opensource/galaxy
3
0
Fork 0
mirror of https://github.com/galaxyproject/galaxy.git synced 2026-09-24 16:30:27 +08:00
Code Issues Packages Projects Releases Wiki Activity

Merge pull request #21986 from dannon/fix/plugin-chat-inference-services

Browse Source 
[26.0] Use inference_services config for plugin chat proxy
This commit is contained in:
John Chilton
2026-03-10 09:02:42 -04:00
committed by GitHub
parent 728a645d5f b733debc3e
commit 570105babe
3 changed files with 143 additions and 30 deletions
Download Patch File Download Diff File Expand all files Collapse all files
lib/galaxy/config/schemas/config_schema.yml
+7 -5
View File
@@ -4094,9 +4094,11 @@ mapping:
type: any
required: false
desc: |
Configuration for AI inference services used by agents. Supports per-agent model, temperature, and token settings.
Agents inherit from 'default' configuration, which itself falls back to global ai_model/ai_api_key settings.
Example: inference_services: { default: { model: gpt-4o-mini, temperature: 0.7 } }
Configuration for AI inference services used by agents and visualization plugins.
Supports per-agent or per-plugin model, temperature, and token settings.
Valid keys include agent types (e.g. router, error_analysis) and plugin names (e.g. jupyterlite).
Agents and plugins inherit from 'default' configuration, which itself falls back to global ai_model/ai_api_key settings.
Example: inference_services: { default: { model: gpt-4o-mini }, jupyterlite: { model: gpt-4o } }
enable_tool_recommendations:
type: bool
@@ -4374,10 +4376,10 @@ mapping:
default: 86400
desc: |
The interval in seconds between attempts to delete all failed Galaxy job's working directories from the filesystem (every 24 hours by default) if enable_failed_jobs_working_directory_cleanup is ``true``. Runs in a Celery task.
enable_beta_tool_formats:
type: bool
default: false
required: false
desc: |
Enable beta tool formats (yaml, cwl, ...) which is a prerequisite for user defined tools.
Enable beta tool formats (yaml, cwl, ...) which is a prerequisite for user defined tools.
lib/galaxy/webapps/galaxy/api/plugins.py
+36 -9
View File
@@ -150,23 +150,49 @@ class FastAPIPlugins:
plugin_specs = plugin and plugin.config.get("specs")
plugin_ai_prompt = plugin_specs and plugin_specs.get("ai_prompt")
if plugin_ai_prompt:
return await self._open_ai_adapter(payload, plugin_ai_prompt)
return await self._open_ai_adapter(payload, plugin_ai_prompt, plugin_name)
else:
return self._create_error("Selected plugin has no AI prompt.")
else:
return self._create_error("Visualization registry is not available.")
def _get_plugin_config(self, plugin_name: str, key: str) -> Optional[str]:
"""Get config for a plugin with fallback through inference_services.
Precedence:
1. Plugin-specific: inference_services.<plugin_name>.<key>
2. Default inference: inference_services.default.<key>
3. Global config: ai_model / ai_api_key / ai_api_base_url
"""
inference_config = getattr(self.config, "inference_services", None)
if isinstance(inference_config, dict):
plugin_specific = inference_config.get(plugin_name)
if isinstance(plugin_specific, dict) and key in plugin_specific:
return plugin_specific[key]
default_config = inference_config.get("default")
if isinstance(default_config, dict) and key in default_config:
return default_config[key]
if key == "model":
return self.config.ai_model
elif key == "api_key":
return self.config.ai_api_key
elif key == "api_base_url":
return self.config.ai_api_base_url
return None
async def _open_ai_adapter(
self,
payload: ChatCompletionRequest,
prompt: str,
plugin_name: str,
):
"""Galaxy managed chat completion adapter with prompt injection"""
# Collect configuration
ai_api_key = self.config.ai_api_key
ai_api_base_url = self.config.ai_api_base_url
ai_model = self.config.ai_model
# Collect configuration via inference_services fallback chain
ai_api_key = self._get_plugin_config(plugin_name, "api_key")
ai_api_base_url = self._get_plugin_config(plugin_name, "api_base_url")
ai_model = self._get_plugin_config(plugin_name, "model")
if ai_api_key is None:
return self._create_error("AI service not configured: API key is required.")
if ai_model is None:
@@ -220,11 +246,12 @@ class FastAPIPlugins:
return self._create_error("Number of tools exceeded or invalid tools list.")
