feat: expose artifact URLs to the model when available

SaveFilesAsArtifactsPlugin now resolves the canonical URI for each saved artifact. When the URI is model-accessible (`gs://`, `https://`, `http://`), we add a `Part(file_data=...)` so the LLM can fetch the filedirectly while still emitting the placeholder. If no model-accessible URI exists, we retain the original inline blob alongside the placeholder to preserve access to the uploaded bytes.

Close #2016

Co-authored-by: George Weale <gweale@google.com>
PiperOrigin-RevId: 828786519
This commit is contained in:
George Weale
2025-11-05 22:17:37 -08:00
committed by Copybara-Service
parent 52db15f133
commit e3caf79139
2 changed files with 188 additions and 63 deletions
@@ -17,6 +17,7 @@ from __future__ import annotations
import copy
import logging
from typing import Optional
import urllib.parse
from google.genai import types
@@ -25,6 +26,11 @@ from .base_plugin import BasePlugin
logger = logging.getLogger('google_adk.' + __name__)
# Schemes supported by our current LLM connectors. Vertex exposes `gs://` while
# hosted endpoints use HTTPS. Expand this list when BaseLlm surfaces provider
# capabilities.
_MODEL_ACCESSIBLE_URI_SCHEMES = {'gs', 'https', 'http'}
class SaveFilesAsArtifactsPlugin(BasePlugin):
"""A plugin that saves files embedded in user messages as artifacts.
@@ -75,8 +81,9 @@ class SaveFilesAsArtifactsPlugin(BasePlugin):
continue
try:
# Use display_name if available; otherwise, generate a filename
file_name = part.inline_data.display_name
# Use display_name if available, otherwise generate a filename
inline_data = part.inline_data
file_name = inline_data.display_name
if not file_name:
file_name = f'artifact_{invocation_context.invocation_id}_{i}'
logger.info(
@@ -87,7 +94,7 @@ class SaveFilesAsArtifactsPlugin(BasePlugin):
display_name = file_name
# Create a copy to stop mutation of the saved artifact if the original part is modified
await invocation_context.artifact_service.save_artifact(
version = await invocation_context.artifact_service.save_artifact(
app_name=invocation_context.app_name,
user_id=invocation_context.user_id,
session_id=invocation_context.session.id,
@@ -95,10 +102,28 @@ class SaveFilesAsArtifactsPlugin(BasePlugin):
artifact=copy.copy(part),
)
# Replace the inline data with a placeholder text (using the clean name)
new_parts.append(
types.Part(text=f'[Uploaded Artifact: "{display_name}"]')
placeholder_part = types.Part(
text=f'[Uploaded Artifact: "{display_name}"]'
)
new_parts.append(placeholder_part)
file_part = await self._build_file_reference_part(
invocation_context=invocation_context,
filename=file_name,
version=version,
mime_type=inline_data.mime_type,
display_name=display_name,
)
if file_part:
new_parts.append(file_part)
else:
logger.debug(
'Artifact %s is not exposed via a model-accessible URI; keeping'
' inline data in user message.',
file_name,
)
new_parts.append(part)
modified = True
logger.info(f'Successfully saved artifact: {file_name}')
@@ -112,3 +137,58 @@ class SaveFilesAsArtifactsPlugin(BasePlugin):
return types.Content(role=user_message.role, parts=new_parts)
else:
return None
async def _build_file_reference_part(
self,
*,
invocation_context: InvocationContext,
filename: str,
version: int,
mime_type: Optional[str],
display_name: str,
) -> Optional[types.Part]:
"""Constructs a file reference part if the artifact URI is model-accessible."""
