lightspeed-stack 0.1.1__tar.gz → 0.1.3__tar.gz

{lightspeed_stack-0.1.1 → lightspeed_stack-0.1.3}/PKG-INFO

@@ -1,7 +1,9 @@
 Metadata-Version: 2.1
 Name: lightspeed-stack
-Version: 0.1.1
+Version: 0.1.3
 Summary: LLM tooling stack
+Keywords: LLM,RAG
+Maintainer-Email: =?utf-8?b?UGF2ZWwgVGnFoW5vdnNrw70=?= <tisnik@centrum.cz>
 License:                                  Apache License
                                     Version 2.0, January 2004
                                  http://www.apache.org/licenses/
@@ -204,6 +206,12 @@ License:                                  Apache License
             See the License for the specific language governing permissions and
             limitations under the License.
          
+Classifier: Development Status :: 4 - Beta
+Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence
+Classifier: Topic :: Software Development
+Classifier: Programming Language :: Python :: 3
+Classifier: Programming Language :: Python :: 3 :: Only
+Classifier: License :: OSI Approved :: Apache Software License
 Project-URL: Homepage, https://github.com/lightspeed-core/lightspeed-stack
 Project-URL: Issues, https://github.com/lightspeed-core/lightspeed-stack/issues
 Requires-Python: <3.14,>=3.12
@@ -228,7 +236,7 @@ Description-Content-Type: text/markdown
 [![License](https://img.shields.io/badge/license-Apache-blue)](https://github.com/lightspeed-core/lightspeed-stack/blob/main/LICENSE)
 [![made-with-python](https://img.shields.io/badge/Made%20with-Python-1f425f.svg)](https://www.python.org/)
 [![Required Python version](https://img.shields.io/python/required-version-toml?tomlFilePath=https%3A%2F%2Fraw.githubusercontent.com%2Flightspeed-core%2Flightspeed-stack%2Frefs%2Fheads%2Fmain%2Fpyproject.toml)](https://www.python.org/)
-[![Tag](https://img.shields.io/github/v/tag/lightspeed-core/lightspeed-stack)](https://github.com/lightspeed-core/lightspeed-stack/releases/tag/0.1.1)
+[![Tag](https://img.shields.io/github/v/tag/lightspeed-core/lightspeed-stack)](https://github.com/lightspeed-core/lightspeed-stack/releases/tag/0.1.3)
 
 Lightspeed Core Stack (LCS) is an AI-powered assistant that provides answers to product questions using backend LLM services, agents, and RAG databases.
 

{lightspeed_stack-0.1.1 → lightspeed_stack-0.1.3}/README.md

@@ -6,7 +6,7 @@
 [![License](https://img.shields.io/badge/license-Apache-blue)](https://github.com/lightspeed-core/lightspeed-stack/blob/main/LICENSE)
 [![made-with-python](https://img.shields.io/badge/Made%20with-Python-1f425f.svg)](https://www.python.org/)
 [![Required Python version](https://img.shields.io/python/required-version-toml?tomlFilePath=https%3A%2F%2Fraw.githubusercontent.com%2Flightspeed-core%2Flightspeed-stack%2Frefs%2Fheads%2Fmain%2Fpyproject.toml)](https://www.python.org/)
-[![Tag](https://img.shields.io/github/v/tag/lightspeed-core/lightspeed-stack)](https://github.com/lightspeed-core/lightspeed-stack/releases/tag/0.1.1)
+[![Tag](https://img.shields.io/github/v/tag/lightspeed-core/lightspeed-stack)](https://github.com/lightspeed-core/lightspeed-stack/releases/tag/0.1.3)
 
 Lightspeed Core Stack (LCS) is an AI-powered assistant that provides answers to product questions using backend LLM services, agents, and RAG databases.
 

{lightspeed_stack-0.1.1 → lightspeed_stack-0.1.3}/pyproject.toml

@@ -3,6 +3,21 @@ name = "lightspeed-stack"
 dynamic = []
 description = "LLM tooling stack"
 authors = []
+maintainers = [
+    { email = "tisnik@centrum.cz", name = "Pavel Tišnovský" },
+]
+classifiers = [
+    "Development Status :: 4 - Beta",
+    "Topic :: Scientific/Engineering :: Artificial Intelligence",
+    "Topic :: Software Development",
+    "Programming Language :: Python :: 3",
+    "Programming Language :: Python :: 3 :: Only",
+    "License :: OSI Approved :: Apache Software License",
+]
+keywords = [
+    "LLM",
+    "RAG",
+]
 requires-python = ">=3.12,<3.14"
 readme = "README.md"
 dependencies = [
@@ -18,7 +33,7 @@ dependencies = [
     "aiohttp>=3.12.14",
     "authlib>=1.6.0",
 ]
-version = "0.1.1"
+version = "0.1.3"
 
 [project.license]
 file = "LICENSE"
@@ -77,6 +92,7 @@ dev = [
     "behave>=1.2.6",
     "types-cachetools>=6.1.0.20250717",
     "build>=1.2.2.post1",
+    "twine>=6.1.0",
 ]
 build = [
     "build>=1.2.2.post1",

{lightspeed_stack-0.1.1 → lightspeed_stack-0.1.3}/src/app/endpoints/conversations.py

@@ -23,11 +23,16 @@ conversation_id_to_agent_id: dict[str, str] = {}
 conversation_responses: dict[int | str, dict[str, Any]] = {
     200: {
         "conversation_id": "123e4567-e89b-12d3-a456-426614174000",
-        "session_data": {
-            "session_id": "123e4567-e89b-12d3-a456-426614174000",
-            "turns": [],
-            "started_at": "2024-01-01T00:00:00Z",
-        },
+        "chat_history": [
+            {
+                "messages": [
+                    {"content": "Hi", "type": "user"},
+                    {"content": "Hello!", "type": "assistant"},
+                ],
+                "started_at": "2024-01-01T00:00:00Z",
+                "completed_at": "2024-01-01T00:00:05Z",
+            }
+        ],
     },
     404: {
         "detail": {

{lightspeed_stack-0.1.1 → lightspeed_stack-0.1.3}/src/app/endpoints/health.py

@@ -11,7 +11,7 @@ from typing import Any
 from llama_stack.providers.datatypes import HealthStatus
 
 from fastapi import APIRouter, status, Response
-from client import LlamaStackClientHolder
+from client import AsyncLlamaStackClientHolder
 from models.responses import (
     LivenessResponse,
     ReadinessResponse,
@@ -22,16 +22,17 @@ logger = logging.getLogger(__name__)
 router = APIRouter(tags=["health"])
 
 
-def get_providers_health_statuses() -> list[ProviderHealthStatus]:
+async def get_providers_health_statuses() -> list[ProviderHealthStatus]:
     """Check health of all providers.
 
     Returns:
         List of provider health statuses.
     """
     try:
-        client = LlamaStackClientHolder().get_client()
+        client = AsyncLlamaStackClientHolder().get_client()
 
-        providers = client.providers.list()
+        # providers = []
+        providers = await client.providers.list()
         logger.debug("Found %d providers", len(providers))
 
         health_results = [
@@ -69,9 +70,9 @@ get_readiness_responses: dict[int | str, dict[str, Any]] = {
 
 
 @router.get("/readiness", responses=get_readiness_responses)
-def readiness_probe_get_method(response: Response) -> ReadinessResponse:
+async def readiness_probe_get_method(response: Response) -> ReadinessResponse:
     """Ready status of service with provider health details."""
-    provider_statuses = get_providers_health_statuses()
+    provider_statuses = await get_providers_health_statuses()
 
     # Check if any provider is unhealthy (not counting not_implemented as unhealthy)
     unhealthy_providers = [

{lightspeed_stack-0.1.1 → lightspeed_stack-0.1.3}/src/app/endpoints/query.py

@@ -80,6 +80,7 @@ def get_agent(  # pylint: disable=too-many-arguments,too-many-positional-argumen
     available_input_shields: list[str],
     available_output_shields: list[str],
     conversation_id: str | None,
+    no_tools: bool = False,
 ) -> tuple[Agent, str]:
     """Get existing agent or create a new one with session persistence."""
     if conversation_id is not None:
@@ -99,7 +100,7 @@ def get_agent(  # pylint: disable=too-many-arguments,too-many-positional-argumen
         instructions=system_prompt,
         input_shields=available_input_shields if available_input_shields else [],
         output_shields=available_output_shields if available_output_shields else [],
-        tool_parser=GraniteToolParser.get_parser(model_id),
+        tool_parser=None if no_tools else GraniteToolParser.get_parser(model_id),
         enable_session_persistence=True,
     )
     conversation_id = agent.create_session(get_suid())
@@ -216,10 +217,12 @@ def select_model_and_provider_id(
                 },
             ) from e
 