# Build openai client with timeout
client_kwargs = dict(api_key=ai_api_key, timeout=TIMEOUT)
if ai_api_base_url:
client_kwargs["base_url"] = ai_api_base_url
try:
client = AsyncOpenAI(**client_kwargs)
client = AsyncOpenAI(
api_key=ai_api_key,
timeout=TIMEOUT,
base_url=ai_api_base_url or None,
)
except Exception as e:
log.debug("Failed to initialize OpenAI client.", exc_info=e)
return self._create_error("Failed to initialize OpenAI client.", 500)
test/integration/test_plugins.py
+100 -16
View File
@@ -19,6 +19,16 @@ openai = pytest.importorskip("openai")
TEST_VISUALIZATION_PLUGINS_DIR = os.path.join(os.path.dirname(__file__), "test_visualization_plugins")
def _create_chat_payload(extra=None):
payload = {
"messages": [{"role": "user", "content": "hi"}],
"tools": [],
}
if extra:
payload.update(extra)
return payload
class TestVisualizationPluginsApi(IntegrationTestCase):
"""Tests for the visualization plugins API endpoints."""
@@ -81,15 +91,6 @@ class TestVisualizationPluginsApi(IntegrationTestCase):
assert "tests" in plugin
assert len(plugin["tests"]) >= 1
def _create_payload(self, extra=None):
payload = {
"messages": [{"role": "user", "content": "hi"}],
"tools": [],
}
if extra:
payload.update(extra)
return payload
def _post_payload(self, payload=None, anon=False):
return self._post("plugins/jupyterlite/chat/completions", payload, json=True, anon=anon)
@@ -100,7 +101,7 @@ class TestVisualizationPluginsApi(IntegrationTestCase):
mock_instance = MagicMock()
mock_instance.chat.completions.create = AsyncMock(return_value=mock_response)
mock_client.return_value = mock_instance
payload = self._create_payload()
payload = _create_chat_payload()
response = self._post_payload(payload, anon=False)
self._assert_status_code_is(response, 200)
assert response.json()["id"] == "test"
@@ -119,7 +120,7 @@ class TestVisualizationPluginsApi(IntegrationTestCase):
mock_instance.chat.completions.create = AsyncMock(return_value=stream_gen())
mock_instance.close = AsyncMock()
mock_client.return_value = mock_instance
payload = self._create_payload({"stream": True})
payload = _create_chat_payload({"stream": True})
response = self._post_payload(payload, anon=False)
self._assert_status_code_is(response, 200)
body = response.text
@@ -131,7 +132,7 @@ class TestVisualizationPluginsApi(IntegrationTestCase):
assert mock_instance.close.called
def test_tools_exceed_max(self):
payload = self._create_payload(
payload = _create_chat_payload(
{"tools": [{"type": "function", "function": {"name": "f", "parameters": {}}}] * 129}
)
response = self._post_payload(payload)
@@ -139,7 +140,7 @@ class TestVisualizationPluginsApi(IntegrationTestCase):
def test_tool_schema_too_large(self):
big_params = {"x": "a" * 20000}
payload = self._create_payload(
payload = _create_chat_payload(
{"tools": [{"type": "function", "function": {"name": "f", "parameters": big_params}}]}
)
response = self._post_payload(payload)
@@ -147,7 +148,7 @@ class TestVisualizationPluginsApi(IntegrationTestCase):
def test_exceed_max_messages(self):
msgs = {"messages": [{"role": "user", "content": "x"}] * (1024 + 1)}
payload = self._create_payload(msgs)
payload = _create_chat_payload(msgs)
response = self._post_payload(payload)
assert "You have exceeded the number of maximum messages" in response.json()["error"]["message"]
@@ -195,7 +196,7 @@ class TestVisualizationPluginsApi(IntegrationTestCase):
mock_instance = MagicMock()
mock_instance.chat.completions.create = AsyncMock(return_value=mock_response)
mock_client.return_value = mock_instance
payload = self._create_payload(
payload = _create_chat_payload(
{
"tools": [
{
@@ -234,8 +235,91 @@ class TestVisualizationPluginsApi(IntegrationTestCase):
mock_instance = MagicMock()
mock_instance.chat.completions.create = AsyncMock(side_effect=MockOpenAIError())
mock_client.return_value = mock_instance
response = self._post_payload(self._create_payload())
response = self._post_payload(_create_chat_payload())
self._assert_status_code_is(response, 404)
body = response.json()
assert body["error"]["message"] == "original error message"
assert body["error"]["type"] == "api_error"
class TestPluginsInferenceServicesConfig(IntegrationTestCase):
"""Tests for inference_services config resolution in plugins."""
