Feat: video to image mode

backend/alembic/versions/582bbd507011_add_youtube_url_to_source_assets.py

@@ -0,0 +1,76 @@
+# Copyright 2026 Google LLC
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+"""add_external_url_to_source_assets
+
+Revision ID: 582bbd507011
+Revises: 7ec3aed70c3f
+Create Date: 2026-07-13 16:06:40.894364
+
+"""
+from typing import Sequence, Union
+
+from alembic import op
+import sqlalchemy as sa
+from sqlalchemy.dialects import postgresql
+
+# revision identifiers, used by Alembic.
+revision: str = "582bbd507011"
+down_revision: Union[str, None] = "7ec3aed70c3f"
+branch_labels: Union[str, Sequence[str], None] = None
+depends_on: Union[str, Sequence[str], None] = None
+
+
+def upgrade() -> None:
+    # ### commands auto generated by Alembic - please adjust! ###
+    op.add_column(
+        "source_assets", sa.Column("external_url", sa.String(), nullable=True)
+    )
+    op.alter_column(
+        "source_assets", "gcs_uri", existing_type=sa.VARCHAR(), nullable=True
+    )
+    op.alter_column(
+        "source_assets",
+        "original_filename",
+        existing_type=sa.VARCHAR(),
+        nullable=True,
+    )
+    op.alter_column(
+        "source_assets", "mime_type", existing_type=sa.VARCHAR(), nullable=True
+    )
+    op.alter_column(
+        "source_assets", "file_hash", existing_type=sa.VARCHAR(), nullable=True
+    )
+    # ### end Alembic commands ###
+
+
+def downgrade() -> None:
+    # ### commands auto generated by Alembic - please adjust! ###
+    op.alter_column(
+        "source_assets", "file_hash", existing_type=sa.VARCHAR(), nullable=False
+    )
+    op.alter_column(
+        "source_assets", "mime_type", existing_type=sa.VARCHAR(), nullable=False
+    )
+    op.alter_column(
+        "source_assets",
+        "original_filename",
+        existing_type=sa.VARCHAR(),
+        nullable=False,
+    )
+    op.alter_column(
+        "source_assets", "gcs_uri", existing_type=sa.VARCHAR(), nullable=False
+    )
+    op.drop_column("source_assets", "external_url")
+    # ### end Alembic commands ###

backend/alembic/versions/7a8b9c0d1e2f_update_unified_gallery_view_external_url.py

