BaseView

Tip

• To understand the general idea better, visit Concept: Views and Concept: Freeform Views
• To learn about structured views, which in most cases are the recommended way to define views, see the reference for the BaseStructuredView class

dbally.views.base.BaseView

Base class for all Views, which are the main building blocks of db-ally. All classes implementing this interface have to be able to execute queries and return the result.

ask abstractmethod async

ask(query: str, llm: LLM, event_tracker: Optional[EventTracker] = None, n_retries: int = 3, dry_run: bool = False, llm_options: Optional[LLMOptions] = None) -> ViewExecutionResult

Executes the query and returns the result.

PARAMETER	DESCRIPTION
query	The natural language query to execute. TYPE: str
llm	The LLM used to execute the query. TYPE: LLM
event_tracker	The event tracker used to audit the query execution. TYPE: Optional[EventTracker] DEFAULT: None
n_retries	The number of retries to execute the query in case of errors. TYPE: int DEFAULT: 3
dry_run	If True, the query will not be used to fetch data from the datasource. TYPE: bool DEFAULT: False
llm_options	Options to use for the LLM. TYPE: Optional[LLMOptions] DEFAULT: None

RETURNS	DESCRIPTION
ViewExecutionResult	The result of the query.

Source code in src/dbally/views/base.py

@abc.abstractmethod
async def ask(
    self,
    query: str,
    llm: LLM,
    event_tracker: Optional[EventTracker] = None,
    n_retries: int = 3,
    dry_run: bool = False,
    llm_options: Optional[LLMOptions] = None,
) -> ViewExecutionResult:
    """
    Executes the query and returns the result.

    Args:
        query: The natural language query to execute.
        llm: The LLM used to execute the query.
        event_tracker: The event tracker used to audit the query execution.
        n_retries: The number of retries to execute the query in case of errors.
        dry_run: If True, the query will not be used to fetch data from the datasource.
        llm_options: Options to use for the LLM.

    Returns:
        The result of the query.
    """

list_similarity_indexes

list_similarity_indexes() -> Dict[AbstractSimilarityIndex, List[IndexLocation]]

Lists all the similarity indexes used by the view.

RETURNS	DESCRIPTION
Dict[AbstractSimilarityIndex, List[IndexLocation]]	Mapping of similarity indexes to their locations.

Source code in src/dbally/views/base.py

def list_similarity_indexes(self) -> Dict[AbstractSimilarityIndex, List[IndexLocation]]:
    """
    Lists all the similarity indexes used by the view.

    Returns:
        Mapping of similarity indexes to their locations.
    """
    return {}

list_few_shots

list_few_shots() -> List[FewShotExample]

List all examples to be injected into few-shot prompt.

RETURNS	DESCRIPTION
List[FewShotExample]	List of few-shot examples

Source code in src/dbally/views/base.py

def list_few_shots(self) -> List[FewShotExample]:
    """
    List all examples to be injected into few-shot prompt.

    Returns:
        List of few-shot examples
    """
    return []

dbally.collection.results.ViewExecutionResult dataclass

ViewExecutionResult(results: List[Dict[str, Any]], context: Dict[str, Any])

Represents the result of the query execution for a single view.

PARAMETER	DESCRIPTION
results	List of dictionaries containing the results of the query execution, each dictionary represents a row in the result set with column names as keys. TYPE: List[Dict[str, Any]]
context	Dictionary containing addtional metadata about the query execution. TYPE: Dict[str, Any]

results instance-attribute

results: List[Dict[str, Any]]

context instance-attribute

context: Dict[str, Any]