@classmethod
def handle_galaxy_config_kwds(cls, config) -> None:
config["ai_api_key"] = "global_key"
config["ai_api_base_url"] = "http://global-url"
config["ai_model"] = "global_model"
config["visualization_plugins_directory"] = TEST_VISUALIZATION_PLUGINS_DIR
config["inference_services"] = {
"default": {
"model": "default_model",
"api_key": "default_key",
"api_base_url": "http://default-url",
},
"jupyterlite": {
"model": "jupyterlite_model",
"api_key": "jupyterlite_key",
"api_base_url": "http://jupyterlite-url",
},
}
def _post_payload(self, payload=None, anon=False):
return self._post("plugins/jupyterlite/chat/completions", payload, json=True, anon=anon)
@patch("galaxy.webapps.galaxy.api.plugins.AsyncOpenAI")
def test_plugin_specific_config_used(self, mock_client):
"""Plugin-specific inference_services config overrides default and global."""
mock_response = MagicMock()
mock_response.model_dump.return_value = {"id": "test", "choices": []}
mock_instance = MagicMock()
mock_instance.chat.completions.create = AsyncMock(return_value=mock_response)
mock_client.return_value = mock_instance
response = self._post_payload(payload=_create_chat_payload())
self._assert_status_code_is(response, 200)
call_kwargs = mock_instance.chat.completions.create.call_args.kwargs
assert call_kwargs["model"] == "jupyterlite_model"
client_kwargs = mock_client.call_args.kwargs
assert client_kwargs["api_key"] == "jupyterlite_key"
assert client_kwargs["base_url"] == "http://jupyterlite-url"
class TestPluginsInferenceServicesDefault(IntegrationTestCase):
"""Tests that inference_services.default is used when no plugin-specific config exists."""
@classmethod
def handle_galaxy_config_kwds(cls, config) -> None:
config["ai_api_key"] = "global_key"
config["ai_api_base_url"] = "http://global-url"
config["ai_model"] = "global_model"
config["visualization_plugins_directory"] = TEST_VISUALIZATION_PLUGINS_DIR
config["inference_services"] = {
"default": {
"model": "default_model",
"api_key": "default_key",
"api_base_url": "http://default-url",
},
}
def _post_payload(self, payload=None, anon=False):
return self._post("plugins/jupyterlite/chat/completions", payload, json=True, anon=anon)
@patch("galaxy.webapps.galaxy.api.plugins.AsyncOpenAI")
def test_default_config_fallback(self, mock_client):
"""inference_services.default is used when no plugin-specific entry exists."""
mock_response = MagicMock()
mock_response.model_dump.return_value = {"id": "test", "choices": []}
mock_instance = MagicMock()
mock_instance.chat.completions.create = AsyncMock(return_value=mock_response)
mock_client.return_value = mock_instance
response = self._post_payload(payload=_create_chat_payload())
self._assert_status_code_is(response, 200)
call_kwargs = mock_instance.chat.completions.create.call_args.kwargs
assert call_kwargs["model"] == "default_model"
client_kwargs = mock_client.call_args.kwargs
assert client_kwargs["api_key"] == "default_key"
assert client_kwargs["base_url"] == "http://default-url"