cupel 1.2.0

Context window management pipeline for LLM applications
[test]
name = "PriorityScorer ranks items by priority value"
stage = "scoring"
scorer = "priority"

# 3 items with distinct priorities.
# countWithPriority = 3, denominator = 2.
# priority 1: rank 0 → 0/2 = 0.0
# priority 5: rank 1 → 1/2 = 0.5
# priority 10: rank 2 → 2/2 = 1.0

[[items]]
content = "low"
tokens = 100
priority = 1

[[items]]
content = "medium"
tokens = 100
priority = 5

[[items]]
content = "high"
tokens = 100
priority = 10

[[expected]]
content = "low"
score_approx = 0.0

[[expected]]
content = "medium"
score_approx = 0.5

[[expected]]
content = "high"
score_approx = 1.0

[tolerance]
score_epsilon = 1e-9