Cohort And Analysis Reference

Cohort construction, preprocessing, descriptive tables, statistical tests, effects, bootstrap intervals, and phenotype rules.

For workflow context, see Cohort And Analysis module guide.

Functions And Classes

fyron.cohort.tables.build_survival_columns
fyron.cohort.tables.build_time_to_event_endpoint
fyron.cohort.tables.cohort_profile
fyron.cohort.tables.join_cohort_tables
fyron.cohort.tables.validate_cohort_table
fyron.dataframe.add_code_indicators
fyron.dataframe.add_date_difference
fyron.dataframe.add_icd10_disease_groups
fyron.dataframe.add_ops_categories
fyron.dataframe.add_time_window_flag
fyron.dataframe.age_at_date
fyron.dataframe.classify_icd10
fyron.dataframe.classify_icd10_string
fyron.dataframe.classify_ops
fyron.dataframe.classify_ops_string
fyron.dataframe.date_difference
fyron.dataframe.explode_code_column
fyron.dataframe.icd10_group_prevalence
fyron.dataframe.normalize_code
fyron.dataframe.ops_category_prevalence
fyron.dataframe.split_code_string
fyron.descriptive.analysis.compare_descriptive_groups
fyron.descriptive.analysis.create_table_one
fyron.descriptive.analysis.describe_categorical
fyron.descriptive.analysis.describe_numeric
fyron.descriptive.analysis.summarize_endpoint
fyron.phenotyping.rules.combine_phenotypes
fyron.phenotyping.rules.define_code_phenotype
fyron.phenotyping.rules.define_lab_threshold_phenotype
fyron.phenotyping.rules.define_medication_exposure
fyron.phenotyping.rules.phenotype_summary
fyron.preprocessing.encoding.apply_categorical_encoder
fyron.preprocessing.encoding.encode_categorical
fyron.preprocessing.encoding.fit_categorical_encoder
fyron.preprocessing.encoding.summarize_encoding_plan
fyron.preprocessing.imputation.apply_imputer
fyron.preprocessing.imputation.fit_imputer
fyron.preprocessing.imputation.impute_dataframe
fyron.preprocessing.imputation.summarize_imputation_plan
fyron.preprocessing.qc.compare_train_test_balance
fyron.preprocessing.qc.detect_potential_leakage
fyron.preprocessing.qc.feature_matrix_report
fyron.preprocessing.qc.find_duplicate_ids
fyron.preprocessing.qc.summarize_categorical_levels
fyron.preprocessing.qc.summarize_missingness
fyron.statistics.bootstrap.BootstrapResult
fyron.statistics.bootstrap.BootstrapResult.to_dict
fyron.statistics.bootstrap.bootstrap_binary_classification_metrics
fyron.statistics.bootstrap.bootstrap_metric
fyron.statistics.bootstrap.bootstrap_summary
fyron.statistics.comparison.bootstrap_auc_difference
fyron.statistics.comparison.mcnemar_test
fyron.statistics.correction.adjust_p_values
fyron.statistics.correction.p_value_correction_table
fyron.statistics.effects.cliffs_delta
fyron.statistics.effects.cohens_d
fyron.statistics.effects.cramers_v
fyron.statistics.effects.odds_ratio_ci
fyron.statistics.effects.risk_ratio_ci
fyron.statistics.effects.standardized_mean_difference
fyron.statistics.qc.correlation_pvalue_table
fyron.statistics.qc.missingness_by_group_table
fyron.statistics.qc.normality_table
fyron.statistics.qc.outlier_table
fyron.statistics.regression.fit_linear_regression_table
fyron.statistics.regression.fit_logistic_regression_table
fyron.statistics.regression.vif_table
fyron.statistics.tests.StatisticalTestResult
fyron.statistics.tests.StatisticalTestResult.to_dict
fyron.statistics.tests.categorical_test
fyron.statistics.tests.compare_multiple_groups
fyron.statistics.tests.compare_two_groups
fyron.statistics.tests.correlation_test
fyron.statistics.tests.normality_test

`fyron.cohort.tables.build_survival_columns`

Build standard survival `time and event` columns from date fields.

Import path: fyron.cohort.tables.build_survival_columns

python

build_survival_columns(df: pd.DataFrame, *, start_col: str, end_col: str, event_col: str, time_col: str = 'time', event_out_col: str = 'event') -> pd.DataFrame

Parameters

Parameter	Required	Type	Default	Description
`df`	yes	`pd.DataFrame`		Cohort table containing start, end, and event columns.
`start_col`	yes	`str`		Baseline or index-date column where follow-up begins.
`end_col`	yes	`str`		Endpoint or censoring column where follow-up ends.
`event_col`	yes	`str`		Existing column indicating whether the event occurred.
`time_col`	no	`str`	`'time'`	Name of the output duration column.
`event_out_col`	no	`str`	`'event'`	Name of the output event-indicator column.

Returns

pandas.DataFrame - Copy of `df with duration and event columns appended. Negative durations raise a ValueError`.

See also: Cohort And Analysis module guide.

`fyron.cohort.tables.build_time_to_event_endpoint`

Build time-to-event columns from index, event, and censor dates.

Import path: fyron.cohort.tables.build_time_to_event_endpoint

python

build_time_to_event_endpoint(df: pd.DataFrame, *, index_date_col: str, event_date_col: str, censor_date_col: str, time_col: str = 'time', event_col: str = 'event', endpoint_date_col: str = 'endpoint_date', endpoint_source_col: str = 'endpoint_source', time_unit: TimeUnit = 'days', require_endpoint: bool = True) -> pd.DataFrame

Parameters

Parameter	Required	Type	Default	Description
`df`	yes	`pd.DataFrame`		Cohort table with one row per patient or analysis unit.
`index_date_col`	yes	`str`		Start date for follow-up, such as diagnosis, baseline CT, or treatment start.
`event_date_col`	yes	`str`		Date of the event of interest. Missing values are treated as censored.
`censor_date_col`	yes	`str`		Last known follow-up or administrative censoring date.
`time_col`	no	`str`	`'time'`	Name of the output follow-up duration column.
`event_col`	no	`str`	`'event'`	Name of the output binary event-indicator column.
`endpoint_date_col`	no	`str`	`'endpoint_date'`	Name of the output column storing the selected event or censoring date.
`endpoint_source_col`	no	`str`	`'endpoint_source'`	Name of the output column storing ``"event",` `"censor", or` `"missing"``.
`time_unit`	no	`TimeUnit`	`'days'`	Unit for the computed follow-up time: ``"days",` `"weeks",` `"months", or` `"years"``.
`require_endpoint`	no	`bool`	`True`	If True, rows without both event and censor dates raise an error.

Returns

pandas.DataFrame - Copy of `df` with time, event, endpoint date, and endpoint source columns.

See also: Cohort And Analysis module guide.

`fyron.cohort.tables.cohort_profile`

Create a compact audit table for a patient-level cohort.

Import path: fyron.cohort.tables.cohort_profile

python

cohort_profile(df: pd.DataFrame, *, id_col: str = 'patient_id', endpoint_cols: Sequence[str] | None = None, group_col: str | None = None) -> pd.DataFrame

Parameters

Parameter	Required	Type	Default	Description
`df`	yes	`pd.DataFrame`		Cohort table to profile.
`id_col`	no	`str`	`'patient_id'`	Patient or analysis-unit identifier.
`endpoint_cols`	no	`Sequence[str] or None`	`None`	Optional binary or numeric endpoint columns to summarize.
`group_col`	no	`str or None`	`None`	Optional subgroup column for distribution checks.

Returns

pandas.DataFrame - Rows named by `metric` with values useful for cohort review.

See also: Cohort And Analysis module guide.

`fyron.cohort.tables.join_cohort_tables`

Join feature and outcome tables on a patient identifier.

Import path: fyron.cohort.tables.join_cohort_tables

python

join_cohort_tables(features: pd.DataFrame, outcomes: pd.DataFrame, *, on: str = 'patient_id', how: str = 'inner', validate: str = 'one_to_one') -> pd.DataFrame

Parameters

Parameter	Required	Type	Default	Description
`features`	yes	`pd.DataFrame`		Patient-level predictors (must include ``on``).
`outcomes`	yes	`pd.DataFrame`		Patient-level outcomes (must include ``on``).
`on`	no	`str`	`'patient_id'`	Join key column name.
`how`	no	`str`	`'inner'`	Merge strategy passed to :meth:`pandas.DataFrame.merge`.
`validate`	no	`str`	`'one_to_one'`	Merge validation (e.g. ``one_to_one``) to catch duplicate keys.

Returns

pd.DataFrame

See also: Cohort And Analysis module guide.

`fyron.cohort.tables.validate_cohort_table`

Validate a cohort DataFrame for analysis-ready use.

Import path: fyron.cohort.tables.validate_cohort_table

python

validate_cohort_table(df: pd.DataFrame, *, required_columns: Sequence[str], id_column: str | None = 'patient_id', allow_duplicate_ids: bool = False) -> pd.DataFrame

Parameters

Parameter	Required	Type	Default	Description
`df`	yes	`pd.DataFrame`		Cohort table with one row per patient, lesion, study, or other declared analysis unit.
`required_columns`	yes	`Sequence[str]`		Column names that must be present before the cohort can be used downstream.
`id_column`	no	`str or None`	`'patient_id'`	Optional identifier column used to detect missing or duplicate analysis units. Pass ``None`` to skip identifier checks.
`allow_duplicate_ids`	no	`bool`	`False`	If False, duplicate values in ``id_column` `raise a` `ValueError``.

Returns

pandas.DataFrame - The original input DataFrame, unchanged, when all validation checks pass.

See also: Cohort And Analysis module guide.

`fyron.dataframe.add_code_indicators`

Add binary indicator columns for selected clinical codes.

Import path: fyron.dataframe.add_code_indicators

python

add_code_indicators(df: pd.DataFrame, code_col: str, codes: Sequence[Any] | Mapping[str, Sequence[Any]], *, sep: str = ';', prefix: str = 'code_', normalize: bool = True) -> pd.DataFrame

Parameters

Parameter	Required	Type	Default	Description
`df`	yes	`pd.DataFrame`		DataFrame containing a code column.
`code_col`	yes	`str`		Column with one or more codes per row.
`codes`	yes	`Sequence[Any] or Mapping[str, Sequence[Any]]`		Either a sequence of codes or a mapping from output names to code lists.
`sep`	no	`str`	`';'`	Delimiter between multiple codes in one cell.
`prefix`	no	`str`	`'code_'`	Prefix used for generated indicator columns.
`normalize`	no	`bool`	`True`	If True, normalize both input and target codes before matching.

Returns

pandas.DataFrame - Copy of `df` with binary indicator columns.

See also: Cohort And Analysis module guide.

`fyron.dataframe.add_date_difference`

Add a date-difference feature from two DataFrame columns.

Import path: fyron.dataframe.add_date_difference

python

add_date_difference(df: pd.DataFrame, start_col: str, end_col: str, output_col: str, *, unit: TimeUnit = 'days', absolute: bool = False, allow_negative: bool = True) -> pd.DataFrame

Parameters

Parameter	Required	Type	Default	Description
`df`	yes	`pd.DataFrame`		DataFrame containing start and end date columns.
`start_col`	yes	`str`		Start date column.
`end_col`	yes	`str`		End date column.
`output_col`	yes	`str`		Name of the created numeric difference column.
`unit`	no	`TimeUnit`	`'days'`	Output unit: ``"days",` `"weeks",` `"months", or` `"years"``.
`absolute`	no	`bool`	`False`	If True, store absolute differences.
`allow_negative`	no	`bool`	`True`	If False, raise when a non-missing negative interval is computed.

Returns

pandas.DataFrame - Copy of `df` with the created difference column.

See also: Cohort And Analysis module guide.

`fyron.dataframe.add_icd10_disease_groups`

Add broad ICD-10 disease-group indicator columns.

Import path: fyron.dataframe.add_icd10_disease_groups

python

add_icd10_disease_groups(df: pd.DataFrame, icd_col: str, *, sep: str = ';', prefix: str = 'icd10_group_', summary_col: str = 'icd10_groups', count_col: str = 'n_icd10_groups', groups: Sequence[str] | None = None) -> pd.DataFrame

Parameters

Parameter	Required	Type	Default	Description
`df`	yes	`pd.DataFrame`		DataFrame containing an ICD-10 code column.
`icd_col`	yes	`str`		Column with one or more ICD-10 codes per row.
`sep`	no	`str`	`';'`	Delimiter between multiple codes in one cell.
`prefix`	no	`str`	`'icd10_group_'`	Prefix for generated binary indicator columns.
`summary_col`	no	`str`	`'icd10_groups'`	Output column containing human-readable disease groups.
`count_col`	no	`str`	`'n_icd10_groups'`	Output column containing the number of disease groups per row.
`groups`	no	`Sequence[str] or None`	`None`	Optional explicit disease-group order. Defaults to Fyron's built-in broad ICD-10 groups.

Returns

pandas.DataFrame - Copy of `df` with binary disease-group indicators and summary columns.

See also: Cohort And Analysis module guide.

`fyron.dataframe.add_ops_categories`

Add broad German OPS procedure-category indicator columns.

Import path: fyron.dataframe.add_ops_categories

python

add_ops_categories(df: pd.DataFrame, ops_col: str, *, sep: str = ';', prefix: str = 'ops_category_', summary_col: str = 'ops_categories', count_col: str = 'n_ops_categories', categories: Sequence[str] | None = None) -> pd.DataFrame

Parameters

Parameter	Required	Type	Default	Description
`df`	yes	`pd.DataFrame`		DataFrame containing an OPS code column.
`ops_col`	yes	`str`		Column with one or more OPS codes per row.
`sep`	no	`str`	`';'`	Delimiter between multiple codes in one cell.
`prefix`	no	`str`	`'ops_category_'`	Prefix for generated binary indicator columns.
`summary_col`	no	`str`	`'ops_categories'`	Output column containing human-readable OPS categories.
`count_col`	no	`str`	`'n_ops_categories'`	Output column containing the number of OPS categories per row.
`categories`	no	`Sequence[str] or None`	`None`	Optional explicit category order. Defaults to Fyron's broad OPS groups.

Returns

pandas.DataFrame - Copy of `df` with binary OPS category indicators and summary columns.

See also: Cohort And Analysis module guide.

`fyron.dataframe.add_time_window_flag`

Flag rows where a date falls within a window around an index date.

Import path: fyron.dataframe.add_time_window_flag

python

add_time_window_flag(df: pd.DataFrame, date_col: str, index_date_col: str, output_col: str, *, before: float | None = None, after: float | None = None, unit: TimeUnit = 'days') -> pd.DataFrame

Parameters

Parameter	Required	Type	Default	Description
`df`	yes	`pd.DataFrame`		DataFrame containing event and index date columns.
`date_col`	yes	`str`		Date to test against the index date.
`index_date_col`	yes	`str`		Index/reference date column.
`output_col`	yes	`str`		Name of the created binary flag column.
`before`	no	`float or None`	`None`	Maximum time before the index date allowed in the window.
`after`	no	`float or None`	`None`	Maximum time after the index date allowed in the window.
`unit`	no	`TimeUnit`	`'days'`	Window unit: ``"days",` `"weeks",` `"months", or` `"years"``.

Returns

pandas.DataFrame - Copy of `df` with an integer 0/1 window flag column.

See also: Cohort And Analysis module guide.

`fyron.dataframe.age_at_date`

Add age in years at a reference date.

Import path: fyron.dataframe.age_at_date

python

age_at_date(df: pd.DataFrame, birth_date_col: str, reference_date_col: str, *, output_col: str = 'age_years') -> pd.DataFrame

Parameters

Parameter	Required	Type	Default	Description
`df`	yes	`pd.DataFrame`		DataFrame containing birth and reference dates.
`birth_date_col`	yes	`str`		Date-of-birth column.
`reference_date_col`	yes	`str`		Date column where age should be calculated, such as diagnosis or index date.
`output_col`	no	`str`	`'age_years'`	Name of the created age column.

Returns

pandas.DataFrame - Copy of `df` with age in years.

See also: Cohort And Analysis module guide.

`fyron.dataframe.classify_icd10`

Classify one ICD-10 code into a broad disease group.

Import path: fyron.dataframe.classify_icd10

python

classify_icd10(code: Any) -> str | None

Parameters

Parameter	Required	Type	Default	Description
`code`	yes	`Any`		ICD-10 code such as ``"C34",` `"I50.1", or` `"Z51"``.

