[Proposal] add `error` function

# Problem Statement

AFAIK, there is no built-in capability to raise errors from within a JMESpath expression. For example, suppose a JSON document has a `status` property, and its value is supposed to be `up` or `down`. The user wants to map `up` to `1` and `down` to `0`, which could be done using the following JMESpath expression:

```
((status == 'up') && `1`) || `0`)
```

The user may want to detect and raise an error when the input document has a value other than `up` or `down`. A workaround is to rely on the fact that some JMESpath expressions are invalid, e.g., `to_number('bad')` would raise an error.
```
((status == 'up') && `1`) || ((status == 'down') && `0`) || to_number('bad')
```
This works but it's very kludgy. In this specific example, one might argue this issue could be validated using a JSON schema, but sometimes there are no JSON schemas.

# Proposal

Add a new `error` function that takes a string expression. If the `error` function is evaluated, an error is raised with the specified message.

```
((status == 'up') && `1`) || ((status == 'down') && `0`) || error(join('invalid-value:', [status]))
```

# Additional Information

The spec states errors are raised when problems are encountered during the evaluation process. Currently, the following errors are defined:
1. `invalid-type`
2. `unknown-function`
3. `invalid-arity`

Related: jmespath/jmespath.site#115 , jmespath/jmespath.site#116

Review the proposal's error-function examples and the specification's existing error definitions first. Then read the related jmespath/jmespath.site#115 and #116 discussions. Done means resolving the design and documenting how the new function and raised error should behave.

Written by the indexing model from the issue text.