Create a test for your product.
Returns a Test object for the given parameters.
See a complete example of creating custom tests in our Create a Custom Test example.
The name of the test.
The type of test. Possible values:
QUALITY
: Tests that evaluate the quality and correctness of outputsRED_TEAMING
: Tests that evaluate security, safety, and bias aspectsSCENARIOS
: Tests that use conversation simulation to evaluate multi-turn dialogue interactionsThe ID of the product you want to evaluate.
Path to a local file (e.g., PDF, TXT, JSON, HTML, Markdown) containing the knowledge base. This file is uploaded to Galtea, which then generates test cases based on its content.
Optional few-shot examples to provide more context to our system about how the test cases should be generated. This can help our system better understand the expected format and style wanted for the test cases. Example:
QUALITY
.The language for generating synthetic test cases if ground_truth_file_path
is provided. Defaults to the language detected in the ground truth file. If provided, it should be written in English and be a valid language name from the ISO 639 standard (e.g., “english”, “spanish”, “french”). More information can be found at https://en.wikipedia.org/wiki/List_of_ISO_639_language_codes.
List of variant types to generate for each test case from the ground_truth_file_path
. For details on variants for Quality tests, see Quality Test Evolutions. Example: ["paraphrased", "typos", "incorrect", "cognitively_diverse", "linguistic_diverse"]
.
For Red Teaming tests, the variants
parameter can be used to specify which threat categories to generate test cases for.
A list of red teaming strategies used when generating test cases for Red Teaming tests. Strategies are techniques for modifying or obfuscating prompts generated for each threat (as specified in variants
for Red Teaming tests). This parameter applies only when type
is RED_TEAMING
.
If specified, this list must include the ‘original’ strategy.
Example: ["original", "base64", "leetspeak", "zero_width_insertion"]
.
An optional integer to limit the maximum number of test cases generated by Galtea when ground_truth_file_path
is used.
Path to a local CSV file containing predefined test cases. This file is uploaded to Galtea.
File Format by Test Type:
input
, expected_output
, tag
, source
columnsgoal
, user_persona
, initial_prompt
, stopping_criterias
, max_iterations
, scenario
columnsSee the Conversation Simulator Tutorial for detailed SCENARIOS CSV format examples.