Returns

str or None - Broad disease group name, `"Other"` for malformed non-empty codes, or None for missing values.

See also: Cohort And Analysis module guide.

`fyron.dataframe.classify_icd10_string`

Classify a delimited ICD-10 code string into disease groups.

Import path: fyron.dataframe.classify_icd10_string

python

classify_icd10_string(value: Any, *, sep: str = ';') -> set[str]

Parameters

Parameter	Required	Type	Default	Description
`value`	yes	`Any`		Scalar cell containing one or more ICD-10 codes.
`sep`	no	`str`	`';'`	Delimiter between codes.

Returns

set[str] - Disease groups present in the code string.

See also: Cohort And Analysis module guide.

`fyron.dataframe.classify_ops`

Classify one German OPS code into a broad procedure category.

Import path: fyron.dataframe.classify_ops

python

classify_ops(code: Any) -> str | None

Parameters

Parameter	Required	Type	Default	Description
`code`	yes	`Any`		OPS code such as ``"1-620",` `"3-20",` `"5-541", or` `"8-980"``.

Returns

str or None - Broad OPS category, `"Other"` for malformed non-empty codes, or None for missing values.

See also: Cohort And Analysis module guide.

`fyron.dataframe.classify_ops_string`

Classify a delimited OPS code string into broad procedure categories.

Import path: fyron.dataframe.classify_ops_string

python

classify_ops_string(value: Any, *, sep: str = ';') -> set[str]

Parameters

Parameter	Required	Type	Default	Description
`value`	yes	`Any`		Scalar cell containing one or more OPS codes.
`sep`	no	`str`	`';'`	Delimiter between codes.

Returns

set[str] - OPS categories present in the code string.

See also: Cohort And Analysis module guide.

`fyron.dataframe.date_difference`

Calculate a date difference in days, weeks, months, or years.

Import path: fyron.dataframe.date_difference

python

date_difference(start: Any, end: Any, *, unit: TimeUnit = 'days', absolute: bool = False) -> pd.Series

Parameters

Parameter	Required	Type	Default	Description
`start`	yes	`Any`		Start date scalar or array-like value.
`end`	yes	`Any`		End date scalar or array-like value.
`unit`	no	`TimeUnit`	`'days'`	Output unit: ``"days",` `"weeks",` `"months", or` `"years"``.
`absolute`	no	`bool`	`False`	If True, return absolute differences.

Returns

pandas.Series - Numeric difference in the requested unit.

See also: Cohort And Analysis module guide.

`fyron.dataframe.explode_code_column`

Explode a delimited clinical code column into one row per code.

Import path: fyron.dataframe.explode_code_column

python

explode_code_column(df: pd.DataFrame, code_col: str, *, sep: str = ';', output_col: str = 'code', keep_empty: bool = False) -> pd.DataFrame

Parameters

Parameter	Required	Type	Default	Description
`df`	yes	`pd.DataFrame`		DataFrame containing a delimited code column.
`code_col`	yes	`str`		Column with one or more codes per row.
`sep`	no	`str`	`';'`	Delimiter between codes.
`output_col`	no	`str`	`'code'`	Name of the exploded output code column.
`keep_empty`	no	`bool`	`False`	If True, keep rows without codes and set ``output_col`` to missing.

Returns

pandas.DataFrame - Exploded DataFrame with one row per parsed code.

See also: Cohort And Analysis module guide.

`fyron.dataframe.icd10_group_prevalence`

Summarize ICD-10 disease-group indicator prevalence.

Import path: fyron.dataframe.icd10_group_prevalence

python

icd10_group_prevalence(df: pd.DataFrame, *, group_cols: Sequence[str] | None = None, prefix: str = 'icd10_group_') -> pd.DataFrame

Parameters

Parameter	Required	Type	Default	Description
`df`	yes	`pd.DataFrame`		DataFrame with ICD-10 group indicator columns.
`group_cols`	no	`Sequence[str] or None`	`None`	Optional explicit indicator columns. If omitted, columns with ``prefix`` are used.
`prefix`	no	`str`	`'icd10_group_'`	Prefix used to discover indicator columns.

Returns

pandas.DataFrame - Prevalence table with group, column, patient count, and percent.

See also: Cohort And Analysis module guide.

`fyron.dataframe.normalize_code`

Normalize one clinical code value for matching.

Import path: fyron.dataframe.normalize_code

python

normalize_code(value: Any, *, uppercase: bool = True, remove_spaces: bool = True, remove_dots: bool = False) -> str | None

Parameters

Parameter	Required	Type	Default	Description
`value`	yes	`Any`		Raw code value such as an ICD-10, OPS, medication, or local registry code.
`uppercase`	no	`bool`	`True`	If True, convert the returned code to uppercase.
`remove_spaces`	no	`bool`	`True`	If True, remove whitespace within the code.
`remove_dots`	no	`bool`	`False`	If True, remove dots from the code.

Returns

str or None - Normalized code string, or None for missing/empty values.

See also: Cohort And Analysis module guide.

`fyron.dataframe.ops_category_prevalence`

Summarize OPS procedure-category indicator prevalence.

Import path: fyron.dataframe.ops_category_prevalence

python

ops_category_prevalence(df: pd.DataFrame, *, category_cols: Sequence[str] | None = None, prefix: str = 'ops_category_') -> pd.DataFrame

Parameters

Parameter	Required	Type	Default	Description
`df`	yes	`pd.DataFrame`		DataFrame with OPS category indicator columns.
`category_cols`	no	`Sequence[str] or None`	`None`	Optional explicit indicator columns. If omitted, columns with ``prefix`` are used.
`prefix`	no	`str`	`'ops_category_'`	Prefix used to discover indicator columns.

Returns

pandas.DataFrame - Prevalence table with category, column, patient count, and percent.

See also: Cohort And Analysis module guide.

`fyron.dataframe.split_code_string`

Split one cell containing one or more clinical codes.

Import path: fyron.dataframe.split_code_string

python

split_code_string(value: Any, *, sep: str = ';', normalize: bool = True) -> list[str]

Parameters

Parameter	Required	Type	Default	Description
`value`	yes	`Any`		Scalar value containing one or more codes, usually separated by semicolons.
`sep`	no	`str`	`';'`	Delimiter used between codes.
`normalize`	no	`bool`	`True`	If True, normalize each code with :func:`normalize_code`.

Returns

list[str] - Ordered list of non-empty codes.

See also: Cohort And Analysis module guide.

`fyron.descriptive.analysis.compare_descriptive_groups`

Compare descriptive variables across groups with explicit statistical tests.

Import path: fyron.descriptive.analysis.compare_descriptive_groups

python

compare_descriptive_groups(df: pd.DataFrame, *, group_col: str, numeric: Sequence[str] | None = None, categorical: Sequence[str] | None = None, include_smd: bool = True) -> pd.DataFrame

Parameters

Parameter	Required	Type	Default	Description
`df`	yes	`pd.DataFrame`		Cohort or analysis table.
`group_col`	yes	`str`		Column defining comparison groups.
`numeric`	no	`Sequence[str] or None`	`None`	Numeric variables to compare. If omitted, numeric columns are selected automatically.
`categorical`	no	`Sequence[str] or None`	`None`	Categorical variables to compare. If omitted, remaining non-numeric columns are selected automatically.
`include_smd`	no	`bool`	`True`	If True, include standardized mean difference summaries.

Returns

pandas.DataFrame - One row per variable with variable type, test result fields, p-value, optional SMD, and error text when a comparison cannot be computed.

See also: Cohort And Analysis module guide.

`fyron.descriptive.analysis.create_table_one`

Create a table-one style long table with optional p-values.

Import path: fyron.descriptive.analysis.create_table_one

python

create_table_one(df: pd.DataFrame, *, group_col: str | None = None, numeric: Sequence[str] | None = None, categorical: Sequence[str] | None = None, include_smd: bool = True) -> pd.DataFrame

Parameters

Parameter	Required	Type	Default	Description
`df`	yes	`pd.DataFrame`		Cohort table to describe.
`group_col`	no	`str or None`	`None`	Optional group column such as treatment arm, cohort, split, or outcome.
`numeric`	no	`Sequence[str] or None`	`None`	Numeric variables to include. If omitted, numeric columns are selected automatically.
`categorical`	no	`Sequence[str] or None`	`None`	Categorical variables to include. If omitted, non-numeric columns are selected automatically.
`include_smd`	no	`bool`	`True`	If True and ``group_col`` is provided, include standardized mean difference values.

