{description}
)}+
Invalid OpenAI API key. Verify or replace the key.
)} diff --git a/src/api/settings.py b/src/api/settings.py index 3e242c4b..c2c7cbd0 100644 --- a/src/api/settings.py +++ b/src/api/settings.py @@ -556,6 +556,19 @@ async def onboarding(request, flows_service): ) # Continue even if setting global variables fails + # Initialize the OpenSearch index now that we have the embedding model configured + try: + # Import here to avoid circular imports + from main import init_index + + logger.info("Initializing OpenSearch index after onboarding configuration") + await init_index() + logger.info("OpenSearch index initialization completed successfully") + except Exception as e: + logger.error("Failed to initialize OpenSearch index after onboarding", error=str(e)) + # Don't fail the entire onboarding process if index creation fails + # The application can still work, but document operations may fail + # Handle sample data ingestion if requested if should_ingest_sample_data: try: diff --git a/src/config/config_manager.py b/src/config/config_manager.py index 055d48a7..0b814470 100644 --- a/src/config/config_manager.py +++ b/src/config/config_manager.py @@ -16,6 +16,8 @@ class ProviderConfig: model_provider: str = "openai" # openai, anthropic, etc. api_key: str = "" + endpoint: str = "" # For providers like Watson/IBM that need custom endpoints + project_id: str = "" # For providers like Watson/IBM that need project IDs @dataclass @@ -129,6 +131,10 @@ class ConfigManager: config_data["provider"]["model_provider"] = os.getenv("MODEL_PROVIDER") if os.getenv("PROVIDER_API_KEY"): config_data["provider"]["api_key"] = os.getenv("PROVIDER_API_KEY") + if os.getenv("PROVIDER_ENDPOINT"): + config_data["provider"]["endpoint"] = os.getenv("PROVIDER_ENDPOINT") + if os.getenv("PROVIDER_PROJECT_ID"): + config_data["provider"]["project_id"] = os.getenv("PROVIDER_PROJECT_ID") # Backward compatibility for OpenAI if os.getenv("OPENAI_API_KEY"): config_data["provider"]["api_key"] = os.getenv("OPENAI_API_KEY") diff --git a/src/config/settings.py b/src/config/settings.py index 5f9b189d..3bf1e6cf 100644 --- a/src/config/settings.py +++ b/src/config/settings.py @@ -78,6 +78,31 @@ INDEX_NAME = "documents" VECTOR_DIM = 1536 EMBED_MODEL = "text-embedding-3-small" +OPENAI_EMBEDDING_DIMENSIONS = { + "text-embedding-3-small": 1536, + "text-embedding-3-large": 3072, + "text-embedding-ada-002": 1536, + } + +OLLAMA_EMBEDDING_DIMENSIONS = { + "nomic-embed-text": 768, + "all-minilm": 384, + "mxbai-embed-large": 1024, +} + +WATSONX_EMBEDDING_DIMENSIONS = { +# IBM Models +"ibm/granite-embedding-107m-multilingual": 384, +"ibm/granite-embedding-278m-multilingual": 1024, +"ibm/slate-125m-english-rtrvr": 768, +"ibm/slate-125m-english-rtrvr-v2": 768, +"ibm/slate-30m-english-rtrvr": 384, +"ibm/slate-30m-english-rtrvr-v2": 384, +# Third Party Models +"intfloat/multilingual-e5-large": 1024, +"sentence-transformers/all-minilm-l6-v2": 384, +} + INDEX_BODY = { "settings": { "index": {"knn": True}, diff --git a/src/main.py b/src/main.py index 90add401..69f2ad9f 100644 --- a/src/main.py +++ b/src/main.py @@ -2,6 +2,7 @@ from connectors.langflow_connector_service import LangflowConnectorService from connectors.service import ConnectorService from services.flows_service import FlowsService +from utils.embeddings import create_dynamic_index_body from utils.logging_config import configure_from_env, get_logger configure_from_env() @@ -52,11 +53,11 @@ from auth_middleware import optional_auth, require_auth from config.settings import ( DISABLE_INGEST_WITH_LANGFLOW, EMBED_MODEL, - INDEX_BODY, INDEX_NAME, SESSION_SECRET, clients, is_no_auth_mode, + get_openrag_config, ) from services.auth_service import AuthService from services.langflow_mcp_service import LangflowMCPService @@ -81,7 +82,6 @@ logger.info( cuda_version=torch.version.cuda, ) - async def wait_for_opensearch(): """Wait for OpenSearch to be ready with retries""" max_retries = 30 @@ -132,12 +132,19 @@ async def init_index(): """Initialize OpenSearch index and security roles""" await wait_for_opensearch() + # Get the configured embedding model from user configuration + config = get_openrag_config() + embedding_model = config.knowledge.embedding_model + + # Create dynamic index body based on the configured embedding model + dynamic_index_body = create_dynamic_index_body(embedding_model) + # Create documents index if not await clients.opensearch.indices.exists(index=INDEX_NAME): - await clients.opensearch.indices.create(index=INDEX_NAME, body=INDEX_BODY) - logger.info("Created OpenSearch index", index_name=INDEX_NAME) + await clients.opensearch.indices.create(index=INDEX_NAME, body=dynamic_index_body) + logger.info("Created OpenSearch index", index_name=INDEX_NAME, embedding_model=embedding_model) else: - logger.info("Index already exists, skipping creation", index_name=INDEX_NAME) + logger.info("Index already exists, skipping creation", index_name=INDEX_NAME, embedding_model=embedding_model) # Create knowledge filters index knowledge_filter_index_name = "knowledge_filters" @@ -391,7 +398,12 @@ async def _ingest_default_documents_openrag(services, file_paths): async def startup_tasks(services): """Startup tasks""" logger.info("Starting startup tasks") - await init_index() + # Only initialize basic OpenSearch connection, not the index + # Index will be created after onboarding when we know the embedding model + await wait_for_opensearch() + + # Configure alerting security + await configure_alerting_security() async def initialize_services(): diff --git a/src/services/flows_service.py b/src/services/flows_service.py index 0d7a7bc8..7397cf6b 100644 --- a/src/services/flows_service.py +++ b/src/services/flows_service.py @@ -1,3 +1,4 @@ +import asyncio from config.settings import ( NUDGES_FLOW_ID, LANGFLOW_URL, @@ -19,6 +20,7 @@ from config.settings import ( WATSONX_LLM_COMPONENT_ID, OLLAMA_EMBEDDING_COMPONENT_ID, OLLAMA_LLM_COMPONENT_ID, + get_openrag_config, ) import json import os @@ -29,6 +31,74 @@ logger = get_logger(__name__) class FlowsService: + def __init__(self): + # Cache for flow file mappings to avoid repeated filesystem scans + self._flow_file_cache = {} + + def _get_flows_directory(self): + """Get the flows directory path""" + current_file_dir = os.path.dirname(os.path.abspath(__file__)) # src/services/ + src_dir = os.path.dirname(current_file_dir) # src/ + project_root = os.path.dirname(src_dir) # project root + return os.path.join(project_root, "flows") + + def _find_flow_file_by_id(self, flow_id: str): + """ + Scan the flows directory and find the JSON file that contains the specified flow ID. + + Args: + flow_id: The flow ID to search for + + Returns: + str: The path to the flow file, or None if not found + """ + if not flow_id: + raise ValueError("flow_id is required") + + # Check cache first + if flow_id in self._flow_file_cache: + cached_path = self._flow_file_cache[flow_id] + if os.path.exists(cached_path): + return cached_path + else: + # Remove stale cache entry + del self._flow_file_cache[flow_id] + + flows_dir = self._get_flows_directory() + + if not os.path.exists(flows_dir): + logger.warning(f"Flows directory not found: {flows_dir}") + return None + + # Scan all JSON files in the flows directory + try: + for filename in os.listdir(flows_dir): + if not filename.endswith('.json'): + continue + + file_path = os.path.join(flows_dir, filename) + + try: + with open(file_path, 'r') as f: + flow_data = json.load(f) + + # Check if this file contains the flow we're looking for + if flow_data.get('id') == flow_id: + # Cache the result + self._flow_file_cache[flow_id] = file_path + logger.info(f"Found flow {flow_id} in file: {filename}") + return file_path + + except (json.JSONDecodeError, FileNotFoundError) as e: + logger.warning(f"Error reading flow file {filename}: {e}") + continue + + except Exception as e: + logger.error(f"Error scanning flows directory: {e}") + return None + + logger.warning(f"Flow with ID {flow_id} not found in flows directory") + return None async def reset_langflow_flow(self, flow_type: str): """Reset a Langflow flow by uploading the corresponding JSON file @@ -41,59 +111,35 @@ class FlowsService: if not LANGFLOW_URL: raise ValueError("LANGFLOW_URL environment variable is required") - # Determine flow file and ID based on type + # Determine flow ID based on type if flow_type == "nudges": - flow_file = "flows/openrag_nudges.json" flow_id = NUDGES_FLOW_ID elif flow_type == "retrieval": - flow_file = "flows/openrag_agent.json" flow_id = LANGFLOW_CHAT_FLOW_ID elif flow_type == "ingest": - flow_file = "flows/ingestion_flow.json" flow_id = LANGFLOW_INGEST_FLOW_ID else: raise ValueError( "flow_type must be either 'nudges', 'retrieval', or 'ingest'" ) + if not flow_id: + raise ValueError(f"Flow ID not configured for flow_type '{flow_type}'") + + # Dynamically find the flow file by ID + flow_path = self._find_flow_file_by_id(flow_id) + if not flow_path: + raise FileNotFoundError(f"Flow file not found for flow ID: {flow_id}") + # Load flow JSON file try: - # Get the project root directory (go up from src/services/ to project root) - # __file__ is src/services/chat_service.py - # os.path.dirname(__file__) is src/services/ - # os.path.dirname(os.path.dirname(__file__)) is src/ - # os.path.dirname(os.path.dirname(os.path.dirname(__file__))) is project root - current_file_dir = os.path.dirname( - os.path.abspath(__file__) - ) # src/services/ - src_dir = os.path.dirname(current_file_dir) # src/ - project_root = os.path.dirname(src_dir) # project root - flow_path = os.path.join(project_root, flow_file) - - if not os.path.exists(flow_path): - # List contents of project root to help debug - try: - contents = os.listdir(project_root) - logger.info(f"Project root contents: {contents}") - - flows_dir = os.path.join(project_root, "flows") - if os.path.exists(flows_dir): - flows_contents = os.listdir(flows_dir) - logger.info(f"Flows directory contents: {flows_contents}") - else: - logger.info("Flows directory does not exist") - except Exception as e: - logger.error(f"Error listing directory contents: {e}") - - raise FileNotFoundError(f"Flow file not found at: {flow_path}") - with open(flow_path, "r") as f: flow_data = json.load(f) - logger.info(f"Successfully loaded flow data from {flow_file}") + logger.info(f"Successfully loaded flow data for {flow_type} from {os.path.basename(flow_path)}") + except json.JSONDecodeError as e: + raise ValueError(f"Invalid JSON in flow file {flow_path}: {e}") except FileNotFoundError: raise ValueError(f"Flow file not found: {flow_path}") - except json.JSONDecodeError as e: - raise ValueError(f"Invalid JSON in flow file {flow_file}: {e}") # Make PATCH request to Langflow API to update the flow using shared client try: @@ -106,8 +152,54 @@ class FlowsService: logger.info( f"Successfully reset {flow_type} flow", flow_id=flow_id, - flow_file=flow_file, + flow_file=os.path.basename(flow_path), ) + + # Now update the flow with current configuration settings + try: + config = get_openrag_config() + + # Check if configuration has been edited (onboarding completed) + if config.edited: + logger.info(f"Updating {flow_type} flow with current configuration settings") + + provider = config.provider.model_provider.lower() + + # Step 1: Assign model provider (replace components) if not OpenAI + if provider != "openai": + logger.info(f"Assigning {provider} components to {flow_type} flow") + provider_result = await self.assign_model_provider(provider) + + if not provider_result.get("success"): + logger.warning(f"Failed to assign {provider} components: {provider_result.get('error', 'Unknown error')}") + # Continue anyway, maybe just value updates will work + + # Step 2: Update model values for the specific flow being reset + single_flow_config = [{ + "name": flow_type, + "flow_id": flow_id, + }] + + logger.info(f"Updating {flow_type} flow model values") + update_result = await self.change_langflow_model_value( + provider=provider, + embedding_model=config.knowledge.embedding_model, + llm_model=config.agent.llm_model, + endpoint=config.provider.endpoint if config.provider.endpoint else None, + flow_configs=single_flow_config + ) + + if update_result.get("success"): + logger.info(f"Successfully updated {flow_type} flow with current configuration") + else: + logger.warning(f"Failed to update {flow_type} flow with current configuration: {update_result.get('error', 'Unknown error')}") + else: + logger.info(f"Configuration not yet edited (onboarding not completed), skipping model updates for {flow_type} flow") + + except Exception as e: + logger.error(f"Error updating {flow_type} flow with current configuration", error=str(e)) + # Don't fail the entire reset operation if configuration update fails + return { "success": True, "message": f"Successfully reset {flow_type} flow", @@ -155,11 +247,10 @@ class FlowsService: logger.info(f"Assigning {provider} components") - # Define flow configurations + # Define flow configurations (removed hardcoded file paths) flow_configs = [ { "name": "nudges", - "file": "flows/openrag_nudges.json", "flow_id": NUDGES_FLOW_ID, "embedding_id": OPENAI_EMBEDDING_COMPONENT_ID, "llm_id": OPENAI_LLM_COMPONENT_ID, @@ -167,7 +258,6 @@ class FlowsService: }, { "name": "retrieval", - "file": "flows/openrag_agent.json", "flow_id": LANGFLOW_CHAT_FLOW_ID, "embedding_id": OPENAI_EMBEDDING_COMPONENT_ID, "llm_id": OPENAI_LLM_COMPONENT_ID, @@ -175,7 +265,6 @@ class FlowsService: }, { "name": "ingest", - "file": "flows/ingestion_flow.json", "flow_id": LANGFLOW_INGEST_FLOW_ID, "embedding_id": OPENAI_EMBEDDING_COMPONENT_ID, "llm_id": None, # Ingestion flow might not have LLM @@ -272,7 +361,6 @@ class FlowsService: async def _update_flow_components(self, config, llm_template, embedding_template, llm_text_template): """Update components in a specific flow""" flow_name = config["name"] - flow_file = config["file"] flow_id = config["flow_id"] old_embedding_id = config["embedding_id"] old_llm_id = config["llm_id"] @@ -281,14 +369,11 @@ class FlowsService: new_llm_id = llm_template["data"]["id"] new_embedding_id = embedding_template["data"]["id"] new_llm_text_id = llm_text_template["data"]["id"] - # Get the project root directory - current_file_dir = os.path.dirname(os.path.abspath(__file__)) - src_dir = os.path.dirname(current_file_dir) - project_root = os.path.dirname(src_dir) - flow_path = os.path.join(project_root, flow_file) - if not os.path.exists(flow_path): - raise FileNotFoundError(f"Flow file not found at: {flow_path}") + # Dynamically find the flow file by ID + flow_path = self._find_flow_file_by_id(flow_id) + if not flow_path: + raise FileNotFoundError(f"Flow file not found for flow ID: {flow_id}") # Load flow JSON with open(flow_path, "r") as f: @@ -527,16 +612,17 @@ class FlowsService: return False async def change_langflow_model_value( - self, provider: str, embedding_model: str, llm_model: str, endpoint: str = None + self, provider: str, embedding_model: str, llm_model: str, endpoint: str = None, flow_configs: list = None ): """ - Change dropdown values for provider-specific components across all flows + Change dropdown values for provider-specific components across flows Args: provider: The provider ("watsonx", "ollama", "openai") embedding_model: The embedding model name to set llm_model: The LLM model name to set endpoint: The endpoint URL (required for watsonx/ibm provider) + flow_configs: Optional list of specific flow configs to update. If None, updates all flows. Returns: dict: Success/error response with details for each flow @@ -552,24 +638,22 @@ class FlowsService: f"Changing dropdown values for provider {provider}, embedding: {embedding_model}, llm: {llm_model}, endpoint: {endpoint}" ) - # Define flow configurations with provider-specific component IDs - flow_configs = [ - { - "name": "nudges", - "file": "flows/openrag_nudges.json", - "flow_id": NUDGES_FLOW_ID, - }, - { - "name": "retrieval", - "file": "flows/openrag_agent.json", - "flow_id": LANGFLOW_CHAT_FLOW_ID, - }, - { - "name": "ingest", - "file": "flows/ingestion_flow.json", - "flow_id": LANGFLOW_INGEST_FLOW_ID, - }, - ] + # Use provided flow_configs or default to all flows + if flow_configs is None: + flow_configs = [ + { + "name": "nudges", + "flow_id": NUDGES_FLOW_ID, + }, + { + "name": "retrieval", + "flow_id": LANGFLOW_CHAT_FLOW_ID, + }, + { + "name": "ingest", + "flow_id": LANGFLOW_INGEST_FLOW_ID, + }, + ] # Determine target component IDs based on provider target_embedding_id, target_llm_id, target_llm_text_id = self._get_provider_component_ids( diff --git a/src/utils/embeddings.py b/src/utils/embeddings.py new file mode 100644 index 00000000..f3c902e7 --- /dev/null +++ b/src/utils/embeddings.py @@ -0,0 +1,64 @@ +from config.settings import OLLAMA_EMBEDDING_DIMENSIONS, OPENAI_EMBEDDING_DIMENSIONS, VECTOR_DIM, WATSONX_EMBEDDING_DIMENSIONS +from utils.logging_config import get_logger + + +logger = get_logger(__name__) + +def get_embedding_dimensions(model_name: str) -> int: + """Get the embedding dimensions for a given model name.""" + + # Check all model dictionaries + all_models = {**OPENAI_EMBEDDING_DIMENSIONS, **OLLAMA_EMBEDDING_DIMENSIONS, **WATSONX_EMBEDDING_DIMENSIONS} + + if model_name in all_models: + dimensions = all_models[model_name] + logger.info(f"Found dimensions for model '{model_name}': {dimensions}") + return dimensions + + logger.warning( + f"Unknown embedding model '{model_name}', using default dimensions: {VECTOR_DIM}" + ) + return VECTOR_DIM + + +def create_dynamic_index_body(embedding_model: str) -> dict: + """Create a dynamic index body configuration based on the embedding model.""" + dimensions = get_embedding_dimensions(embedding_model) + + return { + "settings": { + "index": {"knn": True}, + "number_of_shards": 1, + "number_of_replicas": 1, + }, + "mappings": { + "properties": { + "document_id": {"type": "keyword"}, + "filename": {"type": "keyword"}, + "mimetype": {"type": "keyword"}, + "page": {"type": "integer"}, + "text": {"type": "text"}, + "chunk_embedding": { + "type": "knn_vector", + "dimension": dimensions, + "method": { + "name": "disk_ann", + "engine": "jvector", + "space_type": "l2", + "parameters": {"ef_construction": 100, "m": 16}, + }, + }, + "source_url": {"type": "keyword"}, + "connector_type": {"type": "keyword"}, + "owner": {"type": "keyword"}, + "allowed_users": {"type": "keyword"}, + "allowed_groups": {"type": "keyword"}, + "user_permissions": {"type": "object"}, + "group_permissions": {"type": "object"}, + "created_time": {"type": "date"}, + "modified_time": {"type": "date"}, + "indexed_time": {"type": "date"}, + "metadata": {"type": "object"}, + } + }, + } \ No newline at end of file From a67aaaacd601c54198a43c169cc27396317b8f2f Mon Sep 17 00:00:00 2001 From: Mendon Kissling <59585235+mendonk@users.noreply.github.com> Date: Fri, 26 Sep 2025 12:14:42 -0400 Subject: [PATCH 09/10] remove-multiple-map --- docs/docs/core-components/knowledge.mdx | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/docs/docs/core-components/knowledge.mdx