Developed by | Guardrails AI |
---|---|
Date of development | Feb 15, 2024 |
Validator type | Format |
Blog | - |
License | Apache 2 |
Input/Output | Output |
This validator checks if a string follows valid SQL code. In order to do this, it expects either a conn
string or a schema
file. It checks the validity of generated SQL by performing the following steps:
- If given a connection string, it connects to the DB instance.
- Alternatively, if given a schema, it creates a local instance of the database based on the schema and connects to the local DB.
- If neither a connection string nor schema is provided, then SQL validation is performed using
sqlvalidator
package. - In order to validate any generated LLM string representing a SQL statement, that statement is executed against the connected DB. If the statement runs without raising an exception, then the validator passes.
guardrails hub install hub://guardrails/valid_sql
In this example, we validate a generated SQL code by providing it with a schema file.
# Import Guard and Validator
from guardrails import Guard
from guardrails.hub import ValidSQL
# Setup Guard
guard = Guard().use(ValidSQL, on_fail="exception")
response = guard.validate("SELECT * FROM EMPLOYEES;") # Validator passes
try:
response = guard.validate("SELEKT ID FROM USERS;") # Validator fails
except Exception as e:
print(e)
Output:
Validation failed for field with errors:
__init__(self, conn=None, schema=None, on_fail="noop")
conn
(str): The connection string to connect to a sql schema. E.g.sqlite://
schema
(str): The.sql
file representing the DB schema.on_fail
(str, Callable): The policy to enact when a validator fails. Ifstr
, must be one ofreask
,fix
,filter
,refrain
,noop
,exception
orfix_reask
. Otherwise, must be a function that is called when the validator fails.
Initializes a new instance of the Validator class.
Parameters:
__call__(self, value, metadata={}) → ValidationResult
- This method should not be called directly by the user. Instead, invoke
guard.parse(...)
where this method will be called internally for each associated Validator. - When invoking
guard.parse(...)
, ensure to pass the appropriatemetadata
dictionary that includes keys and values required by this validator. Ifguard
is associated with multiple validators, combine all necessary metadata into a single dictionary. value
(Any): The input value to validate.metadata
(dict): A dictionary containing metadata required for validation. For this validator, no additional metadata keys are required.
Validates the given value
using the rules defined in this validator, relying on the metadata
provided to customize the validation process. This method is automatically invoked by guard.parse(...)
, ensuring the validation logic is applied to the input data.
Note:
Parameters: