Profile Models

1@dataclass(frozen=True, slots=True)
2class ProfileResult:
3    dataset_summary: DatasetSummary
4    column_profiles: Mapping[str, ColumnProfile]
5    numeric_profiles: Mapping[str, NumericProfile]
6    categorical_profiles: Mapping[str, CategoricalProfile]
7    datetime_profiles: Mapping[str, DatetimeProfile]
8    text_profiles: Mapping[str, TextProfile]
9    missing_value_summary: MissingValueSummary
10    duplicate_summary: DuplicateSummary
11    correlation_summary: CorrelationSummary
12    dataset_metadata: DatasetMetadata
13    execution_metadata: ExecutionMetadata
14
15    def to_dict(self) -> dict[str, Any]: ...

1@dataclass(frozen=True, slots=True)
2class DatasetSummary:
3    row_count: int
4    column_count: int
5    size_in_bytes: int | None
6    missing_percentage: float
7    duplicate_percentage: float
8    num_numeric_columns: int
9    num_categorical_columns: int
10    num_datetime_columns: int
11    num_text_columns: int
12    num_constant_columns: int
13    num_fully_empty_columns: int

1@dataclass(frozen=True, slots=True)
2class ColumnProfile:
3    name: str
4    dtype: str
5    logical_type: str  # "numeric" | "categorical" | "datetime" | "text"
6    missing_count: int
7    missing_percentage: float
8    is_constant: bool
9    is_fully_empty: bool

1@dataclass(frozen=True, slots=True)
2class NumericProfile:
3    column_name: str
4    count: int
5    missing_count: int
6    missing_percentage: float
7    unique_count: int
8    mean: float | None
9    median: float | None
10    mode: float | None
11    minimum: float | None
12    maximum: float | None
13    range: float | None
14    variance: float | None
15    std_dev: float | None
16    q1: float | None
17    q2: float | None
18    q3: float | None
19    iqr: float | None
20    sum: float | None
21    zero_count: int
22    negative_count: int
23    positive_count: int
24    skewness: float | None
25    kurtosis: float | None

1@dataclass(frozen=True, slots=True)
2class CategoricalProfile:
3    column_name: str
4    cardinality: int
5    unique_count: int
6    missing_count: int
7    frequency_table: Mapping[str, int]
8    top_values: Sequence[tuple[str, int]]
9    least_frequent_values: Sequence[tuple[str, int]]
10    most_common_category: str | None
11    entropy: float | None  # Shannon entropy, base 2

1@dataclass(frozen=True, slots=True)
2class DatetimeProfile:
3    column_name: str
4    minimum: str | None        # ISO 8601
5    maximum: str | None        # ISO 8601
6    range_days: float | None
7    missing_count: int
8    earliest_record: str | None
9    latest_record: str | None

1@dataclass(frozen=True, slots=True)
2class TextProfile:
3    column_name: str
4    avg_length: float | None
5    min_length: int | None
6    max_length: int | None
7    empty_strings: int
8    whitespace_only: int
9    char_count: int
10    word_count: int

1@dataclass(frozen=True, slots=True)
2class MissingValueSummary:
3    column_missing_counts: Mapping[str, int]
4    column_missing_percentages: Mapping[str, float]
5    total_missing: int
6    dataset_missing_percentage: float
7
8
9@dataclass(frozen=True, slots=True)
10class DuplicateSummary:
11    duplicate_rows_count: int
12    duplicate_percentage: float
13    constant_columns: Sequence[str]
14    fully_empty_columns: Sequence[str]
15
16
17@dataclass(frozen=True, slots=True)
18class CorrelationSummary:
19    pearson: Mapping[str, Mapping[str, float | None]]
20    spearman: Mapping[str, Mapping[str, float | None]]  # reserved
21    kendall: Mapping[str, Mapping[str, float | None]]  # reserved

1@dataclass(frozen=True, slots=True)
2class DatasetMetadata:
3    source: str | None
4    file_size: int | None
5    backend: str  # "pandas" | "polars"
6    custom_metadata: Mapping[str, Any]
7
8
9@dataclass(frozen=True, slots=True)
10class ExecutionMetadata:
11    start_time: str  # ISO 8601
12    duration_seconds: float
13    featuresmith_version: str

1import featuresmith as fs
2
3profile = fs.profile("customers.csv")
4
5print(profile.dataset_summary.row_count)
6print(profile.numeric_profiles["age"].mean)
7print(profile.categorical_profiles["city"].most_common_category)
8print(profile.missing_value_summary.dataset_missing_percentage)