from lunar.evals import exactMatch, contains, jsonValid

from lunar.evals import exactMatch

# Usage
result = client.evals.run(
    name="Exact Match Test",
    dataset=[
        {"input": "What is 2+2?", "expected": "4"},
    ],
    task=task,
    scorers=[exactMatch],
)

# Scores:
# "4" vs "4" → 1.0
# "4" vs "4 " → 1.0 (whitespace trimmed)
# "4" vs "Four" → 0.0

from lunar.evals import exactMatchIgnoreCase

# Scores:
# "Paris" vs "paris" → 1.0
# "PARIS" vs "Paris" → 1.0
# "Paris" vs "London" → 0.0

from lunar.evals import contains

# Scores:
# "The answer is 4" vs "4" → 1.0
# "Paris is the capital" vs "Paris" → 1.0
# "Hello world" vs "foo" → 0.0

from lunar.evals import containsIgnoreCase

# Scores:
# "The answer is FOUR" vs "four" → 1.0
# "PARIS is beautiful" vs "paris" → 1.0

from lunar.evals import startsWith

# Scores:
# "Hello, how are you?" vs "Hello" → 1.0
# "Hi there" vs "Hello" → 0.0

from lunar.evals import endsWith

# Scores:
# "The answer is 42" vs "42" → 1.0
# "42 is the answer" vs "42" → 0.0

from lunar.evals import jsonValid

# Scores:
# '{"name": "John"}' → 1.0
# '[1, 2, 3]' → 1.0
# 'not json' → 0.0
# '{"broken": }' → 0.0

from lunar.evals import notEmpty

# Scores:
# "Hello" → 1.0
# "   text   " → 1.0
# "" → 0.0
# "   " → 0.0

from lunar.evals import isNumeric

# Scores:
# "42" → 1.0
# "3.14" → 1.0
# "-100" → 1.0
# "1e10" → 1.0
# "not a number" → 0.0

from lunar import Lunar
from lunar.evals import (
    exactMatch,
    exactMatchIgnoreCase,
    contains,
    containsIgnoreCase,
    startsWith,
    endsWith,
    jsonValid,
    notEmpty,
    isNumeric,
)

client = Lunar()

# Test different scorers
dataset = [
    {"input": "What is 2+2?", "expected": "4"},
    {"input": "Generate valid JSON", "expected": None},
    {"input": "Say hello", "expected": "Hello"},
]

result = client.evals.run(
    name="Built-in Scorers Test",
    dataset=dataset,
    task=lambda x: client.chat.completions.create(
        model="gpt-4o-mini",
        messages=[{"role": "user", "content": x}]
    ).choices[0].message.content,
    scorers=[
        exactMatch,
        contains,
        jsonValid,
        notEmpty,
    ],
)

# Check results
for scorer_name, summary in result.summary.scores.items():
    print(f"{scorer_name}: {summary.mean:.2f}")

result = client.evals.run(
    name="Comprehensive Test",
    dataset=[
        {"input": "Return JSON with name field", "expected": "name"},
    ],
    task=task,
    scorers=[
        jsonValid,      # Is it valid JSON?
        contains,       # Does it contain "name"?
        notEmpty,       # Is there output?
    ],
)

# A response like '{"name": "John"}' would score:
# jsonValid: 1.0
# contains: 1.0
# notEmpty: 1.0