Returns

pandas.DataFrame - Long Table 1 style output with variable, level, group, formatted summary, and optional p-value, SMD, and test columns.

See also: Cohort And Analysis module guide.

`fyron.descriptive.analysis.describe_categorical`

Describe categorical variables with counts and percentages.

Import path: fyron.descriptive.analysis.describe_categorical

python

describe_categorical(df: pd.DataFrame, *, columns: Sequence[str] | None = None, group_col: str | None = None, include_missing: bool = True, digits: int = 1) -> pd.DataFrame

Parameters

Parameter	Required	Type	Default	Description
`df`	yes	`pd.DataFrame`		Cohort or analysis table.
`columns`	no	`Sequence[str] or None`	`None`	Categorical columns to summarize. If omitted, non-numeric columns are selected automatically.
`group_col`	no	`str or None`	`None`	Optional grouping column for stratified summaries.
`include_missing`	no	`bool`	`True`	Include missing values as an explicit ``"<missing>"`` level.
`digits`	no	`int`	`1`	Decimal places used for percentages.

Returns

pandas.DataFrame - Long table with group, variable, level, count, and percent columns.

See also: Cohort And Analysis module guide.

`fyron.descriptive.analysis.describe_numeric`

Describe numeric variables with n, missing, mean, SD, median, IQR, and range.

Import path: fyron.descriptive.analysis.describe_numeric

python

describe_numeric(df: pd.DataFrame, *, columns: Sequence[str] | None = None, group_col: str | None = None, digits: int = 2) -> pd.DataFrame

Parameters

Parameter	Required	Type	Default	Description
`df`	yes	`pd.DataFrame`		Cohort or analysis table.
`columns`	no	`Sequence[str] or None`	`None`	Numeric columns to summarize. If omitted, numeric columns are selected automatically.
`group_col`	no	`str or None`	`None`	Optional grouping column for stratified summaries.
`digits`	no	`int`	`2`	Decimal places used for rounded numeric summaries.

Returns

pandas.DataFrame - Long table with one row per variable and group containing count, missingness, mean, standard deviation, median, quartiles, and range.

See also: Cohort And Analysis module guide.

`fyron.descriptive.analysis.summarize_endpoint`

Summarize endpoint counts, event rates, and optional follow-up time.

Import path: fyron.descriptive.analysis.summarize_endpoint

python

summarize_endpoint(df: pd.DataFrame, *, event_col: str, time_col: str | None = None, group_col: str | None = None, digits: int = 2) -> pd.DataFrame

Parameters

Parameter	Required	Type	Default	Description
`df`	yes	`pd.DataFrame`		Cohort table containing endpoint and optional follow-up columns.
`event_col`	yes	`str`		Binary or numeric event-indicator column where nonzero values count as events.
`time_col`	no	`str or None`	`None`	Optional duration/follow-up column to summarize.
`group_col`	no	`str or None`	`None`	Optional grouping column for stratified endpoint summaries.
`digits`	no	`int`	`2`	Decimal places used for percentages and follow-up summaries.

Returns

pandas.DataFrame - One row per group with sample size, event count, event rate, and optional median/min/max follow-up.

See also: Cohort And Analysis module guide.

`fyron.phenotyping.rules.combine_phenotypes`

Combine patient-level phenotype tables with an any/all rule.

Import path: fyron.phenotyping.rules.combine_phenotypes

python

combine_phenotypes(tables: Sequence[pd.DataFrame], *, patient_col: str = 'patient_id', strategy: CombineStrategy = 'any', phenotype_name: str = 'combined') -> pd.DataFrame

Parameters

Parameter	Required	Type	Default	Description
`tables`	yes	`Sequence[pd.DataFrame]`		See signature.
`patient_col`	no	`str`	`'patient_id'`	See signature.
`strategy`	no	`CombineStrategy`	`'any'`	See signature.
`phenotype_name`	no	`str`	`'combined'`	See signature.

Returns

pd.DataFrame

See also: Cohort And Analysis module guide.

`fyron.phenotyping.rules.define_code_phenotype`

Define a phenotype from diagnosis/procedure/code rows.

Import path: fyron.phenotyping.rules.define_code_phenotype

python

define_code_phenotype(df: pd.DataFrame, code_col: str, codes: Sequence[Any], *, patient_col: str = 'patient_id', date_col: str | None = None, phenotype_name: str = 'phenotype', code_system_col: str | None = None, code_systems: Sequence[Any] | None = None) -> pd.DataFrame

Parameters

Parameter	Required	Type	Default	Description
`df`	yes	`pd.DataFrame`		See signature.
`code_col`	yes	`str`		See signature.
`codes`	yes	`Sequence[Any]`		See signature.
`patient_col`	no	`str`	`'patient_id'`	See signature.
`date_col`	no	`str or None`	`None`	See signature.
`phenotype_name`	no	`str`	`'phenotype'`	See signature.
`code_system_col`	no	`str or None`	`None`	See signature.
`code_systems`	no	`Sequence[Any] or None`	`None`	See signature.

Returns

pd.DataFrame

See also: Cohort And Analysis module guide.

`fyron.phenotyping.rules.define_lab_threshold_phenotype`

Define a phenotype from numeric lab/measurement thresholds.

Import path: fyron.phenotyping.rules.define_lab_threshold_phenotype

python

define_lab_threshold_phenotype(df: pd.DataFrame, value_col: str, threshold: float, *, direction: ThresholdDirection = '>=', patient_col: str = 'patient_id', date_col: str | None = None, phenotype_name: str = 'phenotype') -> pd.DataFrame

Parameters

Parameter	Required	Type	Default	Description
`df`	yes	`pd.DataFrame`		See signature.
`value_col`	yes	`str`		See signature.
`threshold`	yes	`float`		See signature.
`direction`	no	`ThresholdDirection`	`'>='`	See signature.
`patient_col`	no	`str`	`'patient_id'`	See signature.
`date_col`	no	`str or None`	`None`	See signature.
`phenotype_name`	no	`str`	`'phenotype'`	See signature.

Returns

pd.DataFrame

See also: Cohort And Analysis module guide.

`fyron.phenotyping.rules.define_medication_exposure`

Define medication exposure from medication name/code rows.

Import path: fyron.phenotyping.rules.define_medication_exposure

python

define_medication_exposure(df: pd.DataFrame, medication_col: str, medications: Sequence[Any], *, patient_col: str = 'patient_id', start_col: str | None = None, end_col: str | None = None, phenotype_name: str = 'exposure') -> pd.DataFrame

Parameters

Parameter	Required	Type	Default	Description
`df`	yes	`pd.DataFrame`		See signature.
`medication_col`	yes	`str`		See signature.
`medications`	yes	`Sequence[Any]`		See signature.
`patient_col`	no	`str`	`'patient_id'`	See signature.
`start_col`	no	`str or None`	`None`	See signature.
`end_col`	no	`str or None`	`None`	See signature.
`phenotype_name`	no	`str`	`'exposure'`	See signature.

Returns

pd.DataFrame

See also: Cohort And Analysis module guide.

`fyron.phenotyping.rules.phenotype_summary`

Summarize phenotype prevalence overall or by group.

Import path: fyron.phenotyping.rules.phenotype_summary

python

phenotype_summary(df: pd.DataFrame, phenotype_col: str, *, patient_col: str = 'patient_id', group_col: str | None = None) -> pd.DataFrame

Parameters

Parameter	Required	Type	Default	Description
`df`	yes	`pd.DataFrame`		See signature.
`phenotype_col`	yes	`str`		See signature.
`patient_col`	no	`str`	`'patient_id'`	See signature.
`group_col`	no	`str or None`	`None`	See signature.

Returns

pd.DataFrame

See also: Cohort And Analysis module guide.

`fyron.preprocessing.encoding.apply_categorical_encoder`

Apply a fitted Fyron categorical encoder.

Import path: fyron.preprocessing.encoding.apply_categorical_encoder

python

apply_categorical_encoder(df: pd.DataFrame, encoder: dict[str, Any]) -> pd.DataFrame

Parameters

Parameter	Required	Type	Default	Description
`df`	yes	`pd.DataFrame`		See signature.
`encoder`	yes	`dict[str, Any]`		See signature.

