question

Typed classification questions and heterogeneous batches.

Choice labels retain application-native values and the full distribution. Confidence describes provider concentration evidence, not the probability that an answer is correct. Source definitions are total; use check_* for definitions assembled from runtime data.

Types

Alternative

</>

pub type Alternative(a) =
  @internal Alternative(a)

Batch

</>

pub type Batch(a) =
  @internal Batch(a)

Choice

</>

pub type Choice(a) =
  @internal Choice(a)

Error

opaque </>

One failure boundary for runtime definitions and answer admission.

pub opaque type Error

ErrorKind

</>

pub type ErrorKind {
  InvalidDefinition
  InvalidAnswer
}

Constructors

```
InvalidDefinition
```
```
InvalidAnswer
```

Noul

</>

pub type Noul =
  @internal Noul

Probability

</>

pub type Probability(a) =
  @internal Probability(a)

Question

</>

pub type Question(a) =
  @internal Question(a)

Score

</>

pub type Score =
  @internal Score

Values

alternative

</>

pub fn alternative(
  label: String,
  value: a,
  description: value.Value,
) -> Alternative(a)

ask

</>

pub fn ask(id: String, question: Question(a)) -> Batch(a)

check_ask

</>

pub fn check_ask(
  id: String,
  question: Question(a),
) -> Result(Batch(a), Error)

check_choice

</>

pub fn check_choice(
  instructions: value.Value,
  alternatives: List(Alternative(a)),
) -> Result(Question(Choice(a)), Error)

check_combine

</>

pub fn check_combine(
  left: Batch(a),
  right: Batch(b),
) -> Result(Batch(#(a, b)), Error)

check_noul

</>

pub fn check_noul(
  instructions: value.Value,
  criteria: option.Option(#(value.Value, value.Value)),
) -> Result(Question(Noul), Error)

check_score

</>

pub fn check_score(
  instructions: value.Value,
  levels: List(value.Value),
) -> Result(Question(Score), Error)

choice

</>

pub fn choice(
  instructions: value.Value,
  alternatives: List(Alternative(a)),
) -> Question(Choice(a))

combine

</>

pub fn combine(left: Batch(a), right: Batch(b)) -> Batch(#(a, b))

decode

</>

pub fn decode(
  batch: Batch(a),
  answers: List(#(String, protocol.Answer)),
) -> Result(a, Error)

definitions

</>

pub fn definitions(
  batch: Batch(a),
) -> List(#(String, protocol.QuestionView))

Provider-neutral question vocabulary; a wire may project it differently.

describe_error

</>

pub fn describe_error(error: Error) -> String

error_kind

</>

pub fn error_kind(error: Error) -> ErrorKind

noul

</>

pub fn noul(
  instructions: value.Value,
  criteria: option.Option(#(value.Value, value.Value)),
) -> Question(Noul)

score

</>

pub fn score(
  instructions: value.Value,
  levels: List(value.Value),
) -> Question(Score)