Teradata MCP Server

# -------------------------------------------------------------------------------# # Copyright (C) 2025 by Teradata Corporation. # # All Rights Reserved. # # # # File: tdvs_prompts.yml # # # # Description: # # YAML configuration file containing prompt templates for teradata enterprise # # vector store operations. # # -------------------------------------------------------------------------------# tool_descriptions: tdvs_get_health: "Use this tool to check and return the health or overall status of Teradata Vector Store." tdvs_list: "Use this tool to get the list of all vector stores with their details in Teradata Vector Store. Return the complete dataframe and do not trim any entry." tdvs_get_details: "Use this tool to get details of a specific vector store in Teradata Vector Store" tdvs_destroy: "Use this tool to destroy/delete a vector store in Teradata Vector Store." tdvs_grant_user_permission: "Use this tool to grant user permission to a vector store in Teradata Vector Store. Allowed values for permission are 'ADMIN' or 'USER'." tdvs_revoke_user_permission: "Use this tool to revoke user permission to a vector store in Teradata Vector Store. Allowed values for permission are 'ADMIN' or 'USER'." tdvs_similarity_search: | Use this tool to perform similarity search and retrieve top K similar documents from a vector store in Teradata Vector Store. TOOL PARAMETERS: ----------------- vs_name: str = Field(..., description="Specifies the name of the vector store to be used.") class VectorStoreSimilaritySearch(BaseModel): question: str = Field(..., description="Specifies a string of text for which similarity search needs to be performed.") batch_data: Optional[str] = Field(None, description="Optional Specifies the table name or teradataml DataFrame to be indexed for batch mode") batch_id_column: Optional[str] = Field(None, description="Optional Specifies the ID column to be indexed for batch mode") batch_query_column: Optional[str] = Field(None, description="Optional Specifies the query column to be indexed for batch mode.") tdvs_ask: | Use this tool to find contextual information related to query from specific vector store and return the final reponse. This tool must always be used to retrieve information irrespective of topic of the user_query. TOOL PARAMETERS: ----------------- vs_name: str = Field(..., description="Specifies the name of the vector store to be used.") class VectorStoreAsk(BaseModel): question: str = Field(..., description="The question to ask the VectorStore.") prompt: Optional[str] = Field(None, description="Optional prompt to guide the response.") batch_data: Optional[str] = Field(None, description="Optional Specifies the table name or teradataml DataFrame to be indexed for batch mode") batch_id_column: Optional[str] = Field(None, description="Optional Specifies the ID column to be indexed for batch mode") batch_query_column: Optional[str] = Field(None, description="Optional Specifies the query column to be indexed for batch mode.") tdvs_create: | Use this tool to create a vector store in Teradata Vector Store. TOOL PARAMETERS: ----------------- vs_name: str = Field(..., description="Specifies the name of the vector store to be used.") class VectorStoreCreate(BaseModel): description: str = Field(..., description="Specifies the description of the VectorStore.") target_database: Optional[str] = Field(None, description="Specifies the target database where the VectorStore will be created.") object_names: str = Field(..., description="Specifies the table name(s)/teradataml DataFrame(s) to be indexed for vector store.") key_columns: List[str] = Field(None, description="Optional Specifies the name(s) of the key column(s) to be used for indexing.") data_columns: List[str] = Field(None, description="Optional Specifies the name(s) of the data column(s) to be used for embedding generation(vectorization).") vector_column: Optional[str] = Field(None, description="Specifies the name of the column where the vectorized data will be stored.") chunk_size: Optional[int] = Field(None, description="Optional Specifies the size of each chunk when dividing document files into chunks.") optimized_chunking: Optional[bool] = Field(None, description="Optional Specifies whether an optimized splitting mechanism supplied by Teradata should be used.") header_height: Optional[int] = Field(None, description="Optional Specifies the height of the header in the document file.") footer_height: Optional[int] = Field(None, description="Optional Specifies the height of the footer in the document file.") embeddings_model: str = Field(None, description="Optional Specifies the embedding model to be used for vectorization.") embeddings_dims: Optional[int] = Field(None, description="Optional Specifies the number of dimensions for the embeddings.") metric: Optional[str] = Field(None, description="Optional Specifies the metric to be used for calculating the distance between the vectors.") search_algorithm: Optional[str] = Field(None, description="Optional Specifies the search algorithm to be used for similarity search.") initial_centroids_method: Optional[str] = Field(None, description="Optional Specifies the method to be used for initializing centroids.") train_numclusters: Optional[int] = Field(None, description="Optional Specifies the number of clusters to be used for training.") max_iternum: Optional[int] = Field(None, description="Optional Specifies the maximum number of iterations for training.") stop_threshold: Optional[float] = Field(None, description="Optional Specifies the threshold for stopping the training process.") seed: Optional[int] = Field(None, description="Optional Specifies the seed value for random number generation.") num_init: Optional[int] = Field(None, description="Optional Specifies the number of initializations to be performed for k-means.") top_k: Optional[int] = Field(None, description="Optional Specifies the number of top results to be returned for similarity search.") search_threshold: Optional[float] = Field(None, description="OptionalSpecifies the threshold value to consider for matching tables/views while searching.") search_numcluster: Optional[int] = Field(None, description="Optional Specifies the number of clusters to be used for searching.") prompt: Optional[str] = Field(None, description="Optional Specifies the prompt to be used by language model to generate responses using top matches.") chat_completion_model: Optional[str] = Field(None, description="Optional Specifies the chat completion model to be used for generating responses.") document_files: Optional[List[str]] = Field(None, description="Optional Specifies the list of document files to be indexed for vector store.") ef_search: Optional[int] = Field(None, description="Optional Specifies the number of neighbors to be considered for hnsw algorithm during similarity search.") num_layers: Optional[int] = Field(None, description="Optional Specifies the maximum number of layers to be used for hnsw algorithm during vector store creation.") ef_construction: Optional[int] = Field(None, description="Optional Specifies the number of neighbors to be considered for hnsw algorithm during vector store creation.") num_connpernode: Optional[int] = Field(None, description="Optional Specifies the number of connections per node to be used for hnsw algorithm during vector store creation.") maxnum_connpernode: Optional[int] = Field(None, description="Optional Specifies the maximum number of connections per node to be used for hnsw algorithm during vector store creation.") apply_heuristics: Optional[bool] = Field(None, description="Optional Specifies whether to apply heuristics for hnsw algorithm during vector store creation.") include_objects: Optional[List[str]] = Field(None, description="Optional Specifies the list of tables/views included in the metadata based vector store.") exclude_objects: Optional[List[str]] = Field(None, description="Optional Specifies the list of tables/views excluded from the metadata based vector store.") sample_size: Optional[int] = Field(None, description="Optional Specifies the number of rows to sample tables/views for the metadata based vector store embeddings.") rerank_weight: Optional[float] = Field(None, description="Optional Specifies the weight to be used for reranking the search results.") relevance_top_k: Optional[int] = Field(None, description="Optional Specifies the number of top similarity matches to be considered for reranking.") relevance_search_threshold: Optional[float] = Field(None, description="Optional Specifies the threshold value to be consider matching tables/views while reranking.") include_patterns: Optional[List[str]] = Field(None, description="Optional Specifies the list of patterns to be included in the metadata based vector store.") exclude_patterns: Optional[List[str]] = Field(None, description="Optional Specifies the list of patterns to be excluded from the metadata based vector store.") batch: Optional[bool] = Field(None, description="Optional Specifies whether to use batch processing for embedding generation. Applicable only for AWS.") ignore_embedding_errors: Optional[bool] = Field(None, description="Optional Specifies whether to ignore embedding errors during embedding generation. Applicable only for AWS.") chat_completion_max_tokens: Optional[int] = Field(None, description="Optional Specifies the maximum number of tokens to be generated by chat completion model.") embeddings_base_url: Optional[str] = Field(None, description="Optional Specifies the base URL for the service to be used for embeddings.") completions_base_url: Optional[str] = Field(None, description="Optional Specifies the base URL for the service to be used for completions.") ranking_url: Optional[str] = Field(None, description="Optional Specifies the URL for the service to be used for reranking.") ingest_host: Optional[str] = Field(None, description="Optional Specifies the http host for document parsing.") ingest_port: Optional[int] = Field(None, description="Optional Specifies the port for document parsing.") tdvs_update: | Use this tool to update an existing vector store with the specified parameters in Teradata Vector Store. TOOL PARAMETERS: ----------------- vs_name: str = Field(..., description="Specifies the name of the vector store to be used.") class VectorStoreUpdate(BaseModel): description: str = Field(..., description="Specifies the description of the VectorStore.") target_database: Optional[str] = Field(None, description="Specifies