Returns

pd.DataFrame

See also: Cohort And Analysis module guide.

`fyron.preprocessing.encoding.encode_categorical`

Fit and apply a categorical encoder to one DataFrame.

Import path: fyron.preprocessing.encoding.encode_categorical

python

encode_categorical(df: pd.DataFrame, **kwargs: Any) -> tuple[pd.DataFrame, dict[str, Any]]

Parameters

Parameter	Required	Type	Default	Description
`df`	yes	`pd.DataFrame`		See signature.
`kwargs`	yes	`Any`		See signature.

Returns

tuple[pd.DataFrame, dict[str, Any]]

See also: Cohort And Analysis module guide.

`fyron.preprocessing.encoding.fit_categorical_encoder`

Fit a deterministic one-hot encoder specification.

Import path: fyron.preprocessing.encoding.fit_categorical_encoder

python

fit_categorical_encoder(df: pd.DataFrame, *, columns: Sequence[str] | None = None, drop_first: bool = False, rare_min_count: int | None = None, rare_label: str = '__rare__') -> dict[str, Any]

Parameters

Parameter	Required	Type	Default	Description
`df`	yes	`pd.DataFrame`		See signature.
`columns`	no	`Sequence[str] or None`	`None`	See signature.
`drop_first`	no	`bool`	`False`	See signature.
`rare_min_count`	no	`int or None`	`None`	See signature.
`rare_label`	no	`str`	`'__rare__'`	See signature.

Returns

dict[str, Any]

See also: Cohort And Analysis module guide.

`fyron.preprocessing.encoding.summarize_encoding_plan`

Summarize categorical levels before one-hot encoding.

Import path: fyron.preprocessing.encoding.summarize_encoding_plan

python

summarize_encoding_plan(df: pd.DataFrame, *, columns: Sequence[str] | None = None, max_levels: int = 20) -> pd.DataFrame

Parameters

Parameter	Required	Type	Default	Description
`df`	yes	`pd.DataFrame`		See signature.
`columns`	no	`Sequence[str] or None`	`None`	See signature.
`max_levels`	no	`int`	`20`	See signature.

Returns

pd.DataFrame

See also: Cohort And Analysis module guide.

`fyron.preprocessing.imputation.apply_imputer`

Apply a fitted Fyron imputer to a DataFrame.

Import path: fyron.preprocessing.imputation.apply_imputer

python

apply_imputer(df: pd.DataFrame, imputer: dict[str, Any]) -> pd.DataFrame

Parameters

Parameter	Required	Type	Default	Description
`df`	yes	`pd.DataFrame`		See signature.
`imputer`	yes	`dict[str, Any]`		See signature.

Returns

pd.DataFrame

See also: Cohort And Analysis module guide.

`fyron.preprocessing.imputation.fit_imputer`

Fit a lightweight imputation specification on a training DataFrame.

Import path: fyron.preprocessing.imputation.fit_imputer

python

fit_imputer(df: pd.DataFrame, *, numeric: Sequence[str] | None = None, categorical: Sequence[str] | None = None, strategy_numeric: NumericStrategy = 'median', strategy_categorical: CategoricalStrategy = 'most_frequent', fill_value_numeric: float = 0.0, fill_value_categorical: str = '__missing__', add_missing_indicators: bool = False) -> dict[str, Any]

Parameters

Parameter	Required	Type	Default	Description
`df`	yes	`pd.DataFrame`		See signature.
`numeric`	no	`Sequence[str] or None`	`None`	See signature.
`categorical`	no	`Sequence[str] or None`	`None`	See signature.
`strategy_numeric`	no	`NumericStrategy`	`'median'`	See signature.
`strategy_categorical`	no	`CategoricalStrategy`	`'most_frequent'`	See signature.
`fill_value_numeric`	no	`float`	`0.0`	See signature.
`fill_value_categorical`	no	`str`	`'__missing__'`	See signature.
`add_missing_indicators`	no	`bool`	`False`	See signature.

Returns

dict[str, Any]

See also: Cohort And Analysis module guide.

`fyron.preprocessing.imputation.impute_dataframe`

Fit and apply an imputer to one DataFrame.

Import path: fyron.preprocessing.imputation.impute_dataframe

python

impute_dataframe(df: pd.DataFrame, **kwargs: Any) -> tuple[pd.DataFrame, dict[str, Any]]

Parameters

Parameter	Required	Type	Default	Description
`df`	yes	`pd.DataFrame`		See signature.
`kwargs`	yes	`Any`		See signature.

Returns

tuple[pd.DataFrame, dict[str, Any]]

See also: Cohort And Analysis module guide.

`fyron.preprocessing.imputation.summarize_imputation_plan`

Summarize columns and strategies before fitting an imputer.

Import path: fyron.preprocessing.imputation.summarize_imputation_plan

python

summarize_imputation_plan(df: pd.DataFrame, *, numeric: Sequence[str] | None = None, categorical: Sequence[str] | None = None, strategy_numeric: NumericStrategy = 'median', strategy_categorical: CategoricalStrategy = 'most_frequent') -> pd.DataFrame

Parameters

Parameter	Required	Type	Default	Description
`df`	yes	`pd.DataFrame`		See signature.
`numeric`	no	`Sequence[str] or None`	`None`	See signature.
`categorical`	no	`Sequence[str] or None`	`None`	See signature.
`strategy_numeric`	no	`NumericStrategy`	`'median'`	See signature.
`strategy_categorical`	no	`CategoricalStrategy`	`'most_frequent'`	See signature.

Returns

pd.DataFrame

See also: Cohort And Analysis module guide.

`fyron.preprocessing.qc.compare_train_test_balance`

Compare simple train/test distributions for leakage and split sanity checks.

Import path: fyron.preprocessing.qc.compare_train_test_balance

python

compare_train_test_balance(train: pd.DataFrame, test: pd.DataFrame, *, columns: Sequence[str] | None = None) -> pd.DataFrame

Parameters

Parameter	Required	Type	Default	Description
`train`	yes	`pd.DataFrame`		See signature.
`test`	yes	`pd.DataFrame`		See signature.
`columns`	no	`Sequence[str] or None`	`None`	See signature.

Returns

pd.DataFrame

See also: Cohort And Analysis module guide.

`fyron.preprocessing.qc.detect_potential_leakage`

Flag column names that may encode outcomes, labels, or post-index information.

Import path: fyron.preprocessing.qc.detect_potential_leakage

python

detect_potential_leakage(columns: Sequence[str] | pd.DataFrame, *, outcome_keywords: Sequence[str] = ('event', 'death', 'survival', 'outcome', 'label', 'target'), time_keywords: Sequence[str] = ('followup', 'follow_up', 'last', 'end_date', 'post', 'af...) -> pd.DataFrame

Parameters

Parameter	Required	Type	Default	Description
`columns`	yes	`Sequence[str] or pd.DataFrame`		See signature.
`outcome_keywords`	no	`Sequence[str]`	`('event', 'death', 'survival', 'outcome', 'label', 'target')`	See signature.
`time_keywords`	no	`Sequence[str]`	`('followup', 'follow_up', 'last', 'end_date', 'post', 'af...`	See signature.

Returns

pd.DataFrame

See also: Cohort And Analysis module guide.

`fyron.preprocessing.qc.feature_matrix_report`

Summarize feature-matrix quality before modeling.

Import path: fyron.preprocessing.qc.feature_matrix_report

python

feature_matrix_report(X: pd.DataFrame, y: Sequence[Any] | pd.Series | None = None, *, rare_threshold: float = 0.01, leakage_keywords: Sequence[str] = ('event', 'death', 'survival', 'outcome', 'label', 'target')) -> pd.DataFrame

Parameters

Parameter	Required	Type	Default	Description
`X`	yes	`pd.DataFrame`		Feature matrix with one column per candidate predictor.
`y`	no	`Sequence[Any] or pd.Series or None`	`None`	Optional labels used only for target count metadata.
`rare_threshold`	no	`float`	`0.01`	Binary features with a minority-class frequency below this value are flagged.
`leakage_keywords`	no	`Sequence[str]`	`('event', 'death', 'survival', 'outcome', 'label', 'target')`	Column-name fragments that suggest outcome or post-index leakage.

