Data and I/O

Loading, shaping and saving data: tables, files, arrays, lists, dictionaries and variables.

32 nodes. Right-click any node in the editor to read this documentation in the app.

Array

🔎 Index

id index · Array · Python export: yes

Pick one element (or row) out of an array by integer position.

Inputs

Port	Type	Description
array	ndarray	Array to index into.
index	int	Integer position (negative counts from the end).

Outputs

Port	Type	Description
element	Any	arr[index] — a scalar for 1-D input, a row for 2-D.

🧱 Reshape

id reshape · Array · Python export: yes

Reshape an array to a shape supplied by another node. The shape input is a list/array of ints (e.g. [2, 3]) or a single int for a 1-D result.

Inputs

Port	Type	Description
array	ndarray	Array to reshape.
shape	list[int] | int	Target shape, e.g. [2, 3]. A single int gives a 1-D array.

Outputs

Port	Type	Description
reshaped	ndarray	The reshaped array.

🔀 Transpose

id transpose · Array · Python export: yes

Transpose an array (reverse its axes). For a 2-D array this swaps rows and columns.

Inputs

Port	Type	Description
array	ndarray	Array to transpose.

Outputs

Port	Type	Description
transposed	ndarray	The transposed array.

Data

➕ Append List

id append_list · Data · Python export: yes

Append an item to a list and output the new list. Inside a For Loop it accumulates across iterations: the list you wire in seeds iteration 0 and each later iteration adds to the running result.

Inputs

Port	Type	Description
list	list	Starting list (seeds the first iteration).
item	Any	Value to append.

Outputs

Port	Type	Description
list	list	The list with the item appended.

🏷️ Assign Column

id dataframe_assign_column · Data · Python export: yes

Add (or replace) a column of a DataFrame with a list, array or Series — for example attach a model's predictions or cluster labels to the rows they belong to, then colour a Scatter Plot by that column. A Series is matched to rows by index; a list or array goes in by position and must have one value per row. The input table is not modified.

Inputs

Port	Type	Description
dataframe	DataFrame	Table to add the column to.
values	Series | list | ndarray	The column's values (a Series is aligned by index).

Outputs

Port	Type	Description
dataframe	DataFrame	A copy of the table with the column added / replaced.

Fields

Field	Type	Default	Choices
Column name	text	new_column

🔗 Correlation Matrix

id correlation_matrix · Data · Python export: yes

Pairwise correlation of a table's numeric columns (DataFrame.corr): Pearson (linear), Spearman (rank) or Kendall's tau (rank concordance). Columns: comma-separated, blank = all numeric. Wire into a Heatmap to visualise.

Inputs

Port	Type	Description
dataframe	DataFrame	Table with numeric columns.

Outputs

Port	Type	Description
correlation	DataFrame	Square correlation matrix (columns x columns).

Fields

Field	Type	Default	Choices
Method	select	pearson	pearson, spearman, kendall
Columns (comma-separated, blank = all)	text

📊 DataFrame Describe

id dataframe_describe · Data · Python export: yes

Summary statistics for a DataFrame (count, mean, std, min, quartiles, max) — like df.describe().

Inputs

Port	Type	Description
dataframe	DataFrame	Table to summarise.

Outputs

Port	Type	Description
summary	DataFrame	One column per numeric column, one row per statistic.

🔍 DataFrame Head

id dataframe_head · Data · Python export: yes

Return the first N rows of a DataFrame (like df.head(n)) and show them as a table.

Inputs

Port	Type	Description
dataframe	DataFrame	Table to take the first rows from.

Outputs

Port	Type	Description
head	DataFrame	The first n rows.

Fields

Field	Type	Default	Choices
Rows	int	5

🧮 DataFrame Table

id variable · Data · Python export: yes

An editable table of values (choose float / int / bool / string) that outputs a pandas DataFrame. Use the +R / +C buttons to grow it. For a NumPy array instead, use NumPy Array Table.