artifact_service = invocation_context.artifact_service
if not artifact_service:
return None
try:
artifact_version = await artifact_service.get_artifact_version(
app_name=invocation_context.app_name,
user_id=invocation_context.user_id,
session_id=invocation_context.session.id,
filename=filename,
version=version,
)
except Exception as exc: # pylint: disable=broad-except
logger.warning(
'Failed to resolve artifact version for %s: %s', filename, exc
)
return None
if (
not artifact_version
or not artifact_version.canonical_uri
or not _is_model_accessible_uri(artifact_version.canonical_uri)
):
return None
file_data = types.FileData(
file_uri=artifact_version.canonical_uri,
mime_type=mime_type or artifact_version.mime_type,
display_name=display_name,
)
return types.Part(file_data=file_data)
def _is_model_accessible_uri(uri: str) -> bool:
try:
parsed = urllib.parse.urlparse(uri)
except ValueError:
return False
if not parsed.scheme:
return False
return parsed.scheme.lower() in _MODEL_ACCESSIBLE_URI_SCHEMES
@@ -11,13 +11,13 @@
# 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.
from __future__ import annotations
from unittest.mock import AsyncMock
from unittest.mock import Mock
from google.adk.agents.invocation_context import InvocationContext
from google.adk.artifacts.base_artifact_service import ArtifactVersion
from google.adk.plugins.save_files_as_artifacts_plugin import SaveFilesAsArtifactsPlugin
from google.genai import types
import pytest
@@ -32,17 +32,32 @@ class TestSaveFilesAsArtifactsPlugin:
# Mock invocation context
self.mock_context = Mock(spec=InvocationContext)
self.mock_context.artifact_service = AsyncMock()
self.mock_context.app_name = "test_app"
self.mock_context.user_id = "test_user"
self.mock_context.invocation_id = "test_invocation_123"
self.mock_context.session = Mock()
self.mock_context.session.id = "test_session"
artifact_service = Mock()
artifact_service.save_artifact = AsyncMock(return_value=0)
async def _mock_get_artifact_version(**kwargs):
filename = kwargs.get("filename", "unknown_file")
version = kwargs.get("version", 0)
return ArtifactVersion(
version=version,
canonical_uri=f"gs://mock-bucket/{filename}/versions/{version}",
mime_type="application/pdf",
)
artifact_service.get_artifact_version = AsyncMock(
side_effect=_mock_get_artifact_version
)
self.mock_context.artifact_service = artifact_service
@pytest.mark.asyncio
async def test_save_files_with_display_name(self):
"""Test saving files when inline_data has display_name."""
# Create a message with inline data
inline_data = types.Blob(
display_name="test_document.pdf",
data=b"test data",
@@ -52,12 +67,10 @@ class TestSaveFilesAsArtifactsPlugin:
original_part = types.Part(inline_data=inline_data)
user_message = types.Content(parts=[original_part])
# Execute the plugin
result = await self.plugin.on_user_message_callback(
invocation_context=self.mock_context, user_message=user_message
)
# Verify artifact was saved with correct filename (session-scoped by default)
self.mock_context.artifact_service.save_artifact.assert_called_once_with(
app_name="test_app",
user_id="test_user",
@@ -66,13 +79,20 @@ class TestSaveFilesAsArtifactsPlugin:
artifact=original_part,
)
# Verify message was modified with placeholder (clean name)
assert result
assert len(result.parts) == 2
assert result.parts[0].text == '[Uploaded Artifact: "test_document.pdf"]'
assert result.parts[1].file_data
assert (
result.parts[1].file_data.file_uri
== "gs://mock-bucket/test_document.pdf/versions/0"
)
assert result.parts[1].file_data.display_name == "test_document.pdf"
assert result.parts[1].file_data.mime_type == "application/pdf"
@pytest.mark.asyncio
async def test_save_files_without_display_name(self):
"""Test saving files when inline_data has no display_name."""
# Create inline data without display_name
inline_data = types.Blob(
display_name=None, data=b"test data", mime_type="application/pdf"
)
@@ -80,12 +100,10 @@ class TestSaveFilesAsArtifactsPlugin:
original_part = types.Part(inline_data=inline_data)
user_message = types.Content(parts=[original_part])
# Execute the plugin
result = await self.plugin.on_user_message_callback(
invocation_context=self.mock_context, user_message=user_message
)
# Verify artifact was saved with generated filename (session-scoped by default)
expected_filename = "artifact_test_invocation_123_0"
self.mock_context.artifact_service.save_artifact.assert_called_once_with(
app_name="test_app",
@@ -95,21 +113,22 @@ class TestSaveFilesAsArtifactsPlugin:
artifact=original_part,
)
# Verify message was modified with generated filename (clean name)
generated_display_name = "artifact_test_invocation_123_0"
assert result
assert len(result.parts) == 2
assert result.parts[0].text == f'[Uploaded Artifact: "{expected_filename}"]'
assert result.parts[1].file_data
assert (
result.parts[0].text
== f'[Uploaded Artifact: "{generated_display_name}"]'
result.parts[1].file_data.file_uri
== "gs://mock-bucket/artifact_test_invocation_123_0/versions/0"
)
assert result.parts[1].file_data.display_name == expected_filename
@pytest.mark.asyncio
async def test_multiple_files_in_message(self):
"""Test handling multiple files in a single message."""
