aiqtoolkit 1.1.0a20250429__py3-none-any.whl

aiq/front_ends/console/console_front_end_plugin.py

@@ -0,0 +1,107 @@
+# SPDX-FileCopyrightText: Copyright (c) 2025, NVIDIA CORPORATION & AFFILIATES. All rights reserved.
+# SPDX-License-Identifier: Apache-2.0
+#
+# 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.
+
+import asyncio
+import logging
+from io import StringIO
+
+import click
+from colorama import Fore
+
+from aiq.builder.workflow_builder import WorkflowBuilder
+from aiq.data_models.interactive import HumanPromptModelType
+from aiq.data_models.interactive import HumanResponse
+from aiq.data_models.interactive import HumanResponseText
+from aiq.data_models.interactive import InteractionPrompt
+from aiq.front_ends.console.console_front_end_config import ConsoleFrontEndConfig
+from aiq.front_ends.simple_base.simple_front_end_plugin_base import SimpleFrontEndPluginBase
+from aiq.runtime.session import AIQSessionManager
+
+logger = logging.getLogger(__name__)
+
+
+async def prompt_for_input_cli(question: InteractionPrompt) -> HumanResponse:
+    """
+    A simple CLI-based callback.
+    Takes question as str, returns the typed line as str.
+    """
+
+    if question.content.input_type == HumanPromptModelType.TEXT:
+        user_response = click.prompt(text=question.content.text)
+
+        return HumanResponseText(text=user_response)
+
+    raise ValueError("Unsupported human propmt input type. The run command only supports the 'HumanPromptText' "
+                     "input type. Please use the 'serve' command to ensure full support for all input types.")
+
+
+class ConsoleFrontEndPlugin(SimpleFrontEndPluginBase[ConsoleFrontEndConfig]):
+
+    async def pre_run(self):
+
+        if (not self.front_end_config.input_query and not self.front_end_config.input_file):
+            raise click.UsageError("Must specify either --input_query or --input_file")
+
+    async def run(self):
+
+        # Must yield the workflow function otherwise it cleans up
+        async with WorkflowBuilder.from_config(config=self.full_config) as builder:
+
+            session_manager: AIQSessionManager = None
+
+            if logger.isEnabledFor(logging.INFO):
+                stream = StringIO()
+
+                self.full_config.print_summary(stream=stream)
+
+                click.echo(stream.getvalue())
+
+                workflow = builder.build()
+                session_manager = AIQSessionManager(workflow)
+
+            await self.run_workflow(session_manager)
+
+    async def run_workflow(self, session_manager: AIQSessionManager = None):
+
+        runner_outputs = None
+
+        if (self.front_end_config.input_query):
+
+            async def run_single_query(query):
+
+                async with session_manager.session(user_input_callback=prompt_for_input_cli) as session:
+                    async with session.run(query) as runner:
+                        base_output = await runner.result(to_type=str)
+
+                        return base_output
+
+            # Convert to a list
+            input_list = list(self.front_end_config.input_query)
+            logger.debug("Processing input: %s", self.front_end_config.input_query)
+
+            runner_outputs = await asyncio.gather(*[run_single_query(query) for query in input_list])
+
+        elif (self.front_end_config.input_file):
+
+            # Run the workflow
+            with open(self.front_end_config.input_file, "r", encoding="utf-8") as f:
+
+                async with session_manager.workflow.run(f) as runner:
+                    runner_outputs = await runner.result(to_type=str)
+        else:
+            assert False, "Should not reach here. Should have been caught by pre_run"
+
+        # Print result
+        logger.info(f"\n{'-' * 50}\n{Fore.GREEN}Workflow Result:\n%s{Fore.RESET}\n{'-' * 50}", runner_outputs)

aiq/front_ends/console/register.py

@@ -0,0 +1,25 @@
+# SPDX-FileCopyrightText: Copyright (c) 2025, NVIDIA CORPORATION & AFFILIATES. All rights reserved.
+# SPDX-License-Identifier: Apache-2.0
+#
+# 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.
+
+from aiq.cli.register_workflow import register_front_end
+from aiq.data_models.config import AIQConfig
+from aiq.front_ends.console.console_front_end_config import ConsoleFrontEndConfig
+
+
+@register_front_end(config_type=ConsoleFrontEndConfig)
+async def register_fastapi_front_end(config: ConsoleFrontEndConfig, full_config: AIQConfig):
+    from aiq.front_ends.console.console_front_end_plugin import ConsoleFrontEndPlugin
+
+    yield ConsoleFrontEndPlugin(full_config=full_config)

aiq/front_ends/cron/__init__.py

@@ -0,0 +1,14 @@
+# SPDX-FileCopyrightText: Copyright (c) 2024-2025, NVIDIA CORPORATION & AFFILIATES. All rights reserved.
+# SPDX-License-Identifier: Apache-2.0
+#
+# 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.

