systemOneTextInput()
function systemOneTextInput(request): (string | JSONType[] | Record<string, JSONType>)[];The content of a request the model reads as text: its state and each question’s instructions and criteria. Question ids and assets are not part of it.
Parameters
request
assets?
Record<string, {
data_base64?: string;
format?: "png" | "jpg" | "jpeg";
part?: string;
type?: "image";
}> = ...
Images that state, instructions or criteria name by their key: inline as base64, or in
the multipart file part part names.
model
string = ...
questions
Record<string,
| {
criteria: Record<string, string | JSONType[] | Record<string, JSONType> | null>;
instructions: string | JSONType[] | Record<string, JSONType>;
type: "choice";
}
| {
criteria: (string | JSONType[] | Record<string, JSONType>)[];
instructions: string | JSONType[] | Record<string, JSONType>;
type: "score";
}
| {
criteria?: {
false: string | JSONType[] | Record<string, JSONType>;
true: string | JSONType[] | Record<string, JSONType>;
};
instructions: string | JSONType[] | Record<string, JSONType>;
type: "noul";
}
| {
instructions: string | JSONType[] | Record<string, JSONType>;
type: "box";
}> = ...
Keyed by ids the caller chooses; the answers come back under the same keys.
state
string | JSONType[] | Record<string, JSONType> = SystemOneContentSchema