Plan the next question

Install Relativities before running the synthetic examples. rslc plan keeps one person’s answers in a journey file and prints the next question as a card. The planner asks only for inputs that still change a result. Rate with missing answers shows the result that each rater gives.

Start a journey

Save the shared renters example, a rater from the synthetic carrier Example Mutual:

rslc examples get renters

WROTE renters/rater.rsl
WROTE renters/tables/BaseRate.csv
WROTE renters/tables/BuildingFactor.csv
WROTE renters/tables/DeductibleFactor.csv

Start a journey for the rater:

rslc plan start journey.json --rater renters=renters

yesno Policy.insured_now: Do you have a renters policy now?
  section: Insurance
ready: none

journey.json is the journey file. Each later command reads it and writes it back.
--rater renters=renters gives the rater an ID and names the folder that holds rater.rsl.
The first line of a card holds its kind, the path of the input, and the question.
ready lists each rater that is settled.

Answer a card

Pass the path of the card and the answer as JSON:

rslc plan answer journey.json Policy.insured_now false

number Policy.contents: What are your belongings worth?
  section: Home
  help: Count furniture, clothing and electronics.
  min: 10000
  max: 100000
ready: none

A number card shows the smallest and the largest answer. The help line comes from the rater’s help annotation.

rslc plan answer journey.json Policy.contents 30000

choice Policy.territory: Where is the home?
  section: Home
  - "metro": Metro
  - "suburban": Suburban
  - "rural": Rural
ready: none

A choice card lists each answer as its JSON value, then its label. A text answer keeps its JSON quotes, so wrap it in single quotes for the shell:

rslc plan answer journey.json Policy.territory '"rural"'

yesno Policy.has_dog: Do you have a dog?
  section: Household
ready: none

The planner leaves construction out. Each rural row of the building table holds the same factor.

rslc plan answer journey.json Policy.has_dog true

yesno Policy.dog_has_bitten: Has the dog bitten anyone?
  section: Household
ready: none

dog_has_bitten is a follow-up input. The planner asks it only after has_dog is true.

rslc plan answer journey.json Policy.dog_has_bitten false

done: every rater is settled
ready: renters

Five answers settle the rater. It declares seven inputs.

Read the premium

The journey file holds the answers under facts. Rate them with rslc.journey.Journey:

import json
from pathlib import Path

import rslc
from rslc.journey import Journey

state = json.loads(Path("journey.json").read_text())
journey = Journey({name: rslc.PreparedRater(folder) for name, folder in state["raters"].items()})
print(state["facts"])
# {'Policy': {'insured_now': False, 'contents': 30000, 'territory': 'rural', 'has_dog': True, 'dog_has_bitten': False}}
print(journey.evaluations(state["facts"]))
# {'renters': {'settled': {'ok': True, 'outputs': {'Policy': {'premium': 241}}}}}

The annual premium is $241. Results lists each field of an evaluation.

Plan for several raters

A journey takes one --rater for each rater. Save a second, smaller rater as basic/rater.rsl:

Complete example: basic/rater.rsl

basic/rater.rsl

spec {
    rsl: "0.4.1",
    carrier: "Example Mutual",
    state: AZ,
    product: "Renters Basic",
    effective_date: 2026-01-01
}

Sources = {
    manual: Source = {
        artifact: artifact("00000000-0000-4000-8000-000000000000"),
        kind: pdf
    }
}

Constants = {
    rate_per_1000 [cite: manual:p2] = 6,
    dog_surcharge [cite: manual:p3] = 40
}

Records = {
    Policy = {
        inputs {
            contents: Num {precision: int} [ask: "What are your belongings worth?", group: "Home"]
                where self in [10000, 50000] [code: "CONTENTS-LIMIT", message: "The basic program covers belongings up to $50,000.", cite: manual:p2],
            has_dog: Bool [ask: "Do you have a dog?", group: "Household"]
        },
        step rating [cite: manual:p2] {
            dog_charge: Num = if self.has_dog then dog_surcharge else 0,
            premium: Premium [cite: manual:p2] =
                round(self.contents / 1000 * rate_per_1000 + self.dog_charge, 0, half_up)
        }
    }
}

RatingOutcomes = {
    premium: Policy.premium
}

Start a journey for both raters, and give each a title:

rslc plan start basket.json --rater renters=renters --rater basic=basic --title "renters=Renters" --title "basic=Renters Basic"

yesno Policy.insured_now: Do you have a renters policy now?
  section: Insurance
  asked by: Renters
  still settled if skipped: 1
ready: none

asked by names the raters that declare the input, when some raters leave it out. still settled if skipped counts the other raters whose result ignores the answer. The two raters share the paths Policy.contents and Policy.has_dog, so one answer serves both.

