evalaudit 0.2.0__py3-none-any.whl
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- evalaudit/__init__.py +56 -0
- evalaudit/_types.py +1572 -0
- evalaudit/agreement.py +600 -0
- evalaudit/audit.py +821 -0
- evalaudit/compare.py +708 -0
- evalaudit/judge.py +744 -0
- evalaudit/pairwise.py +682 -0
- evalaudit/power.py +528 -0
- evalaudit/scores.py +136 -0
- evalaudit-0.2.0.dist-info/METADATA +313 -0
- evalaudit-0.2.0.dist-info/RECORD +13 -0
- evalaudit-0.2.0.dist-info/WHEEL +4 -0
- evalaudit-0.2.0.dist-info/licenses/LICENSE +21 -0
evalaudit/__init__.py
ADDED
|
@@ -0,0 +1,56 @@
|
|
|
1
|
+
"""evalaudit: statistical validity checks for AI evaluations."""
|
|
2
|
+
|
|
3
|
+
__version__ = "0.1.0"
|
|
4
|
+
|
|
5
|
+
from ._types import (
|
|
6
|
+
AgreementResult,
|
|
7
|
+
AuditConfig,
|
|
8
|
+
AuditReport,
|
|
9
|
+
BTResult,
|
|
10
|
+
ComparisonResult,
|
|
11
|
+
Finding,
|
|
12
|
+
JudgeValidation,
|
|
13
|
+
KappaResult,
|
|
14
|
+
LengthBias,
|
|
15
|
+
PositionBias,
|
|
16
|
+
PowerResult,
|
|
17
|
+
ScoreCI,
|
|
18
|
+
SkippedCheck,
|
|
19
|
+
)
|
|
20
|
+
from .agreement import cohens_kappa, fleiss_kappa, rater_agreement
|
|
21
|
+
from .compare import compare_independent, compare_paired
|
|
22
|
+
from .pairwise import bradley_terry, to_elo
|
|
23
|
+
from .judge import judge_validation, length_bias, position_bias
|
|
24
|
+
from .power import detectable_effect, min_sample_size
|
|
25
|
+
from .scores import score_ci
|
|
26
|
+
from .audit import audit
|
|
27
|
+
|
|
28
|
+
__all__ = [
|
|
29
|
+
"score_ci",
|
|
30
|
+
"compare_paired",
|
|
31
|
+
"compare_independent",
|
|
32
|
+
"rater_agreement",
|
|
33
|
+
"cohens_kappa",
|
|
34
|
+
"fleiss_kappa",
|
|
35
|
+
"bradley_terry",
|
|
36
|
+
"to_elo",
|
|
37
|
+
"judge_validation",
|
|
38
|
+
"position_bias",
|
|
39
|
+
"length_bias",
|
|
40
|
+
"detectable_effect",
|
|
41
|
+
"min_sample_size",
|
|
42
|
+
"audit",
|
|
43
|
+
"ScoreCI",
|
|
44
|
+
"ComparisonResult",
|
|
45
|
+
"AgreementResult",
|
|
46
|
+
"KappaResult",
|
|
47
|
+
"BTResult",
|
|
48
|
+
"JudgeValidation",
|
|
49
|
+
"PositionBias",
|
|
50
|
+
"LengthBias",
|
|
51
|
+
"PowerResult",
|
|
52
|
+
"AuditConfig",
|
|
53
|
+
"AuditReport",
|
|
54
|
+
"Finding",
|
|
55
|
+
"SkippedCheck",
|
|
56
|
+
]
|