# Create message with multiple inline data parts
inline_data1 = types.Blob(
display_name="file1.txt", data=b"file1 content", mime_type="text/plain"
)
inline_data2 = types.Blob(
display_name="file2.jpg", data=b"file2 content", mime_type="image/jpeg"
)
@@ -122,49 +141,89 @@ class TestSaveFilesAsArtifactsPlugin:
]
)
# Execute the plugin
result = await self.plugin.on_user_message_callback(
invocation_context=self.mock_context, user_message=user_message
)
# Verify both artifacts were saved
assert self.mock_context.artifact_service.save_artifact.call_count == 2
# Check first file
first_call = (
self.mock_context.artifact_service.save_artifact.call_args_list[0]
)
assert first_call[1]["filename"] == "file1.txt"
# Check second file
second_call = (
self.mock_context.artifact_service.save_artifact.call_args_list[1]
)
assert first_call[1]["filename"] == "file1.txt"
assert second_call[1]["filename"] == "file2.jpg"
# Verify message parts were modified correctly (clean names)
assert result
assert len(result.parts) == 5
assert result.parts[0].text == '[Uploaded Artifact: "file1.txt"]'
assert result.parts[1].text == "Some text between files" # Unchanged
assert result.parts[2].text == '[Uploaded Artifact: "file2.jpg"]'
assert result.parts[1].file_data
assert (
result.parts[1].file_data.file_uri
== "gs://mock-bucket/file1.txt/versions/0"
)
assert result.parts[1].file_data.display_name == "file1.txt"
assert result.parts[2].text == "Some text between files"
assert result.parts[3].text == '[Uploaded Artifact: "file2.jpg"]'
assert result.parts[4].file_data
assert (
result.parts[4].file_data.file_uri
== "gs://mock-bucket/file2.jpg/versions/0"
)
assert result.parts[4].file_data.display_name == "file2.jpg"
@pytest.mark.asyncio
async def test_unsupported_canonical_uri_keeps_inline_data(self):
"""Fallback to inline data when artifact URI is not model-accessible."""
inline_data = types.Blob(
display_name="local_only.png",
data=b"image data",
mime_type="image/png",
)
artifact_service = self.mock_context.artifact_service
original_side_effect = artifact_service.get_artifact_version.side_effect
async def _memory_only_version(**kwargs):
return ArtifactVersion(
version=kwargs.get("version", 0),
canonical_uri=(
"memory://apps/test_app/users/test_user/sessions/test_session/"
"artifacts/local_only.png/versions/0"
),
mime_type="image/png",
)
artifact_service.get_artifact_version.side_effect = _memory_only_version
try:
user_message = types.Content(parts=[types.Part(inline_data=inline_data)])
result = await self.plugin.on_user_message_callback(
invocation_context=self.mock_context, user_message=user_message
)
assert result
assert len(result.parts) == 2
assert result.parts[0].text == '[Uploaded Artifact: "local_only.png"]'
assert result.parts[1].inline_data == inline_data
finally:
artifact_service.get_artifact_version.side_effect = original_side_effect
@pytest.mark.asyncio
async def test_no_artifact_service(self):
"""Test behavior when artifact service is not available."""