Put a card off

A person who lacks an answer puts the card off. Skip the path of the card:

rslc plan skip basket.json Policy.insured_now

number Policy.contents: What are your belongings worth?
  section: Home
  help: Count furniture, clothing and electronics.
  min: 10000
  max: 100000
ready: none

rslc plan answer basket.json Policy.contents 30000

yesno Policy.has_dog: Do you have a dog?
  section: Household
ready: none

Renters waits on the skipped answer, so the planner first asks what Renters Basic needs.

rslc plan answer basket.json Policy.has_dog false

choice Policy.territory: Where is the home?
  section: Home
  still settled if skipped: 1
  - "metro": Metro
  - "suburban": Suburban
  - "rural": Rural
ready: basic

Renters Basic is ready after two answers. The remaining cards belong to Renters. A skipped card comes back after the other cards, marked (skipped before):

rslc plan answer basket.json Policy.territory '"rural"'

yesno Policy.insured_now: Do you have a renters policy now?
  section: Insurance
  still settled if skipped: 1
  (skipped before)
ready: basic

To print the current card again, run rslc plan next. It leaves the journey as it is:

rslc plan next basket.json

yesno Policy.insured_now: Do you have a renters policy now?
  section: Insurance
  still settled if skipped: 1
  (skipped before)
ready: basic

Search the answers of a card

To find an answer in a long list, search the card’s path:

rslc plan search basket.json sub --path Policy.territory

"suburban": Suburban

Each query word matches the start of a word in a label. A search card holds a list too long to print, and this command finds its answers.

Decline a document

The renters rater names a document with two annotations. gives says who has a declarations page, and found_on says that the page prints contents. The planner offers the document as a card:

rslc plan start documents.json --rater renters=renters
rslc plan answer documents.json Policy.insured_now true

yesno Policy.insured_now: Do you have a renters policy now?
  section: Insurance
ready: none
document document:declarations_page:
  section: Home
  help: declarations page
ready: none

A host that reads documents saves the answers it finds as facts. To go on with the questions, dismiss the document:

rslc plan dismiss documents.json declarations_page

number Policy.contents: What are your belongings worth?
  section: Home
  help: Count furniture, clothing and electronics.
  min: 10000
  max: 100000
ready: none

Drive a journey from Python

Journey.plan returns the same plan as rslc plan --json, for facts that your program holds:

from rslc.journey import answer

person = {
    "Policy.insured_now": False,
    "Policy.contents": 30000,
    "Policy.territory": "metro",
    "Policy.construction": "masonry",
    "Policy.has_dog": False,
}
facts = {"Policy": {}}
step = journey.plan(facts)
while step["next"]:
    card = step["next"]
    print(card["kind"], card["path"], card["text"])
    facts = answer(facts, card["path"], person[card["path"]])
    step = journey.plan(facts, after=card["path"])
print(step["ready"], journey.evaluations(facts)["renters"]["settled"]["outputs"])
# yesno Policy.insured_now Do you have a renters policy now?
# number Policy.contents What are your belongings worth?
# choice Policy.territory Where is the home?
# choice Policy.construction How is the building built?
# yesno Policy.has_dog Do you have a dog?
# ['renters'] {'Policy': {'premium': 207}}

answer returns a copy of the facts with one more answer.
after names the card answered last, so the planner stays in its section.
step["next"] is None when the plan has zero open cards. step["standings"] then gives the state of each rater.

Planner lists each command, each Journey method and each field of a plan.
Missing answers explains why a rater names the inputs that still matter.
Ask the questions a rater needs writes the ask, help and group annotations that a card shows.

Get started

Tutorials

Implement a filing

Test a rater

Quote and underwrite

Concepts

Reference

Plan the next question

Start a journey

Answer a card

Read the premium

Plan for several raters

Put a card off

Search the answers of a card

Decline a document

Drive a journey from Python

​Start a journey

​Answer a card

​Read the premium

​Plan for several raters

​Put a card off

​Search the answers of a card

​Decline a document

​Drive a journey from Python

​Related

Start a journey

Answer a card

Read the premium

Plan for several raters

Put a card off

Search the answers of a card

Decline a document

Drive a journey from Python

Related