Expand description
A rubric followed across a conversation: every question asked again after each turn, and what it said drawn as one line per question.
A single call over a thread says where the answer ended up. It does not say when it got there — whether urgency was clear from the first message or only arrived with the third — and that is usually the thing worth knowing about a rubric meant to run on a live conversation. This module is the pure half: the prefixes to ask, and the lines to draw from what came back.
Structs§
- Series
- One question across the turns.
Functions§
- changes
turn 3 yes · turn 4 no, orno throughoutwhen it never changed its mind.- prefixes
- The session once per turn: the first turn, the first two, and so on up to the whole thread. Empty when the state is not a conversation.
- series
- Line each question up across the turns.
- sparkline
- One block per value, as tall as the value is of
top. The scale is fixed rather than fitted to the values, so a noul that sits at 0.9 all the way through reads as high and flat, not as noise. - trend_
lines - One line per question: the spark, where it started and ended, and every turn it changed.