Returns

pandas.DataFrame - One row per feature with dtype, missingness, uniqueness, constant, rare-binary, and leakage-name flags.

See also: Cohort And Analysis module guide.

`fyron.preprocessing.qc.find_duplicate_ids`

Return rows whose patient identifier appears more than once.

Import path: fyron.preprocessing.qc.find_duplicate_ids

python

find_duplicate_ids(df: pd.DataFrame, *, id_col: str = 'patient_id') -> pd.DataFrame

Parameters

Parameter	Required	Type	Default	Description
`df`	yes	`pd.DataFrame`		See signature.
`id_col`	no	`str`	`'patient_id'`	See signature.

Returns

pd.DataFrame

See also: Cohort And Analysis module guide.

`fyron.preprocessing.qc.summarize_categorical_levels`

Summarize counts and percentages for categorical columns.

Import path: fyron.preprocessing.qc.summarize_categorical_levels

python

summarize_categorical_levels(df: pd.DataFrame, *, columns: Sequence[str] | None = None, max_levels: int = 20, include_missing: bool = True) -> pd.DataFrame

Parameters

Parameter	Required	Type	Default	Description
`df`	yes	`pd.DataFrame`		See signature.
`columns`	no	`Sequence[str] or None`	`None`	See signature.
`max_levels`	no	`int`	`20`	See signature.
`include_missing`	no	`bool`	`True`	See signature.

Returns

pd.DataFrame

See also: Cohort And Analysis module guide.

`fyron.preprocessing.qc.summarize_missingness`

Summarize missing values by column, optionally within groups.

Import path: fyron.preprocessing.qc.summarize_missingness

python

summarize_missingness(df: pd.DataFrame, *, columns: Sequence[str] | None = None, group_col: str | None = None) -> pd.DataFrame

Parameters

Parameter	Required	Type	Default	Description
`df`	yes	`pd.DataFrame`		See signature.
`columns`	no	`Sequence[str] or None`	`None`	See signature.
`group_col`	no	`str or None`	`None`	See signature.

Returns

pd.DataFrame

See also: Cohort And Analysis module guide.

`fyron.statistics.bootstrap.BootstrapResult`

Result for one bootstrapped statistic.

Import path: fyron.statistics.bootstrap.BootstrapResult

python

BootstrapResult()

Parameters

No parameters documented.

Returns

Class constructor.

See also: Cohort And Analysis module guide.

Public Methods

`fyron.statistics.bootstrap.BootstrapResult.to_dict`

See the signature and linked module guide for context.

Import path: fyron.statistics.bootstrap.BootstrapResult.to_dict

python

BootstrapResult.to_dict(self, *, include_distribution: bool = False) -> dict[str, Any]

Parameters

Parameter	Required	Type	Default	Description
`include_distribution`	no	`bool`	`False`	See signature.

Returns

dict[str, Any]

See also: Cohort And Analysis module guide.

`fyron.statistics.bootstrap.bootstrap_binary_classification_metrics`

Bootstrap common binary classification metrics.

Import path: fyron.statistics.bootstrap.bootstrap_binary_classification_metrics

python

bootstrap_binary_classification_metrics(y_true: Sequence[Any], y_pred: Sequence[Any], y_prob: Sequence[Any] | None = None, *, metrics: Sequence[str] | None = None, pos_label: Any = 1, n_rounds: int = 1000, alpha: float = 0.05, random_state: int | None = None, drop_invalid: bool = True) -> pd.DataFrame

Parameters

Parameter	Required	Type	Default	Description
`y_true`	yes	`Sequence[Any]`		See signature.
`y_pred`	yes	`Sequence[Any]`		See signature.
`y_prob`	no	`Sequence[Any] or None`	`None`	See signature.
`metrics`	no	`Sequence[str] or None`	`None`	See signature.
`pos_label`	no	`Any`	`1`	See signature.
`n_rounds`	no	`int`	`1000`	See signature.
`alpha`	no	`float`	`0.05`	See signature.
`random_state`	no	`int or None`	`None`	See signature.
`drop_invalid`	no	`bool`	`True`	See signature.

Returns

pd.DataFrame

See also: Cohort And Analysis module guide.

`fyron.statistics.bootstrap.bootstrap_metric`

Bootstrap an arbitrary metric over one or more aligned arrays.

Import path: fyron.statistics.bootstrap.bootstrap_metric

python

bootstrap_metric(*arrays: Sequence[Any], metric_func: MetricFunc, metric_name: str | None = None, n_rounds: int = 1000, alpha: float = 0.05, random_state: int | None = None, drop_invalid: bool = True, **metric_kwargs: Any) -> BootstrapResult

Parameters

Parameter	Required	Type	Default	Description
`arrays`	yes	`Sequence[Any]`		See signature.
`metric_func`	yes	`MetricFunc`		See signature.
`metric_name`	no	`str or None`	`None`	See signature.
`n_rounds`	no	`int`	`1000`	See signature.
`alpha`	no	`float`	`0.05`	See signature.
`random_state`	no	`int or None`	`None`	See signature.
`drop_invalid`	no	`bool`	`True`	See signature.
`metric_kwargs`	yes	`Any`		See signature.

Returns

BootstrapResult

See also: Cohort And Analysis module guide.

`fyron.statistics.bootstrap.bootstrap_summary`

Bootstrap a one-sample summary statistic such as mean or median.

Import path: fyron.statistics.bootstrap.bootstrap_summary

python

bootstrap_summary(values: Sequence[Any], *, statistic: SummaryStatistic | Callable[[Any], float] = 'mean', n_rounds: int = 1000, alpha: float = 0.05, random_state: int | None = None) -> BootstrapResult

Parameters

Parameter	Required	Type	Default	Description
`values`	yes	`Sequence[Any]`		See signature.
`statistic`	no	`SummaryStatistic or Callable[[Any], float]`	`'mean'`	See signature.
`n_rounds`	no	`int`	`1000`	See signature.
`alpha`	no	`float`	`0.05`	See signature.
`random_state`	no	`int or None`	`None`	See signature.

Returns

BootstrapResult

See also: Cohort And Analysis module guide.

`fyron.statistics.comparison.bootstrap_auc_difference`

Bootstrap the paired AUC difference between two models.

Import path: fyron.statistics.comparison.bootstrap_auc_difference

python

bootstrap_auc_difference(y_true: Sequence[Any], prob_a: Sequence[Any], prob_b: Sequence[Any], *, pos_label: Any = 1, n_rounds: int = 1000, alpha: float = 0.05, random_state: int | None = None, drop_invalid: bool = True) -> pd.DataFrame

Parameters

Parameter	Required	Type	Default	Description
`y_true`	yes	`Sequence[Any]`		Ground-truth binary labels.
`prob_a`	yes	`Sequence[Any]`		Predicted probabilities or scores from model A.
`prob_b`	yes	`Sequence[Any]`		Predicted probabilities or scores from model B on the same samples.
`pos_label`	no	`Any`	`1`	Label treated as the positive class.
`n_rounds`	no	`int`	`1000`	Number of bootstrap resamples.
`alpha`	no	`float`	`0.05`	Two-sided confidence interval width, where ``0.05`` gives 95% intervals.
`random_state`	no	`int or None`	`None`	Optional random seed for reproducible resampling.
`drop_invalid`	no	`bool`	`True`	If True, skip bootstrap rounds that contain only one class.

Returns

pandas.DataFrame - One-row table with AUC for both models, paired AUC difference, bootstrap confidence interval, requested rounds, valid rounds, dropped rounds, and alpha.

See also: Cohort And Analysis module guide.

`fyron.statistics.comparison.mcnemar_test`

Run McNemar's paired test for two classifiers.

Import path: fyron.statistics.comparison.mcnemar_test

python

mcnemar_test(y_true: Sequence[Any], pred_a: Sequence[Any], pred_b: Sequence[Any], *, exact: bool = False) -> pd.DataFrame

Parameters

Parameter	Required	Type	Default	Description
`y_true`	yes	`Sequence[Any]`		Ground-truth class labels.
`pred_a`	yes	`Sequence[Any]`		Predicted class labels from model A.
`pred_b`	yes	`Sequence[Any]`		Predicted class labels from model B on the same samples.
`exact`	no	`bool`	`False`	If True, use the exact binomial version; otherwise use the continuity corrected chi-square approximation.