+    llama_stack_model_id = f"{provider_id}/{model_id}"
     # Validate that the model_id and provider_id are in the available models
     logger.debug("Searching for model: %s, provider: %s", model_id, provider_id)
     if not any(
-        m.identifier == model_id and m.provider_id == provider_id for m in models
+        m.identifier == llama_stack_model_id and m.provider_id == provider_id
+        for m in models
     ):
         message = f"Model {model_id} from provider {provider_id} not found in available models"
         logger.error(message)
@@ -231,7 +234,7 @@ def select_model_and_provider_id(
             },
         )
 
-    return model_id, provider_id
+    return llama_stack_model_id, provider_id
 
 
 def _is_inout_shield(shield: Shield) -> bool:
@@ -286,36 +289,47 @@ def retrieve_response(  # pylint: disable=too-many-locals
         available_input_shields,
         available_output_shields,
         query_request.conversation_id,
+        query_request.no_tools or False,
     )
 
-    # preserve compatibility when mcp_headers is not provided
-    if mcp_headers is None:
+    # bypass tools and MCP servers if no_tools is True
+    if query_request.no_tools:
         mcp_headers = {}
-    mcp_headers = handle_mcp_headers_with_toolgroups(mcp_headers, configuration)
-    if not mcp_headers and token:
-        for mcp_server in configuration.mcp_servers:
-            mcp_headers[mcp_server.url] = {
-                "Authorization": f"Bearer {token}",
-            }
-
-    agent.extra_headers = {
-        "X-LlamaStack-Provider-Data": json.dumps(
-            {
-                "mcp_headers": mcp_headers,
-            }
-        ),
-    }
+        agent.extra_headers = {}
+        toolgroups = None
+    else:
+        # preserve compatibility when mcp_headers is not provided
+        if mcp_headers is None:
+            mcp_headers = {}
+        mcp_headers = handle_mcp_headers_with_toolgroups(mcp_headers, configuration)
+        if not mcp_headers and token:
+            for mcp_server in configuration.mcp_servers:
+                mcp_headers[mcp_server.url] = {
+                    "Authorization": f"Bearer {token}",
+                }
+
+        agent.extra_headers = {
+            "X-LlamaStack-Provider-Data": json.dumps(
+                {
+                    "mcp_headers": mcp_headers,
+                }
+            ),
+        }
+
+        vector_db_ids = [vector_db.identifier for vector_db in client.vector_dbs.list()]
+        toolgroups = (get_rag_toolgroups(vector_db_ids) or []) + [
+            mcp_server.name for mcp_server in configuration.mcp_servers
+        ]
+        # Convert empty list to None for consistency with existing behavior
+        if not toolgroups:
+            toolgroups = None
 
-    vector_db_ids = [vector_db.identifier for vector_db in client.vector_dbs.list()]
-    toolgroups = (get_rag_toolgroups(vector_db_ids) or []) + [
-        mcp_server.name for mcp_server in configuration.mcp_servers
-    ]
     response = agent.create_turn(
         messages=[UserMessage(role="user", content=query_request.query)],
         session_id=conversation_id,
         documents=query_request.get_documents(),
         stream=False,
-        toolgroups=toolgroups or None,
+        toolgroups=toolgroups,
     )
 
     # Check for validation errors in the response

{lightspeed_stack-0.1.1 → lightspeed_stack-0.1.3}/src/app/endpoints/streaming_query.py