# Set artifact service to None
self.mock_context.artifact_service = None
inline_data = types.Blob(
display_name="test.pdf", data=b"test data", mime_type="application/pdf"
)
user_message = types.Content(parts=[types.Part(inline_data=inline_data)])
# Execute the plugin
result = await self.plugin.on_user_message_callback(
invocation_context=self.mock_context, user_message=user_message
)
# Should return original message unchanged
assert result == user_message
assert result.parts[0].inline_data == inline_data
@@ -173,15 +232,11 @@ class TestSaveFilesAsArtifactsPlugin:
"""Test behavior when message has no parts."""
user_message = types.Content(parts=[])
# Execute the plugin
result = await self.plugin.on_user_message_callback(
invocation_context=self.mock_context, user_message=user_message
)
# Should return None to proceed with original message
assert result is None
# Should not try to save any artifacts
self.mock_context.artifact_service.save_artifact.assert_not_called()
@pytest.mark.asyncio
@@ -191,21 +246,16 @@ class TestSaveFilesAsArtifactsPlugin:
parts=[types.Part(text="Hello world"), types.Part(text="No files here")]
)
# Execute the plugin
result = await self.plugin.on_user_message_callback(
invocation_context=self.mock_context, user_message=user_message
)
# Should return None to proceed with original message
assert result is None
# Should not try to save any artifacts
self.mock_context.artifact_service.save_artifact.assert_not_called()
@pytest.mark.asyncio
async def test_save_artifact_failure(self):
"""Test behavior when saving artifact fails."""
# Mock save_artifact to raise an exception
self.mock_context.artifact_service.save_artifact.side_effect = Exception(
"Storage error"
)
@@ -213,33 +263,28 @@ class TestSaveFilesAsArtifactsPlugin:
inline_data = types.Blob(
display_name="test.pdf", data=b"test data", mime_type="application/pdf"
)
user_message = types.Content(parts=[types.Part(inline_data=inline_data)])
original_part = types.Part(inline_data=inline_data)
user_message = types.Content(parts=[original_part])
# Execute the plugin - should not raise exception
result = await self.plugin.on_user_message_callback(
invocation_context=self.mock_context, user_message=user_message
)
# Should return None when saving fails (no modifications made)
assert result is None
@pytest.mark.asyncio
async def test_mixed_success_and_failure(self):
"""Test behavior when some files save successfully and others fail."""
# Mock save_artifact to succeed on first call, fail on second
save_calls = 0
def mock_save_artifact(*_args, **_kwargs):
async def _save_side_effect(*_args, **_kwargs):
nonlocal save_calls
save_calls += 1
if save_calls == 2:
raise Exception("Storage error on second file")
return AsyncMock()
return 0
self.mock_context.artifact_service.save_artifact.side_effect = (
mock_save_artifact
_save_side_effect
)
inline_data1 = types.Blob(
@@ -247,7 +292,6 @@ class TestSaveFilesAsArtifactsPlugin:
data=b"success data",
mime_type="application/pdf",
)
inline_data2 = types.Blob(
display_name="failure.pdf",
data=b"failure data",
@@ -259,17 +303,16 @@ class TestSaveFilesAsArtifactsPlugin:
parts=[types.Part(inline_data=inline_data1), original_part2]
)
# Execute the plugin
result = await self.plugin.on_user_message_callback(
invocation_context=self.mock_context, user_message=user_message
)
# First file should be replaced with placeholder (clean name)
assert result
assert len(result.parts) == 3
assert result.parts[0].text == '[Uploaded Artifact: "success.pdf"]'
# Second file should remain unchanged due to failure
assert result.parts[1] == original_part2
assert result.parts[1].inline_data == inline_data2
assert result.parts[1].file_data
assert result.parts[2] == original_part2
assert result.parts[2].inline_data == inline_data2
@pytest.mark.asyncio
async def test_placeholder_text_format(self):
@@ -277,19 +320,21 @@ class TestSaveFilesAsArtifactsPlugin:
inline_data = types.Blob(
display_name="test file with spaces.docx",
data=b"document data",
mime_type="application/vnd.openxmlformats-officedocument.wordprocessingml.document",
mime_type=(
"application/vnd.openxmlformats-officedocument."
"wordprocessingml.document"
),
)
user_message = types.Content(parts=[types.Part(inline_data=inline_data)])
# Execute the plugin
result = await self.plugin.on_user_message_callback(
invocation_context=self.mock_context, user_message=user_message
)
# Verify exact format of placeholder text (clean name)
expected_text = '[Uploaded Artifact: "test file with spaces.docx"]'
assert result.parts[0].text == expected_text
assert result.parts[1].file_data
def test_plugin_name_default(self):
"""Test that plugin has correct default name."""