Returns

pandas.DataFrame - One-row table with the test name, statistic, p-value, total sample size, discordant-pair counts, and exact-test flag.

See also: Cohort And Analysis module guide.

`fyron.statistics.correction.adjust_p_values`

Adjust p-values for multiple testing and return a review table.

Import path: fyron.statistics.correction.adjust_p_values

python

adjust_p_values(p_values: Sequence[Any], *, method: PValueCorrectionMethod = 'fdr_bh', alpha: float = 0.05) -> pd.DataFrame

Parameters

Parameter	Required	Type	Default	Description
`p_values`	yes	`Sequence[Any]`		Sequence of raw p-values. Missing values are preserved.
`method`	no	`PValueCorrectionMethod`	`'fdr_bh'`	Correction method: ``"fdr_bh"/Benjamini-Hochberg,` `"bonferroni",` `"holm", or` `"none"``.
`alpha`	no	`float`	`0.05`	Significance level used to compute the rejected flag.

Returns

pandas.DataFrame - One row per p-value with raw p-value, adjusted p-value, rank, rejected flag, correction method, and alpha.

See also: Cohort And Analysis module guide.

`fyron.statistics.correction.p_value_correction_table`

Append adjusted p-values and rejection flags to a result DataFrame.

Import path: fyron.statistics.correction.p_value_correction_table

python

p_value_correction_table(df: pd.DataFrame, *, p_col: str = 'p_value', method: PValueCorrectionMethod = 'fdr_bh', alpha: float = 0.05) -> pd.DataFrame

Parameters

Parameter	Required	Type	Default	Description
`df`	yes	`pd.DataFrame`		Result table containing one raw p-value column, for example a regression, correlation, or feature-screening table.
`p_col`	no	`str`	`'p_value'`	Name of the column containing raw p-values.
`method`	no	`PValueCorrectionMethod`	`'fdr_bh'`	Multiple-testing correction method. Use ``"fdr_bh"` `for Benjamini-Hochberg FDR,` `"bonferroni",` `"holm", or` `"none"``.
`alpha`	no	`float`	`0.05`	Significance threshold used to compute the ``rejected`` flag.

Returns

pandas.DataFrame - Copy of `df with p_adjusted, p_rank, rejected, p_adjust_method, and alpha` columns appended.

See also: Cohort And Analysis module guide.

`fyron.statistics.effects.cliffs_delta`

Return Cliff's delta for ordinal/non-parametric two-group comparisons.

Import path: fyron.statistics.effects.cliffs_delta

python

cliffs_delta(group_a: Sequence[Any], group_b: Sequence[Any]) -> float

Parameters

Parameter	Required	Type	Default	Description
`group_a`	yes	`Sequence[Any]`		See signature.
`group_b`	yes	`Sequence[Any]`		See signature.

Returns

float

See also: Cohort And Analysis module guide.

`fyron.statistics.effects.cohens_d`

Return Cohen's d for two independent numeric groups.

Import path: fyron.statistics.effects.cohens_d

python

cohens_d(group_a: Sequence[Any], group_b: Sequence[Any]) -> float

Parameters

Parameter	Required	Type	Default	Description
`group_a`	yes	`Sequence[Any]`		See signature.
`group_b`	yes	`Sequence[Any]`		See signature.

Returns

float

See also: Cohort And Analysis module guide.

`fyron.statistics.effects.cramers_v`

Return Cramer's V for a contingency table.

Import path: fyron.statistics.effects.cramers_v

python

cramers_v(table: Sequence[Sequence[Any]]) -> float

Parameters

Parameter	Required	Type	Default	Description
`table`	yes	`Sequence[Sequence[Any]]`		See signature.

Returns

float

See also: Cohort And Analysis module guide.

`fyron.statistics.effects.odds_ratio_ci`

Return odds ratio and Wald confidence interval for a 2x2 table.

Import path: fyron.statistics.effects.odds_ratio_ci

python

odds_ratio_ci(table: Sequence[Sequence[Any]], *, alpha: float = 0.05, correction: float = 0.5) -> dict[str, float]

Parameters

Parameter	Required	Type	Default	Description
`table`	yes	`Sequence[Sequence[Any]]`		See signature.
`alpha`	no	`float`	`0.05`	See signature.
`correction`	no	`float`	`0.5`	See signature.

Returns

dict[str, float]

See also: Cohort And Analysis module guide.

`fyron.statistics.effects.risk_ratio_ci`

Return risk ratio and Wald confidence interval for a 2x2 table.

Import path: fyron.statistics.effects.risk_ratio_ci

python

risk_ratio_ci(table: Sequence[Sequence[Any]], *, alpha: float = 0.05, correction: float = 0.5) -> dict[str, float]

Parameters

Parameter	Required	Type	Default	Description
`table`	yes	`Sequence[Sequence[Any]]`		See signature.
`alpha`	no	`float`	`0.05`	See signature.
`correction`	no	`float`	`0.5`	See signature.

Returns

dict[str, float]

See also: Cohort And Analysis module guide.

`fyron.statistics.effects.standardized_mean_difference`

Return SMD for numeric values or binary categorical proportions.

Import path: fyron.statistics.effects.standardized_mean_difference

python

standardized_mean_difference(group_a: Sequence[Any], group_b: Sequence[Any], *, categorical: bool = False) -> float

Parameters

Parameter	Required	Type	Default	Description
`group_a`	yes	`Sequence[Any]`		See signature.
`group_b`	yes	`Sequence[Any]`		See signature.
`categorical`	no	`bool`	`False`	See signature.

Returns

float

See also: Cohort And Analysis module guide.

`fyron.statistics.qc.correlation_pvalue_table`

Return pairwise correlations, p-values, and optional adjusted p-values.

Import path: fyron.statistics.qc.correlation_pvalue_table

python

correlation_pvalue_table(df: pd.DataFrame, columns: Sequence[str], *, method: str = 'spearman', p_adjust: str | None = 'fdr_bh') -> pd.DataFrame

Parameters

Parameter	Required	Type	Default	Description
`df`	yes	`pd.DataFrame`		Analysis table containing numeric variables.
`columns`	yes	`Sequence[str]`		Variables for pairwise correlation testing.
`method`	no	`str`	`'spearman'`	Correlation method, typically ``"spearman"` `or` `"pearson"``.
`p_adjust`	no	`str or None`	`'fdr_bh'`	Optional multiple-testing correction method. Pass ``None`` to keep only raw p-values.

Returns

pandas.DataFrame - Pairwise table with variable names, method, correlation coefficient, p-value, sample size, and optional adjusted p-value columns.

See also: Cohort And Analysis module guide.

`fyron.statistics.qc.missingness_by_group_table`

Summarize missingness overall and by group for manuscript supplements.

Import path: fyron.statistics.qc.missingness_by_group_table

python

missingness_by_group_table(df: pd.DataFrame, *, group_col: str, columns: Sequence[str] | None = None) -> pd.DataFrame

Parameters

Parameter	Required	Type	Default	Description
`df`	yes	`pd.DataFrame`		Cohort or feature table to audit.
`group_col`	yes	`str`		Column defining groups such as cohort, center, split, or endpoint status.
`columns`	no	`Sequence[str] or None`	`None`	Columns to summarize. If omitted, all columns except ``group_col`` are included.

Returns

pandas.DataFrame - Long table with variable, group, row count, missing count, and missing percent.

See also: Cohort And Analysis module guide.

`fyron.statistics.qc.normality_table`

Run normality tests for multiple columns and return one table.

Import path: fyron.statistics.qc.normality_table

python

normality_table(df: pd.DataFrame, columns: Sequence[str], *, method: str = 'shapiro') -> pd.DataFrame

Parameters

Parameter	Required	Type	Default	Description
`df`	yes	`pd.DataFrame`		Analysis table containing numeric columns to test.
`columns`	yes	`Sequence[str]`		Column names to test.
`method`	no	`str`	`'shapiro'`	Normality test passed to ``normality_test. The default is` `"shapiro"``.

Returns

pandas.DataFrame - One row per variable with test name, statistic, p-value, non-missing sample size, and an error column when a test could not be computed.

See also: Cohort And Analysis module guide.

`fyron.statistics.qc.outlier_table`

Flag numeric outliers by IQR fences and return one summary row per column.

