Transform Library¶

Choregraph provides 48 built-in transform functions registered in TRANSFORM_REGISTRY. The Builder looks up functions by name when constructing Kedro pipeline nodes from the XML specification.

Categories¶

Filtering¶

Function	Description
`filter_less_than`	Keep rows where column < value
`filter_greater_than`	Keep rows where column > value
`filter_equal`	Keep rows where column == value
`filter_not_equal`	Keep rows where column != value
`filter_in_range`	Keep rows where min_value ≤ column ≤ max_value

Top / Bottom¶

Function	Description
`get_top_n`	Top N rows by column value
`get_top_percentage`	Top percentage of rows by column value
`get_bottom_n`	Bottom N rows by column value
`get_bottom_percentage`	Bottom percentage of rows by column value

Aggregation¶

Function	Description
`aggregate_mean`	Mean of numeric columns, optionally grouped
`aggregate_count`	Row count, optionally grouped
`aggregate_sum`	Sum of numeric columns, optionally grouped
`aggregate_median`	Median of numeric columns, optionally grouped

Min / Max¶

Function	Description
`calculate_min`	Minimum value from a column or list
`calculate_max`	Maximum value from a column or list

Column Operations¶

Function	Description
`select_columns`	Keep only specified columns
`drop_columns`	Remove specified columns
`rename_column`	Rename a single column
`add_label`	Add a constant-value label column

Row Operations¶

Function	Description
`slice_rows`	Keep a range of rows by index
`sort_values`	Sort by one or more columns
`sample_rows`	Random sample of rows
`count_rows`	Count rows, optionally grouped

Calculations¶

Function	Description
`calc_distance`	Euclidean distance from a reference point
`calc_ratio`	Ratio between two columns
`arithmetic_op`	Arithmetic between columns or constants (+, -, *, /)

Reshaping¶

Function	Description
`melt`	Unpivot columns into rows
`hierarchical_rollup`	Aggregate hierarchical data at multiple levels
`concat_partitions`	Concatenate partitioned datasets

Advanced¶

Function	Description
`normalize_column`	Min-max or z-score normalization
`discretize`	Bin continuous values (uniform or quantile)
`execute_code`	Execute user-provided Python code on a DataFrame

Time Series¶

Function	Description
`extract_date_part`	Extract year, month, day, etc. from datetime column
`rolling_statistics`	Rolling window aggregation (mean, sum, etc.)
`lag_lead`	Shift column values forward or backward
`offset_datetime`	Offset datetime column by a time delta
`forecast_time_series`	Simple time series forecasting

Multi-Input¶

Function	Description
`join`	Join multiple DataFrames on a key
`union`	Vertically stack (concatenate) DataFrames

JSON¶

Function	Description
`flatten_json`	Flatten nested JSON structures

Image¶

Function	Description
`image_to_dataframe`	Convert image pixels to a DataFrame
`extract_channel`	Extract a single color channel from image data
`image_metadata`	Extract image metadata (dimensions, format, etc.)

Geolocation¶

Function	Description
`geocode_location`	Enrich with lat/lon from location names
`get_country_contours`	Join country boundary geometries

NLP¶

Function	Description
`nlp_binarize_labels_auto`	Unsupervised multi-label binarization
`nlp_binarize_labels_hinted`	Supervised binarization with fuzzy hint matching

Excel¶

Function	Description
`tidy_excel_data`	LLM-assisted multi-table Excel tidying

Detailed Reference¶

Core Transforms — Filtering, aggregation, column/row ops, calculations, advanced, multi-input, JSON
Excel Transforms — LLM-assisted Excel processing
Geo Collection — Geocoding and country boundaries
NLP Collection — Text label binarization