b/docs/docs/core-components/knowledge.mdx index d8160f8b..b5507eef 100644 --- a/docs/docs/core-components/knowledge.mdx +++ b/docs/docs/core-components/knowledge.mdx @@ -30,7 +30,7 @@ OpenRAG supports knowledge ingestion through direct file uploads and OAuth conne The **Knowledge Ingest** flow uses Langflow's [**File** component](https://docs.langflow.org/components-data#file) to split and embed files loaded from your local machine into the OpenSearch database. -The default path to your local folder is mounted from the `./documents` folder in your OpenRAG project directory to the `/app/documents/` directory inside the Docker container. Files added to the host or the container will be visible in both locations. To configure this location, modify the **Documents Paths** variable in either the TUI's [Advanced Setup](/install#advanced-setup) or in the `.env` used by Docker Compose. Add multiple paths in a comma-separated list with no spaces. For example, `./documents,/Users/username/Documents`. +The default path to your local folder is mounted from the `./documents` folder in your OpenRAG project directory to the `/app/documents/` directory inside the Docker container. Files added to the host or the container will be visible in both locations. To configure this location, modify the **Documents Paths** variable in either the TUI's [Advanced Setup](/install#advanced-setup) or in the `.env` used by Docker Compose. To load and process a single file from the mapped location, click **Add Knowledge**, and then click **Add File**. The file is loaded into your OpenSearch database, and appears in the Knowledge page. From f424e08801b34054f82d1b3586e02d4f2872d3fe Mon Sep 17 00:00:00 2001 From: Mendon Kissling <59585235+mendonk@users.noreply.github.com> Date: Fri, 26 Sep 2025 13:35:42 -0400 Subject: [PATCH 10/10] no-static-default-onboarding-behavior --- docs/docs/core-components/knowledge.mdx | 13 ++++++------- 1 file changed, 6 insertions(+), 7 deletions(-) diff --git a/docs/docs/core-components/knowledge.mdx b/docs/docs/core-components/knowledge.mdx index b5507eef..255b0b68 100644 --- a/docs/docs/core-components/knowledge.mdx +++ b/docs/docs/core-components/knowledge.mdx @@ -134,15 +134,14 @@ A new filter is created with default settings that match everything. ## OpenRAG default configuration -OpenRAG creates a specialized OpenSearch index called `documents` with the values defined at `src/config/settings.py`. -* **Vector Dimensions**: 1536-dimensional embeddings using OpenAI's `text-embedding-3-small` model. -* **KNN Vector Type**: Uses `knn_vector` field type with `disk_ann` method and `jvector` engine. -* **Distance Metric**: L2 (Euclidean) distance for vector similarity. -* **Performance Optimization**: Configured with `ef_construction: 100` and `m: 16` parameters. - -OpenRAG supports hybrid search, which combines semantic and keyword search. +OpenRAG automatically detects and configures the correct vector dimensions for embedding models, ensuring optimal search performance and compatibility. +The complete list of supported models is available at [/src/services/models_service.py](https://github.com/langflow-ai/openrag/blob/main/src/services/models_service.py). +You can use custom embedding models by specifying them in your configuration. +If you use an unknown embedding model, OpenRAG will automatically fall back to `1536` dimensions and log a warning. The system will continue to work, but search quality may be affected if the actual model dimensions differ from `1536`. +The default embedding dimension is `1536` and the default model is `text-embedding-3-small`. +For models with known vector dimensions, see [/src/config/settings.py](https://github.com/langflow-ai/openrag/blob/main/src/config/settings.py). \ No newline at end of file