fix: Handle HTTP/HTTPS URLs for media files in LiteLLM content conversion

For providers that typically require file IDs (like OpenAI and Azure), if a file URI is an HTTP/HTTPS URL and the MIME type is image, video, or audio, convert it to the corresponding URL-based content type (e.g., "image_url") instead of using the generic "file" type

Close #4112

Co-authored-by: George Weale <gweale@google.com>
PiperOrigin-RevId: 863311703

tests/unittests/models/test_litellm.py

@@ -2330,6 +2330,55 @@ async def test_get_content_file_uri_file_id_required_falls_back_to_text(
   ]
 
 
+@pytest.mark.asyncio
+@pytest.mark.parametrize(
+    "provider,model",
+    [
+        ("openai", "openai/gpt-4o"),
+        ("azure", "azure/gpt-4"),
+    ],
+)
+@pytest.mark.parametrize(
+    "file_uri,mime_type,expected_type",
+    [
+        pytest.param(
+            "https://example.com/image.png",
+            "image/png",
+            "image_url",
+            id="image",
+        ),
+        pytest.param(
+            "https://example.com/video.mp4",
+            "video/mp4",
+            "video_url",
+            id="video",
+        ),
+        pytest.param(
+            "https://example.com/audio.mp3",
+            "audio/mpeg",
+            "audio_url",
+            id="audio",
+        ),
+    ],
+)
+async def test_get_content_file_uri_media_url_file_id_required_uses_url_type(
+    provider, model, file_uri, mime_type, expected_type
+):
+  parts = [
+      types.Part(
+          file_data=types.FileData(
+              file_uri=file_uri,
+              mime_type=mime_type,
+          )
+      )
+  ]
+  content = await _get_content(parts, provider=provider, model=model)
+  assert content == [{
+      "type": expected_type,
+      expected_type: {"url": file_uri},
+  }]
+
+
 @pytest.mark.asyncio
 @pytest.mark.parametrize(
     "provider,model",
@@ -2353,6 +2402,53 @@ async def test_get_content_file_uri_file_id_required_preserves_file_id(
   assert content == [{"type": "file", "file": {"file_id": "file-abc123"}}]
 
 
+@pytest.mark.asyncio
+@pytest.mark.parametrize(
+    "provider,model",
+    [
+        ("openai", "openai/gpt-4o"),
+        ("azure", "azure/gpt-4"),
+    ],
+)
+async def test_get_content_file_uri_http_pdf_file_id_required_falls_back_to_text(
+    provider, model
+):
+  file_uri = "https://example.com/document.pdf"
+  parts = [
+      types.Part(
+          file_data=types.FileData(
+              file_uri=file_uri,
+              mime_type="application/pdf",
+              display_name="document.pdf",
+          )
+      )
+  ]
+  content = await _get_content(parts, provider=provider, model=model)
+  assert content == [
+      {"type": "text", "text": '[File reference: "document.pdf"]'}
+  ]
+
+
+@pytest.mark.asyncio
+async def test_get_content_file_uri_http_pdf_non_file_id_provider_uses_file():
+  file_uri = "https://example.com/document.pdf"
+  parts = [
+      types.Part(
+          file_data=types.FileData(
+              file_uri=file_uri,
+              mime_type="application/pdf",
+          )
+      )
+  ]
+  content = await _get_content(
+      parts, provider="vertex_ai", model="vertex_ai/gemini-2.5-flash"
+  )
+  assert content == [{
+      "type": "file",
+      "file": {"file_id": file_uri, "format": "application/pdf"},
+  }]
+
+
 @pytest.mark.asyncio
 async def test_get_content_file_uri_anthropic_falls_back_to_text():
   parts = [