truthound-dashboard 1.2.1__py3-none-any.whl → 1.3.0__py3-none-any.whl

truthound_dashboard/schemas/validators/distribution_validators.py

@@ -0,0 +1,422 @@
+"""Distribution validators.
+
+Validators for value ranges, statistical distributions, and outlier detection.
+"""
+
+from .base import (
+    ParameterDefinition,
+    ParameterType,
+    ValidatorCategory,
+    ValidatorDefinition,
+)
+
+DISTRIBUTION_VALIDATORS: list[ValidatorDefinition] = [
+    ValidatorDefinition(
+        name="Between",
+        display_name="Value Between",
+        category=ValidatorCategory.DISTRIBUTION,
+        description="Validates that values fall within a specified range (inclusive).",
+        parameters=[
+            ParameterDefinition(
+                name="column",
+                label="Column",
+                type=ParameterType.COLUMN,
+                required=True,
+            ),
+            ParameterDefinition(
+                name="min_value",
+                label="Minimum Value",
+                type=ParameterType.FLOAT,
+                required=True,
+            ),
+            ParameterDefinition(
+                name="max_value",
+                label="Maximum Value",
+                type=ParameterType.FLOAT,
+                required=True,
+            ),
+            ParameterDefinition(
+                name="mostly",
+                label="Mostly (Threshold)",
+                type=ParameterType.FLOAT,
+                description="Acceptable pass ratio (0.0-1.0)",
+                min_value=0,
+                max_value=1,
+            ),
+        ],
+        tags=["distribution", "range", "bounds"],
+        severity_default="medium",
+    ),
+    ValidatorDefinition(
+        name="Range",
+        display_name="Range",
+        category=ValidatorCategory.DISTRIBUTION,
+        description="Validates value ranges with optional inclusivity control.",
+        parameters=[
+            ParameterDefinition(
+                name="column",
+                label="Column",
+                type=ParameterType.COLUMN,
+                required=True,
+            ),
+            ParameterDefinition(
+                name="min_value",
+                label="Minimum Value",
+                type=ParameterType.FLOAT,
+            ),
+            ParameterDefinition(
+                name="max_value",
+                label="Maximum Value",
+                type=ParameterType.FLOAT,
+            ),
+            ParameterDefinition(
+                name="inclusive_min",
+                label="Include Minimum",
+                type=ParameterType.BOOLEAN,
+                default=True,
+            ),
+            ParameterDefinition(
+                name="inclusive_max",
+                label="Include Maximum",
+                type=ParameterType.BOOLEAN,
+                default=True,
+            ),
+        ],
+        tags=["distribution", "range", "bounds"],
+        severity_default="medium",
+    ),
+    ValidatorDefinition(
+        name="Positive",
+        display_name="Positive Values",
+        category=ValidatorCategory.DISTRIBUTION,
+        description="Ensures all values are strictly positive (> 0).",
+        parameters=[
+            ParameterDefinition(
+                name="column",
+                label="Column",
+                type=ParameterType.COLUMN,
+                required=True,
+            ),
+        ],
+        tags=["distribution", "positive", "numeric"],
+        severity_default="high",
+    ),
+    ValidatorDefinition(
+        name="NonNegative",
+        display_name="Non-Negative Values",
+        category=ValidatorCategory.DISTRIBUTION,
+        description="Ensures all values are non-negative (>= 0).",
+        parameters=[
+            ParameterDefinition(
+                name="column",
+                label="Column",
+                type=ParameterType.COLUMN,
+                required=True,
+            ),
+        ],
+        tags=["distribution", "non-negative", "numeric"],
+        severity_default="high",
+    ),
+    ValidatorDefinition(
+        name="InSet",
+        display_name="In Set",
+        category=ValidatorCategory.DISTRIBUTION,
+        description="Validates that values belong to a predefined set.",
+        parameters=[
+            ParameterDefinition(
+                name="column",
+                label="Column",
+                type=ParameterType.COLUMN,
+                required=True,
+            ),
+            ParameterDefinition(
+                name="value_set",
+                label="Allowed Values",
+                type=ParameterType.STRING_LIST,
+                description="Set of allowed values",
+                required=True,
+            ),
+            ParameterDefinition(
+                name="mostly",
+                label="Mostly (Threshold)",
+                type=ParameterType.FLOAT,
+                min_value=0,
+                max_value=1,
+            ),
+        ],
+        tags=["distribution", "set", "enum"],