@@ -58,6 +58,7 @@ async def get_agent(
     available_input_shields: list[str],
     available_output_shields: list[str],
     conversation_id: str | None,
+    no_tools: bool = False,
 ) -> tuple[AsyncAgent, str]:
     """Get existing agent or create a new one with session persistence."""
     if conversation_id is not None:
@@ -76,7 +77,7 @@ async def get_agent(
         instructions=system_prompt,
         input_shields=available_input_shields if available_input_shields else [],
         output_shields=available_output_shields if available_output_shields else [],
-        tool_parser=GraniteToolParser.get_parser(model_id),
+        tool_parser=None if no_tools else GraniteToolParser.get_parser(model_id),
         enable_session_persistence=True,
     )
     conversation_id = await agent.create_session(get_suid())
@@ -532,41 +533,53 @@ async def retrieve_response(
         available_input_shields,
         available_output_shields,
         query_request.conversation_id,
+        query_request.no_tools or False,
     )
 
-    # preserve compatibility when mcp_headers is not provided
-    if mcp_headers is None:
+    # bypass tools and MCP servers if no_tools is True
+    if query_request.no_tools:
         mcp_headers = {}
+        agent.extra_headers = {}
+        toolgroups = None
+    else:
+        # preserve compatibility when mcp_headers is not provided
+        if mcp_headers is None:
+            mcp_headers = {}
 
-    mcp_headers = handle_mcp_headers_with_toolgroups(mcp_headers, configuration)
+        mcp_headers = handle_mcp_headers_with_toolgroups(mcp_headers, configuration)
 
-    if not mcp_headers and token:
-        for mcp_server in configuration.mcp_servers:
-            mcp_headers[mcp_server.url] = {
-                "Authorization": f"Bearer {token}",
-            }
+        if not mcp_headers and token:
+            for mcp_server in configuration.mcp_servers:
+                mcp_headers[mcp_server.url] = {
+                    "Authorization": f"Bearer {token}",
+                }
 
-    agent.extra_headers = {
-        "X-LlamaStack-Provider-Data": json.dumps(
-            {
-                "mcp_headers": mcp_headers,
-            }
-        ),
-    }
+        agent.extra_headers = {
+            "X-LlamaStack-Provider-Data": json.dumps(
+                {
+                    "mcp_headers": mcp_headers,
+                }
+            ),
+        }
+
+        logger.debug("Session ID: %s", conversation_id)
+        vector_db_ids = [
+            vector_db.identifier for vector_db in await client.vector_dbs.list()
+        ]
+        toolgroups = (get_rag_toolgroups(vector_db_ids) or []) + [
+            mcp_server.name for mcp_server in configuration.mcp_servers
+        ]
+        # Convert empty list to None for consistency with existing behavior
+        if not toolgroups:
+            toolgroups = None
 
     logger.debug("Session ID: %s", conversation_id)
-    vector_db_ids = [
-        vector_db.identifier for vector_db in await client.vector_dbs.list()
-    ]
-    toolgroups = (get_rag_toolgroups(vector_db_ids) or []) + [
-        mcp_server.name for mcp_server in configuration.mcp_servers
-    ]
     response = await agent.create_turn(
         messages=[UserMessage(role="user", content=query_request.query)],
         session_id=conversation_id,
         documents=query_request.get_documents(),
         stream=True,
-        toolgroups=toolgroups or None,
+        toolgroups=toolgroups,
     )
 
     return response, conversation_id

{lightspeed_stack-0.1.1 → lightspeed_stack-0.1.3}/src/models/requests.py

@@ -69,6 +69,7 @@ class QueryRequest(BaseModel):
         model: The optional model.
         system_prompt: The optional system prompt.
         attachments: The optional attachments.
+        no_tools: Whether to bypass all tools and MCP servers (default: False).
 
     Example:
         ```python
@@ -82,6 +83,7 @@ class QueryRequest(BaseModel):
     model: Optional[str] = None
     system_prompt: Optional[str] = None
     attachments: Optional[list[Attachment]] = None
+    no_tools: Optional[bool] = False
     # media_type is not used in 'lightspeed-stack' that only supports application/json.
     # the field is kept here to enable compatibility with 'road-core' clients.
     media_type: Optional[str] = None
@@ -97,6 +99,7 @@ class QueryRequest(BaseModel):
                     "provider": "openai",
                     "model": "model-name",
                     "system_prompt": "You are a helpful assistant",
+                    "no_tools": False,
                     "attachments": [
                         {
                             "attachment_type": "log",