Blocks API Reference

This page provides auto-generated API documentation from source docstrings. When reading raw markdown, refer to the quick reference below.

Quick Reference

Class	Import	Description
BaseBlock	`from sdg_hub import BaseBlock`	Abstract base class for all blocks with Pydantic validation
BlockRegistry	`from sdg_hub import BlockRegistry`	Registry for block discovery and metadata

Class	Import	Description
LLMChatBlock	`from sdg_hub.core.blocks import LLMChatBlock`	Unified LLM chat block supporting 100+ providers via LiteLLM
PromptBuilderBlock	`from sdg_hub.core.blocks import PromptBuilderBlock`	Formats prompts into structured chat messages or plain text
LLMResponseExtractorBlock	`from sdg_hub.core.blocks import LLMResponseExtractorBlock`	Extracts specified fields from LLM response objects
LLMErrorHandler	`from sdg_hub.core.blocks.llm.error_handler import LLMErrorHandler`	Centralized error handling for LLM operations across all providers

Class	Import	Description
TagParserBlock	`from sdg_hub.core.blocks import TagParserBlock`	Parses text content using start/end tags
JSONParserBlock	`from sdg_hub.core.blocks import JSONParserBlock`	Parses JSON from text and expands fields into separate columns
RegexParserBlock	`from sdg_hub.core.blocks import RegexParserBlock`	Parses text content using regex patterns

Class	Import	Description
TextConcatBlock	`from sdg_hub.core.blocks import TextConcatBlock`	Combines multiple columns into a single column using a separator
DuplicateColumnsBlock	`from sdg_hub.core.blocks import DuplicateColumnsBlock`	Duplicates existing columns with new names according to a mapping
RenameColumnsBlock	`from sdg_hub.core.blocks import RenameColumnsBlock`	Renames columns in a dataset according to a mapping
MeltColumnsBlock	`from sdg_hub.core.blocks import MeltColumnsBlock`	Transforms wide dataset format into long format by melting columns into rows
IndexBasedMapperBlock	`from sdg_hub.core.blocks import IndexBasedMapperBlock`	Maps values from source columns to output based on choice columns
UniformColumnValueSetter	`from sdg_hub.core.blocks import UniformColumnValueSetter`	Replaces all values in a column with a summary statistic (mode, mean, median)
JSONStructureBlock	`from sdg_hub.core.blocks.transform.json_structure_block import JSONStructureBlock`	Combines multiple columns into a single JSON object column
RowMultiplierBlock	`from sdg_hub.core.blocks.transform.row_multiplier import RowMultiplierBlock`	Duplicates each row a configurable number of times
SamplerBlock	`from sdg_hub.core.blocks.transform.sampler import SamplerBlock`	Randomly samples n values from a list column

Class	Import	Description
ColumnValueFilterBlock	`from sdg_hub.core.blocks import ColumnValueFilterBlock`	Filters datasets based on column values using comparison operations

Class	Import	Description
AgentBlock	`from sdg_hub.core.blocks import AgentBlock`	Execute agent frameworks (Langflow, etc.) on DataFrame rows
AgentResponseExtractorBlock	`from sdg_hub.core.blocks.agent.agent_response_extractor_block import AgentResponseExtractorBlock`	Extracts text content from agent framework responses

Class	Import	Description
MCPAgentBlock	`from sdg_hub.core.blocks import MCPAgentBlock`	LLM agent with remote MCP tools in an agentic loop

::: sdg_hub.core.blocks.base.BaseBlock options: members_order: source show_source: false

::: sdg_hub.core.blocks.registry.BlockRegistry options: members_order: source show_source: false

::: sdg_hub.core.blocks.llm.llm_chat_block.LLMChatBlock options: members_order: source show_source: false

::: sdg_hub.core.blocks.llm.prompt_builder_block.PromptBuilderBlock options: members_order: source show_source: false

::: sdg_hub.core.blocks.llm.llm_response_extractor_block.LLMResponseExtractorBlock options: members_order: source show_source: false

::: sdg_hub.core.blocks.llm.error_handler.LLMErrorHandler options: members_order: source show_source: false

::: sdg_hub.core.blocks.parsing.tag_parser_block.TagParserBlock options: members_order: source show_source: false

::: sdg_hub.core.blocks.parsing.json_parser_block.JSONParserBlock options: members_order: source show_source: false

::: sdg_hub.core.blocks.parsing.regex_parser_block.RegexParserBlock options: members_order: source show_source: false

::: sdg_hub.core.blocks.transform.text_concat.TextConcatBlock options: members_order: source show_source: false

::: sdg_hub.core.blocks.transform.duplicate_columns.DuplicateColumnsBlock options: members_order: source show_source: false

::: sdg_hub.core.blocks.transform.rename_columns.RenameColumnsBlock options: members_order: source show_source: false

::: sdg_hub.core.blocks.transform.melt_columns.MeltColumnsBlock options: members_order: source show_source: false

::: sdg_hub.core.blocks.transform.index_based_mapper.IndexBasedMapperBlock options: members_order: source show_source: false

::: sdg_hub.core.blocks.transform.uniform_col_val_setter.UniformColumnValueSetter options: members_order: source show_source: false

::: sdg_hub.core.blocks.transform.json_structure_block.JSONStructureBlock options: members_order: source show_source: false

::: sdg_hub.core.blocks.transform.row_multiplier.RowMultiplierBlock options: members_order: source show_source: false

::: sdg_hub.core.blocks.transform.sampler.SamplerBlock options: members_order: source show_source: false

::: sdg_hub.core.blocks.filtering.column_value_filter.ColumnValueFilterBlock options: members_order: source show_source: false

::: sdg_hub.core.blocks.agent.agent_block.AgentBlock options: members_order: source show_source: false

::: sdg_hub.core.blocks.agent.agent_response_extractor_block.AgentResponseExtractorBlock options: members_order: source show_source: false

::: sdg_hub.core.blocks.mcp.mcp_agent_block.MCPAgentBlock options: members_order: source show_source: false