse, PageDetail, PageListResponse
+from .models import GyazoOEmbedResponse, GyazoOEmbedResponsePhoto, PageDetail, PageListResponse
 
 if TYPE_CHECKING:
     from types import TracebackType
@@ -139,15 +139,26 @@ def get_file(self, file_id: str) -> bytes:
 
         parsed_url = urlparse(url)
         if parsed_url.hostname and "gyazo.com" in parsed_url.hostname:
-            oembed_url = f"{self.BASE_URL}/oembed-proxy/gyazo"
-            response = self.client.get(oembed_url, params={"url": url})
-            response.raise_for_status()
-            json = response.json()
-            if oembed_type := json.get("type") != "photo":
-                msg = f"Unsupported Gyazo oEmbed type: {oembed_type}"
-                raise ValueError(msg)
-            oembed_data = GyazoOEmbedResponse.model_validate(json)
-            url = oembed_data.root.url
+            # If URL already has a file extension (e.g., .mp4, .jpg), directly convert to i.gyazo.com
+            path = parsed_url.path.strip("/")
+            if "." in path.split("/")[-1]:  # Check if last path segment has extension
+                url = f"https://i.gyazo.com/{path}"
+            else:
+                # Use oEmbed API to get the actual file URL
+                oembed_url = f"{self.BASE_URL}/oembed-proxy/gyazo"
+                response = self.client.get(oembed_url, params={"url": url})
+                response.raise_for_status()
+                json = response.json()
+                if (oembed_type := json.get("type")) not in ("photo", "video"):
+                    msg = f"Unsupported Gyazo oEmbed type: {oembed_type}"
+                    raise ValueError(msg)
+                oembed_data = GyazoOEmbedResponse.model_validate(json)
+                if isinstance(oembed_data.root, GyazoOEmbedResponsePhoto):
+                    url = oembed_data.root.url
+                else:  # video
+                    # Extract Gyazo ID from the original URL and construct direct video URL
+                    gyazo_id = parsed_url.path.strip("/")
+                    url = f"https://i.gyazo.com/{gyazo_id}.mp4"
         response = self.client.get(url)
         response.raise_for_status()
 

scrapbox/models.py

@@ -107,7 +107,28 @@ class GyazoOEmbedResponsePhoto(BaseModel):
     title: str
 
 
-class GyazoOEmbedResponse(RootModel[GyazoOEmbedResponsePhoto]):
+class GyazoOEmbedResponseVideo(BaseModel):
+    """Video information in the Gyazo oEmbed response."""
+
+    model_config = ConfigDict(alias_generator=to_camel, from_attributes=True, populate_by_name=True)
+
+    type: Literal["video"]
+    version: Literal["1.0"]
+    provider_name: str
+    provider_url: str
+    html: str
+    thumbnail_url: str
+    thumbnail_width: int
+    thumbnail_height: int
+    has_audio_track: bool
+    video_length_ms: int
+    width: int
+    height: int
+    scale: float
+    title: str
+
+
+class GyazoOEmbedResponse(RootModel[GyazoOEmbedResponsePhoto | GyazoOEmbedResponseVideo]):
     """Response from the Gyazo oEmbed API.
 
     See: https://gyazo.com/api/docs/image#oembed