@@ -0,0 +1,208 @@
+# Copyright 2026 Google LLC
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+"""update_unified_gallery_view_external_url
+
+Revision ID: 7a8b9c0d1e2f
+Revises: 582bbd507011
+Create Date: 2026-07-15 03:40:00.000000
+
+"""
+from typing import Sequence, Union
+
+from alembic import op
+
+# revision identifiers, used by Alembic.
+revision: str = "7a8b9c0d1e2f"
+down_revision: Union[str, None] = "582bbd507011"
+branch_labels: Union[str, Sequence[str], None] = None
+depends_on: Union[str, Sequence[str], None] = None
+
+
+def upgrade() -> None:
+    op.execute("DROP VIEW IF EXISTS unified_gallery_view;")
+    op.execute(
+        """
+    CREATE VIEW unified_gallery_view AS
+    WITH unified_base AS (
+        SELECT
+            mi.id,
+            mi.workspace_id,
+            mi.user_id,
+            mi.created_at,
+            'media_item'::text AS item_type,
+            mi.status,
+            mi.gcs_uris,
+            mi.thumbnail_uris,
+            mi.deleted_at,
+            mi.titles,
+            mi.descriptions,
+            jsonb_build_object(
+                'model', mi.model,
+                'prompt', mi.prompt,
+                'original_prompt', mi.original_prompt,
+                'negative_prompt', mi.negative_prompt,
+                'aspect_ratio', mi.aspect_ratio,
+                'mime_type', mi.mime_type,
+                'style', mi.style,
+                'lighting', mi.lighting,
+                'num_media', mi.num_media,
+                'generation_time', mi.generation_time,
+                'file_name', mi.comment,
+                'source_assets', mi.source_assets,
+                'source_media_items', mi.source_media_items,
+                'is_video', (mi.mime_type LIKE 'video%'),
+                'is_audio', (mi.mime_type LIKE 'audio%'),
+                'tags', (
+                    SELECT jsonb_agg(jsonb_build_object('id', t.id, 'name', t.name, 'color', t.color, 'workspace_id', t.workspace_id))
+                    FROM media_item_tags mit
+                    JOIN tags t ON mit.tag_id = t.id
+                    WHERE mit.media_item_id = mi.id
+                )
+            ) AS metadata
+        FROM media_items mi
+        UNION ALL
+        SELECT
+            sa.id,
+            sa.workspace_id,
+            sa.user_id,
+            sa.created_at,
+            'source_asset'::text AS item_type,
+            'completed'::text AS status,
+            CASE
+                WHEN (sa.gcs_uri IS NOT NULL) THEN ARRAY[sa.gcs_uri]
+                ELSE '{}'::text[]
+            END AS gcs_uris,
+            CASE
+                WHEN (sa.thumbnail_gcs_uri IS NOT NULL) THEN ARRAY[sa.thumbnail_gcs_uri]
+                ELSE '{}'::text[]
+            END AS thumbnail_uris,
+            sa.deleted_at,
+            sa.titles,
+            sa.descriptions,
+            jsonb_build_object(
+                'file_name', sa.original_filename,
+                'original_filename', sa.original_filename,
+                'mime_type', sa.mime_type,
+                'aspect_ratio', sa.aspect_ratio,
+                'asset_type', sa.asset_type,
+                'external_url', sa.external_url,
+                'is_video', (sa.mime_type LIKE 'video%' OR sa.asset_type = 'youtube_video' OR sa.external_url IS NOT NULL),
+                'is_audio', (sa.mime_type LIKE 'audio%'),
+                'tags', (
+                    SELECT jsonb_agg(jsonb_build_object('id', t.id, 'name', t.name, 'color', t.color, 'workspace_id', t.workspace_id))
+                    FROM source_asset_tags sat
+                    JOIN tags t ON sat.tag_id = t.id
+                    WHERE sat.source_asset_id = sa.id
+                )
+            ) AS metadata
+        FROM source_assets sa
+    )
+    SELECT 
+        ub.*,
+        w.name AS workspace_name,
+        u.picture AS user_picture,
+        u.email AS user_email
+    FROM unified_base ub
+    LEFT JOIN workspaces w ON ub.workspace_id = w.id
+    LEFT JOIN users u ON ub.user_id = u.id;
+    """
+    )
+
+
+def downgrade() -> None:
+    op.execute("DROP VIEW IF EXISTS unified_gallery_view;")
+    op.execute(
+        """
+    CREATE VIEW unified_gallery_view AS
+    SELECT
+        mi.id,
+        mi.workspace_id,
+        mi.user_id,
+        mi.created_at,
+        'media_item'::text AS item_type,
+        mi.status,
+        mi.gcs_uris,
+        mi.thumbnail_uris,
+        mi.deleted_at,
+        w.name AS workspace_name,
+        u.picture AS user_picture,
+        u.email AS user_email,
+        mi.titles,
+        mi.descriptions,
+        jsonb_build_object(
+            'model', mi.model,
+            'prompt', mi.prompt,
+            'original_prompt', mi.original_prompt,
+            'negative_prompt', mi.negative_prompt,
+            'aspect_ratio', mi.aspect_ratio,
+            'mime_type', mi.mime_type,
+            'style', mi.style,
+            'lighting', mi.lighting,
+            'num_media', mi.num_media,
+            'generation_time', mi.generation_time,
+            'file_name', mi.comment,
+            'source_assets', mi.source_assets,
+            'source_media_items', mi.source_media_items,
+            'is_video', (mi.mime_type like 'video%'),
+            'is_audio', (mi.mime_type like 'audio%'),
+            'tags', (
+                SELECT jsonb_agg(jsonb_build_object('id', t.id, 'name', t.name, 'color', t.color, 'workspace_id', t.workspace_id))
+                FROM media_item_tags mit
+                JOIN tags t ON mit.tag_id = t.id
+                WHERE mit.media_item_id = mi.id
+            )
+        ) AS metadata
+    FROM media_items mi
+    LEFT JOIN workspaces w ON mi.workspace_id = w.id
+    LEFT JOIN users u ON mi.user_id = u.id
+    UNION ALL
+    SELECT
+        sa.id,
+        sa.workspace_id,
+        sa.user_id,
+        sa.created_at,
+        'source_asset'::text AS item_type,
+        'completed'::text AS status,
+        ARRAY[sa.gcs_uri] AS gcs_uris,
+        CASE
+            WHEN (sa.thumbnail_gcs_uri IS NOT NULL) THEN ARRAY[sa.thumbnail_gcs_uri]
+            ELSE '{}'::text[]
+        END AS thumbnail_uris,
+        sa.deleted_at,
+        w.name AS workspace_name,
+        u.picture AS user_picture,
+        u.email AS user_email,
+        sa.titles,
+        sa.descriptions,
+        jsonb_build_object(
+            'file_name', sa.original_filename,
+            'original_filename', sa.original_filename,
+            'mime_type', sa.mime_type,
+            'aspect_ratio', sa.aspect_ratio,
+            'asset_type', sa.asset_type,
+            'is_video', (sa.mime_type like 'video%'),
+            'is_audio', (sa.mime_type like 'audio%'),
+            'tags', (
+                SELECT jsonb_agg(jsonb_build_object('id', t.id, 'name', t.name, 'color', t.color, 'workspace_id', t.workspace_id))
+                FROM source_asset_tags sat
+                JOIN tags t ON sat.tag_id = t.id
+                WHERE sat.source_asset_id = sa.id
+            )
+        ) AS metadata
+    FROM source_assets sa
+    LEFT JOIN workspaces w ON sa.workspace_id = w.id
+    LEFT JOIN users u ON sa.user_id = u.id;
+    """
+    )