the target database where the VectorStore will be created.") object_names: str = Field(..., description="Specifies the table name(s)/teradataml DataFrame(s) to be indexed for vector store.") alter_operation: Literal["ADD", "DELETE"] = Field(..., description="Optional Specifies the alter operation such as ADD or DELETE to be performed on the VectorStore.") update_style: Optional[Literal["MINOR", "MAJOR"]] = Field(None, description="Optional Specifies the update style to be used for the VectorStore.") embeddings_model: str = Field(None, description="Optional Specifies the embedding model to be used for vectorization.") embeddings_dims: Optional[int] = Field(None, description="Optional Specifies the number of dimensions for the embeddings.") metric: Optional[str] = Field(None, description="Optional Specifies the metric to be used for calculating the distance between the vectors.") search_algorithm: Optional[str] = Field(None, description="Optional Specifies the search algorithm to be used for similarity search.") initial_centroids_method: Optional[str] = Field(None, description="Optional Specifies the method to be used for initializing centroids.") train_numclusters: Optional[int] = Field(None, description="Optional Specifies the number of clusters to be used for training.") max_iternum: Optional[int] = Field(None, description="Optional Specifies the maximum number of iterations for training.") stop_threshold: Optional[float] = Field(None, description="Optional Specifies the threshold for stopping the training process.") seed: Optional[int] = Field(None, description="Optional Specifies the seed value for random number generation.") num_init: Optional[int] = Field(None, description="Optional Specifies the number of initializations to be performed for k-means.") top_k: Optional[int] = Field(None, description="Optional Specifies the number of top results to be returned for similarity search.") search_threshold: Optional[float] = Field(None, description="OptionalSpecifies the threshold value to consider for matching tables/views while searching.") search_numcluster: Optional[int] = Field(None, description="Optional Specifies the number of clusters to be used for searching.") prompt: Optional[str] = Field(None, description="Optional Specifies the prompt to be used by language model to generate responses using top matches.") chat_completion_model: Optional[str] = Field(None, description="Optional Specifies the chat completion model to be used for generating responses.") document_files: Optional[List[str]] = Field(None, description="Optional Specifies the list of document files to be indexed for vector store.") ef_search: Optional[int] = Field(None, description="Optional Specifies the number of neighbors to be considered for hnsw algorithm during similarity search.") num_layers: Optional[int] = Field(None, description="Optional Specifies the maximum number of layers to be used for hnsw algorithm during vector store creation.") ef_construction: Optional[int] = Field(None, description="Optional Specifies the number of neighbors to be considered for hnsw algorithm during vector store creation.") num_connpernode: Optional[int] = Field(None, description="Optional Specifies the number of connections per node to be used for hnsw algorithm during vector store creation.") maxnum_connpernode: Optional[int] = Field(None, description="Optional Specifies the maximum number of connections per node to be used for hnsw algorithm during vector store creation.") apply_heuristics: Optional[bool] = Field(None, description="Optional Specifies whether to apply heuristics for hnsw algorithm during vector store creation.") include_objects: Optional[List[str]] = Field(None, description="Optional Specifies the list of tables/views included in the metadata based vector store.") exclude_objects: Optional[List[str]] = Field(None, description="Optional Specifies the list of tables/views excluded from the metadata based vector store.") sample_size: Optional[int] = Field(None, description="Optional Specifies the number of rows to sample tables/views for the metadata based vector store embeddings.") rerank_weight: Optional[float] = Field(None, description="Optional Specifies the weight to be used for reranking the search results.") relevance_top_k: Optional[int] = Field(None, description="Optional Specifies the number of top similarity matches to be considered for reranking.") relevance_search_threshold: Optional[float] = Field(None, description="Optional Specifies the threshold value to be consider matching tables/views while reranking.") include_patterns: Optional[List[str]] = Field(None, description="Optional Specifies the list of patterns to be included in the metadata based vector store.") exclude_patterns: Optional[List[str]] = Field(None, description="Optional Specifies the list of patterns to be excluded from the metadata based vector store.") ignore_embedding_errors: Optional[bool] = Field(None, description="Optional Specifies whether to ignore embedding errors during embedding generation. Applicable only for AWS.") chat_completion_max_tokens: Optional[int] = Field(None, description="Optional Specifies the maximum number of tokens to be generated by chat completion model.")