+        severity_default="medium",
+    ),
+    ValidatorDefinition(
+        name="NotInSet",
+        display_name="Not In Set",
+        category=ValidatorCategory.DISTRIBUTION,
+        description="Ensures values do not belong to a forbidden set.",
+        parameters=[
+            ParameterDefinition(
+                name="column",
+                label="Column",
+                type=ParameterType.COLUMN,
+                required=True,
+            ),
+            ParameterDefinition(
+                name="forbidden_set",
+                label="Forbidden Values",
+                type=ParameterType.STRING_LIST,
+                description="Set of forbidden values",
+                required=True,
+            ),
+        ],
+        tags=["distribution", "set", "forbidden"],
+        severity_default="high",
+    ),
+    ValidatorDefinition(
+        name="Increasing",
+        display_name="Increasing",
+        category=ValidatorCategory.DISTRIBUTION,
+        description="Validates that values are monotonically increasing.",
+        parameters=[
+            ParameterDefinition(
+                name="column",
+                label="Column",
+                type=ParameterType.COLUMN,
+                required=True,
+            ),
+            ParameterDefinition(
+                name="strict",
+                label="Strictly Increasing",
+                type=ParameterType.BOOLEAN,
+                description="If true, values must be strictly increasing (no equal values)",
+                default=False,
+            ),
+        ],
+        tags=["distribution", "monotonic", "increasing"],
+        severity_default="medium",
+    ),
+    ValidatorDefinition(
+        name="Decreasing",
+        display_name="Decreasing",
+        category=ValidatorCategory.DISTRIBUTION,
+        description="Validates that values are monotonically decreasing.",
+        parameters=[
+            ParameterDefinition(
+                name="column",
+                label="Column",
+                type=ParameterType.COLUMN,
+                required=True,
+            ),
+            ParameterDefinition(
+                name="strict",
+                label="Strictly Decreasing",
+                type=ParameterType.BOOLEAN,
+                default=False,
+            ),
+        ],
+        tags=["distribution", "monotonic", "decreasing"],
+        severity_default="medium",
+    ),
+    ValidatorDefinition(
+        name="Outlier",
+        display_name="Outlier (IQR)",
+        category=ValidatorCategory.DISTRIBUTION,
+        description="Detects statistical outliers using the IQR method.",
+        parameters=[
+            ParameterDefinition(
+                name="column",
+                label="Column",
+                type=ParameterType.COLUMN,
+                required=True,
+            ),
+            ParameterDefinition(
+                name="iqr_multiplier",
+                label="IQR Multiplier",
+                type=ParameterType.FLOAT,
+                description="Multiplier for IQR (1.5 = standard, 3.0 = extreme)",
+                default=1.5,
+                min_value=0,
+            ),
+        ],
+        tags=["distribution", "outlier", "iqr", "anomaly"],
+        severity_default="medium",
+    ),
+    ValidatorDefinition(
+        name="ZScoreOutlier",
+        display_name="Outlier (Z-Score)",
+        category=ValidatorCategory.DISTRIBUTION,
+        description="Detects outliers using Z-score methodology.",
+        parameters=[
+            ParameterDefinition(
+                name="column",
+                label="Column",
+                type=ParameterType.COLUMN,
+                required=True,
+            ),
+            ParameterDefinition(
+                name="threshold",
+                label="Z-Score Threshold",
+                type=ParameterType.FLOAT,
+                description="Z-score threshold (default: 3.0)",
+                default=3.0,
+                min_value=0,
+            ),
+        ],
+        tags=["distribution", "outlier", "zscore", "anomaly"],
+        severity_default="medium",
+    ),
+    ValidatorDefinition(
+        name="Quantile",
+        display_name="Quantile",
+        category=ValidatorCategory.DISTRIBUTION,
+        description="Validates that values fall within specified quantile bounds.",
+        parameters=[
+            ParameterDefinition(
+                name="column",
+                label="Column",
+                type=ParameterType.COLUMN,
+                required=True,
+            ),
+            ParameterDefinition(
+                name="quantile",
+                label="Quantile",
+                type=ParameterType.FLOAT,
+                description="Quantile value (0.0-1.0, e.g., 0.95 for 95th percentile)",
+                required=True,
+                min_value=0,
+                max_value=1,
+            ),
+            ParameterDefinition(
+                name="max_value",