aiq/front_ends/fastapi/__init__.py

@@ -0,0 +1,14 @@
+# SPDX-FileCopyrightText: Copyright (c) 2024-2025, NVIDIA CORPORATION & AFFILIATES. All rights reserved.
+# SPDX-License-Identifier: Apache-2.0
+#
+# 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.

aiq/front_ends/fastapi/fastapi_front_end_config.py

@@ -0,0 +1,150 @@
+# SPDX-FileCopyrightText: Copyright (c) 2025, NVIDIA CORPORATION & AFFILIATES. All rights reserved.
+# SPDX-License-Identifier: Apache-2.0
+#
+# 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.
+
+import logging
+import typing
+from datetime import datetime
+
+from pydantic import BaseModel
+from pydantic import Field
+
+from aiq.data_models.front_end import FrontEndBaseConfig
+from aiq.data_models.step_adaptor import StepAdaptorConfig
+
+logger = logging.getLogger(__name__)
+
+
+class AIQEvaluateRequest(BaseModel):
+    """Request model for the evaluate endpoint."""
+    config_file: str = Field(description="Path to the configuration file for evaluation")
+    job_id: str | None = Field(default=None, description="Unique identifier for the evaluation job")
+    reps: int = Field(default=1, description="Number of repetitions for the evaluation, defaults to 1")
+    expiry_seconds: int = Field(
+        default=3600,
+        description="Optional time (in seconds) before the job expires. Clamped between 600 (10 min) and 86400 (24h).")
+
+
+class AIQEvaluateResponse(BaseModel):
+    """Response model for the evaluate endpoint."""
+    job_id: str = Field(description="Unique identifier for the evaluation job")
+    status: str = Field(description="Current status of the evaluation job")
+
+
+class AIQEvaluateStatusResponse(BaseModel):
+    """Response model for the evaluate status endpoint."""
+    job_id: str = Field(description="Unique identifier for the evaluation job")
+    status: str = Field(description="Current status of the evaluation job")
+    config_file: str = Field(description="Path to the configuration file used for evaluation")
+    error: str | None = Field(default=None, description="Error message if the job failed")
+    output_path: str | None = Field(default=None,
+                                    description="Path to the output file if the job completed successfully")
+    created_at: datetime = Field(description="Timestamp when the job was created")
+    updated_at: datetime = Field(description="Timestamp when the job was last updated")
+    expires_at: datetime | None = Field(default=None, description="Timestamp when the job will expire")
+
+
+class FastApiFrontEndConfig(FrontEndBaseConfig, name="fastapi"):
+    """
+    A FastAPI based front end that allows an AgentIQ workflow to be served as a microservice.
+    """
+
+    class EndpointBase(BaseModel):
+
+        method: typing.Literal["GET", "POST", "PUT", "DELETE"]
+        description: str
+        path: str | None = Field(
+            default=None,
+            description=("Path for the default workflow. If None, no workflow endpoint is created."),
+        )
+        websocket_path: str | None = Field(
+            default=None,
+            description=("Path for the websocket. If None, no websocket is created."),
+        )
+        openai_api_path: str | None = Field(
+            default=None,
+            description=("Path for the default workflow using the OpenAI API Specification. "
+                         "If None, no workflow endpoint with the OpenAI API Specification is created."),
+        )
+
+    class Endpoint(EndpointBase):
+        function_name: str = Field(description="The name of the function to call for this endpoint")
+
+    class CrossOriginResourceSharing(BaseModel):
+        allow_origins: list[str] | None = Field(
+            default=None, description=" A list of origins that should be permitted to make cross-origin requests.")
+        allow_origin_regex: str | None = Field(
+            default=None,
+            description="A permitted regex string to match against origins to make cross-origin requests",
+        )
+        allow_methods: list[str] | None = Field(
+            default_factory=lambda: ['GET'],
+            description="A list of HTTP methods that should be allowed for cross-origin requests.")
+        allow_headers: list[str] | None = Field(
+            default_factory=list,
+            description="A list of HTTP request headers that should be supported for cross-origin requests.")
+        allow_credentials: bool | None = Field(
+            default=False,
+            description="Indicate that cookies should be supported for cross-origin requests.",
+        )
+        expose_headers: list[str] | None = Field(
+            default_factory=list,
+            description="Indicate any response headers that should be made accessible to the browser.",
+        )
+        max_age: int | None = Field(
+            default=600,
+            description="Sets a maximum time in seconds for browsers to cache CORS responses.",
+        )
+
+    root_path: str = Field(default="", description="The root path for the API")
+    host: str = Field(default="localhost", description="Host to bind the server to")
+    port: int = Field(default=8000, description="Port to bind the server to", ge=0, le=65535)
+    reload: bool = Field(default=False, description="Enable auto-reload for development")
+    workers: int = Field(default=1, description="Number of workers to run", ge=1)
+    step_adaptor: StepAdaptorConfig = StepAdaptorConfig()
+
+    workflow: typing.Annotated[EndpointBase, Field(description="Endpoint for the default workflow.")] = EndpointBase(
+        method="POST",
+        path="/generate",
+        websocket_path="/websocket",
+        openai_api_path="/chat",
+        description="Executes the default AgentIQ workflow from the loaded configuration ",
+    )
+
+    evaluate: typing.Annotated[EndpointBase, Field(description="Endpoint for evaluating workflows.")] = EndpointBase(
+        method="POST",
+        path="/evaluate",
+        description="Evaluates the performance and accuracy of the workflow on a dataset",
+    )
+
+    endpoints: list[Endpoint] = Field(
+        default_factory=list,
+        description=(
+            "Additional endpoints to add to the FastAPI app which run functions within the AgentIQ configuration. "
+            "Each endpoint must have a unique path."))
+
+    cors: CrossOriginResourceSharing = Field(
+        default_factory=CrossOriginResourceSharing,
+        description="Cross origin resource sharing configuration for the FastAPI app")
+
+    use_gunicorn: bool = Field(
+        default=False,
+        description="Use Gunicorn to run the FastAPI app",
+    )
+    runner_class: str | None = Field(
+        default=None,
+        description=("The AgentIQ runner class to use when launching the FastAPI app from multiple processes. "
+                     "Each runner is responsible for loading and running the AgentIQ workflow. "
+                     "Note: This is different from the worker class used by Gunicorn."),
+    )