Import path: fyron.statistics.qc.outlier_table

python

outlier_table(df: pd.DataFrame, columns: Sequence[str], *, method: str = 'iqr', iqr_multiplier: float = 1.5) -> pd.DataFrame

Parameters

Parameter	Required	Type	Default	Description
`df`	yes	`pd.DataFrame`		Analysis table containing numeric variables.
`columns`	yes	`Sequence[str]`		Numeric columns to inspect.
`method`	no	`str`	`'iqr'`	Outlier method. Currently only ``"iqr"`` is supported.
`iqr_multiplier`	no	`float`	`1.5`	Multiplier for the IQR fence, commonly ``1.5``.

Returns

pandas.DataFrame - One row per variable with non-missing count, outlier count, outlier percent, lower fence, and upper fence.

See also: Cohort And Analysis module guide.

`fyron.statistics.regression.fit_linear_regression_table`

Fit an ordinary least-squares model and return a manuscript table.

Import path: fyron.statistics.regression.fit_linear_regression_table

python

fit_linear_regression_table(df: pd.DataFrame, *, outcome_col: str, covariates: Sequence[str], add_intercept: bool = True, alpha: float = 0.05, p_adjust: str | None = None) -> pd.DataFrame

Parameters

Parameter	Required	Type	Default	Description
`df`	yes	`pd.DataFrame`		Analysis table containing the outcome and covariate columns.
`outcome_col`	yes	`str`		Numeric continuous outcome column.
`covariates`	yes	`Sequence[str]`		Predictor columns. Categorical columns are dummy encoded with the first level dropped.
`add_intercept`	no	`bool`	`True`	If True, include an intercept row in the model and output table.
`alpha`	no	`float`	`0.05`	Two-sided confidence interval width, where ``0.05`` gives 95% intervals.
`p_adjust`	no	`str or None`	`None`	Optional multiple-testing correction method applied to non-intercept p-values, such as ``"fdr_bh",` `"bonferroni", or` `"holm"``.

Returns

pandas.DataFrame - One row per model term with beta estimate, standard error, confidence interval, test statistic, p-value, sample size, model name, outcome, and formula.

See also: Cohort And Analysis module guide.

`fyron.statistics.regression.fit_logistic_regression_table`

Fit a binary logistic regression model and return OR/CI/p-value rows.

Import path: fyron.statistics.regression.fit_logistic_regression_table

python

fit_logistic_regression_table(df: pd.DataFrame, *, outcome_col: str, covariates: Sequence[str], add_intercept: bool = True, alpha: float = 0.05, p_adjust: str | None = None, max_iter: int = 1000) -> pd.DataFrame

Parameters

Parameter	Required	Type	Default	Description
`df`	yes	`pd.DataFrame`		Analysis table containing a binary outcome and covariate columns.
`outcome_col`	yes	`str`		Binary outcome column coded as ``0` `and` `1``.
`covariates`	yes	`Sequence[str]`		Predictor columns. Categorical columns are dummy encoded with the first level dropped.
`add_intercept`	no	`bool`	`True`	If True, include an intercept row in the model and output table.
`alpha`	no	`float`	`0.05`	Two-sided confidence interval width, where ``0.05`` gives 95% intervals.
`p_adjust`	no	`str or None`	`None`	Optional multiple-testing correction method applied to non-intercept p-values.
`max_iter`	no	`int`	`1000`	Maximum optimizer iterations for the logistic likelihood fit.

Returns

pandas.DataFrame - One row per model term with odds ratio, log-odds estimate, standard error, confidence interval, z statistic, p-value, sample size, model name, outcome, and formula.

See also: Cohort And Analysis module guide.

`fyron.statistics.regression.vif_table`

Return variance inflation factors for covariates after dummy encoding.

Import path: fyron.statistics.regression.vif_table

python

vif_table(df: pd.DataFrame, covariates: Sequence[str]) -> pd.DataFrame

Parameters

Parameter	Required	Type	Default	Description
`df`	yes	`pd.DataFrame`		Analysis table containing the covariate columns.
`covariates`	yes	`Sequence[str]`		Predictor columns to inspect for multicollinearity. Categorical columns are dummy encoded before VIF calculation.

Returns

pandas.DataFrame - Table with one row per encoded term and columns `term, vif, and n`.

See also: Cohort And Analysis module guide.

`fyron.statistics.tests.StatisticalTestResult`

Named result returned by Fyron statistical tests.

Import path: fyron.statistics.tests.StatisticalTestResult

python

StatisticalTestResult()

Parameters

No parameters documented.

Returns

Class constructor.

See also: Cohort And Analysis module guide.

Public Methods

`fyron.statistics.tests.StatisticalTestResult.to_dict`

See the signature and linked module guide for context.

Import path: fyron.statistics.tests.StatisticalTestResult.to_dict

python

StatisticalTestResult.to_dict(self) -> dict[str, Any]

Parameters

No parameters documented.

Returns

dict[str, Any]

See also: Cohort And Analysis module guide.

`fyron.statistics.tests.categorical_test`

Run a chi-square or Fisher exact test on a contingency table.

Import path: fyron.statistics.tests.categorical_test

python

categorical_test(table: Sequence[Sequence[Any]], *, method: CategoricalMethod = 'auto') -> StatisticalTestResult

Parameters

Parameter	Required	Type	Default	Description
`table`	yes	`Sequence[Sequence[Any]]`		See signature.
`method`	no	`CategoricalMethod`	`'auto'`	See signature.

Returns

StatisticalTestResult

See also: Cohort And Analysis module guide.

`fyron.statistics.tests.compare_multiple_groups`

Compare three or more numeric groups with ANOVA or Kruskal-Wallis.

Import path: fyron.statistics.tests.compare_multiple_groups

python

compare_multiple_groups(groups: Sequence[Sequence[Any]], *, method: MultiGroupMethod = 'auto') -> StatisticalTestResult

Parameters

Parameter	Required	Type	Default	Description
`groups`	yes	`Sequence[Sequence[Any]]`		See signature.
`method`	no	`MultiGroupMethod`	`'auto'`	See signature.

Returns

StatisticalTestResult

See also: Cohort And Analysis module guide.

`fyron.statistics.tests.compare_two_groups`

Compare two numeric groups.

Import path: fyron.statistics.tests.compare_two_groups

python

compare_two_groups(group_a: Sequence[Any], group_b: Sequence[Any], *, paired: bool = False, method: TwoGroupMethod = 'auto', alternative: Alternative = 'two-sided', equal_var: bool = False) -> StatisticalTestResult

Parameters

Parameter	Required	Type	Default	Description
`group_a`	yes	`Sequence[Any]`		See signature.
`group_b`	yes	`Sequence[Any]`		See signature.
`paired`	no	`bool`	`False`	See signature.
`method`	no	`TwoGroupMethod`	`'auto'`	See signature.
`alternative`	no	`Alternative`	`'two-sided'`	See signature.
`equal_var`	no	`bool`	`False`	See signature.

Returns

StatisticalTestResult

See also: Cohort And Analysis module guide.

`fyron.statistics.tests.correlation_test`

Test association between two numeric variables.

Import path: fyron.statistics.tests.correlation_test

python

correlation_test(x: Sequence[Any], y: Sequence[Any], *, method: CorrelationMethod = 'spearman', alternative: Alternative = 'two-sided') -> StatisticalTestResult

Parameters

Parameter	Required	Type	Default	Description
`x`	yes	`Sequence[Any]`		See signature.
`y`	yes	`Sequence[Any]`		See signature.
`method`	no	`CorrelationMethod`	`'spearman'`	See signature.
`alternative`	no	`Alternative`	`'two-sided'`	See signature.

Returns

StatisticalTestResult

See also: Cohort And Analysis module guide.

`fyron.statistics.tests.normality_test`

Test whether a numeric vector is compatible with a normal distribution.

Import path: fyron.statistics.tests.normality_test

python

normality_test(values: Sequence[Any], *, method: Literal['shapiro', 'normaltest'] = 'shapiro') -> StatisticalTestResult

Parameters

Parameter	Required	Type	Default	Description
`values`	yes	`Sequence[Any]`		See signature.
`method`	no	`Literal['shapiro', 'normaltest']`	`'shapiro'`	See signature.

Returns

StatisticalTestResult

See also: Cohort And Analysis module guide.