+                label="Maximum Value at Quantile",
+                type=ParameterType.FLOAT,
+                description="Maximum allowed value at the specified quantile",
+                required=True,
+            ),
+        ],
+        tags=["distribution", "quantile", "percentile"],
+        severity_default="medium",
+    ),
+    ValidatorDefinition(
+        name="Distribution",
+        display_name="Distribution",
+        category=ValidatorCategory.DISTRIBUTION,
+        description="Validates that data follows an expected distribution pattern.",
+        parameters=[
+            ParameterDefinition(
+                name="column",
+                label="Column",
+                type=ParameterType.COLUMN,
+                required=True,
+            ),
+            ParameterDefinition(
+                name="distribution",
+                label="Expected Distribution",
+                type=ParameterType.SELECT,
+                description="Expected distribution type",
+                required=True,
+                options=[
+                    {"value": "normal", "label": "Normal (Gaussian)"},
+                    {"value": "uniform", "label": "Uniform"},
+                    {"value": "exponential", "label": "Exponential"},
+                    {"value": "poisson", "label": "Poisson"},
+                    {"value": "binomial", "label": "Binomial"},
+                ],
+            ),
+        ],
+        tags=["distribution", "statistical", "pattern"],
+        severity_default="low",
+    ),
+    ValidatorDefinition(
+        name="KLDivergence",
+        display_name="KL Divergence",
+        category=ValidatorCategory.DISTRIBUTION,
+        description="Validates distribution similarity using Kullback-Leibler divergence.",
+        parameters=[
+            ParameterDefinition(
+                name="column",
+                label="Column",
+                type=ParameterType.COLUMN,
+                required=True,
+            ),
+            ParameterDefinition(
+                name="reference_distribution",
+                label="Reference Distribution",
+                type=ParameterType.SCHEMA,
+                description="Reference probability distribution as JSON",
+                required=True,
+                placeholder='{"A": 0.5, "B": 0.3, "C": 0.2}',
+            ),
+            ParameterDefinition(
+                name="threshold",
+                label="Threshold",
+                type=ParameterType.FLOAT,
+                description="Maximum allowed KL divergence",
+                required=True,
+                default=0.1,
+                min_value=0,
+            ),
+        ],
+        tags=["distribution", "kl", "divergence", "statistical"],
+        severity_default="medium",
+    ),
+    ValidatorDefinition(
+        name="ChiSquare",
+        display_name="Chi-Square Test",
+        category=ValidatorCategory.DISTRIBUTION,
+        description="Performs chi-square goodness-of-fit testing.",
+        parameters=[
+            ParameterDefinition(
+                name="column",
+                label="Column",
+                type=ParameterType.COLUMN,
+                required=True,
+            ),
+            ParameterDefinition(
+                name="expected_frequencies",
+                label="Expected Frequencies",
+                type=ParameterType.SCHEMA,
+                description="Expected frequency distribution as JSON",
+                required=True,
+            ),
+            ParameterDefinition(
+                name="alpha",
+                label="Significance Level",
+                type=ParameterType.FLOAT,
+                description="Significance level (default: 0.05)",
+                default=0.05,
+                min_value=0,
+                max_value=1,
+            ),
+        ],
+        tags=["distribution", "chi-square", "statistical", "test"],
+        severity_default="medium",
+    ),
+    ValidatorDefinition(
+        name="MostCommonValue",
+        display_name="Most Common Value",
+        category=ValidatorCategory.DISTRIBUTION,
+        description="Validates the most common value and its frequency.",
+        parameters=[
+            ParameterDefinition(
+                name="column",
+                label="Column",
+                type=ParameterType.COLUMN,
+                required=True,
+            ),
+            ParameterDefinition(
+                name="expected_value",
+                label="Expected Most Common Value",
+                type=ParameterType.STRING,
+            ),
+            ParameterDefinition(
+                name="min_frequency",
+                label="Minimum Frequency",
+                type=ParameterType.FLOAT,
+                description="Minimum frequency ratio (0.0-1.0)",
+                min_value=0,
+                max_value=1,
+            ),
+        ],
+        tags=["distribution", "frequency", "mode"],
+        severity_default="low",
+    ),
+]