aiq/front_ends/fastapi/fastapi_front_end_plugin.py

@@ -0,0 +1,103 @@
+# SPDX-FileCopyrightText: Copyright (c) 2025, NVIDIA CORPORATION & AFFILIATES. All rights reserved.
+# SPDX-License-Identifier: Apache-2.0
+#
+# 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.
+
+import os
+import tempfile
+import typing
+
+from aiq.builder.front_end import FrontEndBase
+from aiq.front_ends.fastapi.fastapi_front_end_config import FastApiFrontEndConfig
+from aiq.front_ends.fastapi.fastapi_front_end_plugin_worker import FastApiFrontEndPluginWorkerBase
+from aiq.front_ends.fastapi.main import get_app
+from aiq.utils.io.yaml_tools import yaml_dump
+
+
+class FastApiFrontEndPlugin(FrontEndBase[FastApiFrontEndConfig]):
+
+    def get_worker_class(self) -> type[FastApiFrontEndPluginWorkerBase]:
+        from aiq.front_ends.fastapi.fastapi_front_end_plugin_worker import FastApiFrontEndPluginWorker
+
+        return FastApiFrontEndPluginWorker
+
+    @typing.final
+    def get_worker_class_name(self) -> str:
+
+        if (self.front_end_config.runner_class):
+            return self.front_end_config.runner_class
+
+        worker_class = self.get_worker_class()
+
+        return f"{worker_class.__module__}.{worker_class.__qualname__}"
+
+    async def run(self):
+
+        # Write the entire config to a temporary file
+        with tempfile.NamedTemporaryFile(mode="w", prefix="aiq_config", suffix=".yml", delete=True) as config_file:
+
+            # Get as dict
+            config_dict = self.full_config.model_dump(mode="json", by_alias=True, round_trip=True)
+
+            # Write to YAML file
+            yaml_dump(config_dict, config_file)
+
+            # Set the config file in the environment
+            os.environ["AIQ_CONFIG_FILE"] = str(config_file.name)
+
+            # Set the worker class in the environment
+            os.environ["AIQ_FRONT_END_WORKER"] = self.get_worker_class_name()
+
+            if not self.front_end_config.use_gunicorn:
+                import uvicorn
+
+                reload_excludes = ["./.*"]
+
+                uvicorn.run("aiq.front_ends.fastapi.main:get_app",
+                            host=self.front_end_config.host,
+                            port=self.front_end_config.port,
+                            workers=self.front_end_config.workers,
+                            reload=self.front_end_config.reload,
+                            factory=True,
+                            reload_excludes=reload_excludes)
+
+            else:
+                app = get_app()
+
+                from gunicorn.app.wsgiapp import WSGIApplication
+
+                class StandaloneApplication(WSGIApplication):
+
+                    def __init__(self, app, options=None):
+                        self.options = options or {}
+                        self.app = app
+                        super().__init__()
+
+                    def load_config(self):
+                        config = {
+                            key: value
+                            for key, value in self.options.items() if key in self.cfg.settings and value is not None
+                        }
+                        for key, value in config.items():
+                            self.cfg.set(key.lower(), value)
+
+                    def load(self):
+                        return self.app
+
+                options = {
+                    "bind": f"{self.front_end_config.host}:{self.front_end_config.port}",
+                    "workers": self.front_end_config.workers,
+                    "worker_class": "uvicorn.workers.UvicornWorker",
+                }
+
+                StandaloneApplication(app, options=options).run()