backend/src/common/base_dto.py

@@ -16,7 +16,7 @@
 
 from enum import Enum
 
-from pydantic import BaseModel, ConfigDict
+from pydantic import BaseModel, ConfigDict, Field
 from pydantic.alias_generators import to_camel
 
 
@@ -267,3 +267,14 @@ class BaseDto(BaseModel):
         populate_by_name=True,
         from_attributes=True,
     )
+
+
+class AssetReferenceDto(BaseDto):
+    id: int = Field(description="The ID of the asset.")
+    type: str = Field(
+        description="The type of asset: 'source_asset' or 'media_item'."
+    )
+    index: int | None = Field(
+        default=0,
+        description="The index of the media in the media item (if applicable).",
+    )

backend/src/common/media_utils.py

@@ -19,6 +19,7 @@
 import logging
 import os
 import pathlib
+import re
 import subprocess
 
 from PIL import Image as PILImage
@@ -237,3 +238,52 @@ def get_video_dimensions(video_path: str) -> tuple[int, int]:
     width = data["streams"][0]["width"]
     height = data["streams"][0]["height"]
     return width, height
+
+
+def extract_youtube_video_id(url: str | None) -> str | None:
+    """Extracts the 11-character video ID from a YouTube URL."""
+    if not url:
+        return None
+    trimmed = url.strip()
+    try:
+        from urllib.parse import urlparse, parse_qs
+
+        parsed = urlparse(trimmed)
+        hostname = parsed.hostname.lower() if parsed.hostname else ""
+
+        is_youtube = any(
+            h in hostname for h in ("youtube.com", "youtube-nocookie.com")
+        )
+        is_short = "youtu.be" in hostname
+
+        if is_youtube:
+            if any(p in parsed.path for p in ("/embed/", "/shorts/", "/v/")):
+                parts = parsed.path.split("/")
+                for p in parts:
+                    if len(p) == 11:
+                        return p
+            query_params = parse_qs(parsed.query)
+            v_list = query_params.get("v")
+            if v_list and len(v_list[0]) == 11:
+                return v_list[0]
+        elif is_short:
+            path_parts = parsed.path.strip("/").split("/")
+            if path_parts and len(path_parts[0]) == 11:
+                return path_parts[0]
+    except Exception:
+        pass
+    # Regex fallback
+    pattern = r"(?:youtu\.be\/|v\/|u\/\w\/|embed\/|shorts\/|watch\?v=|&v=)([a-zA-Z0-9_-]{11})"
+    match = re.search(pattern, trimmed)
+    return match.group(1) if match else None
+
+
+def get_youtube_aspect_ratio(url: str | None) -> str:
+    """Determines the aspect ratio of a YouTube video based on its URL (e.g., 9:16 for Shorts)."""
+    from src.common.base_dto import AspectRatioEnum
+
+    if not url:
+        return AspectRatioEnum.RATIO_16_9.value
+    if "/shorts/" in url:
+        return AspectRatioEnum.RATIO_9_16.value
+    return AspectRatioEnum.RATIO_16_9.value

backend/src/common/schema/media_item_model.py

@@ -83,6 +83,9 @@ class AssetRoleEnum(str, Enum):
     )
     VIDEO_REFERENCE = "video_reference"  # Video used as omni input reference
     AUDIO_REFERENCE = "audio_reference"  # Audio used as omni input reference
+    YOUTUBE_VIDEO_REFERENCE = (
+        "youtube_video_reference"  # YouTube video used as input reference
+    )
 
 
 class SourceAssetLink(BaseModel):

backend/src/galleries/dto/gallery_response_dto.py

@@ -25,9 +25,10 @@ class SourceAssetLinkResponse(SourceAssetLink):
     """Extends the source asset link with a presigned URL and GCS URI for frontend display."""
 
     presigned_url: str
-    gcs_uri: str
+    gcs_uri: str | None = None
     presigned_thumbnail_url: str | None = None
     mime_type: str | None = None
+    external_url: str | None = None
 
 
 class SourceMediaItemLinkResponse(SourceMediaItemLink):