Outputs

Port	Type	Description
variable	DataFrame	The table as a DataFrame (columns numbered from 0).

🔑 Dict Get

id dict_get · Data · Python export: yes

Look up a key in a dictionary; gives None if the key is missing. The key is read as JSON first (so 2 finds the int key 2), falling back to plain text.

Inputs

Port	Type	Description
dict	dict	Dictionary to read from.
key	str | int	Key to look up (overrides the Key field). Overrides the Key field when connected.

Outputs

Port	Type	Description
value	Any	dict[key], or None if the key is absent.

Fields

Field	Type	Default	Choices
Key	text

🗝️ Dict Keys

id dict_keys · Data · Python export: yes

The keys of a dictionary, as a list.

Inputs

Port	Type	Description
dict	dict	Dictionary to list the keys of.

Outputs

Port	Type	Description
keys	list	The dictionary's keys, in insertion order.

📝 Dict Set

id dict_set · Data · Python export: yes

Return a copy of a dictionary with one key set (or replaced). The original is not modified. A missing dict input starts from {}.

Inputs

Port	Type	Description
dict	dict	Dictionary to copy (optional — defaults to empty).
key	str | int	Key to set (overrides the Key field). Overrides the Key field when connected.
value	Any	Value to store under the key.

Outputs

Port	Type	Description
dict	dict	New dictionary with the key set.

Fields

Field	Type	Default	Choices
Key	text

📦 Dictionary

id dict_literal · Data · Python export: yes

A literal dictionary, typed as a JSON object ({"a": 1, "b": 2}).

Outputs

Port	Type	Description
dict	dict	The parsed dictionary.

Fields

Field	Type	Default	Choices
Dictionary (JSON format)	textarea	{ "a": 1, "b": 2 }

🔢 Float

id scalar_float · Data · Python export: yes

A single floating-point number. Wire it into any port that takes a number to drive a parameter from the graph.

Outputs

Port	Type	Description
scalar_float	float	The number.

Fields

Field	Type	Default	Choices
Value	float	0.0

🖼️ Image Show

id image_show · Data · Python export: yes

Display a NumPy array as an image on the node (like imshow) and pass the array through unchanged. Greyscale (H x W), H x W x 1/3/4 and channel-first C x H x W arrays are understood; uint8 is shown as-is, floats are read as 0-1 (or 0-255). Tick 'Auto-scale' to stretch the min..max range across the display - useful for heatmaps and other non-image data.

Inputs

Port	Type	Description
array	ndarray	Image-like array to display.

Outputs

Port	Type	Description
array	ndarray	The input array, unchanged.

Fields

Field	Type	Default	Choices
Auto-scale (stretch min..max)	checkbox	false

🔢 Integer

id scalar_int · Data · Python export: yes

A single integer. Wire it into any port that takes a whole number (sizes, counts, indices) to drive a parameter from the graph.

Outputs

Port	Type	Description
scalar_int	int	The integer.

Fields

Field	Type	Default	Choices
Value	int	0

📏 Length

id list_length · Data · Python export: yes

The number of items in a list, tuple, dict, string or array — len(x).

Inputs

Port	Type	Description
sequence	list | tuple | dict | str | ndarray	Anything with a length.

Outputs

Port	Type	Description
n	int	Number of items.

📜 List

id list_obj · Data · Python export: yes

A literal list. Type it as JSON ([1, 2, 3] or ["a", "b"]) or as comma-separated numbers (1, 2, 3). Every run gets a fresh copy.

Outputs

Port	Type	Description
list	list	The parsed list.

Fields

Field	Type	Default	Choices
List (JSON or CSV)	textarea	[0, 1, 2, 3, 4]

🔎 List Index

id list_index · Data · Python export: yes

Pick one element out of a list, tuple or array by integer position (0-based; negative counts from the end).

Inputs

Port	Type	Description
sequence	list | tuple | ndarray	The sequence to index into.
index	int	Position to read (overrides the Index field). Overrides the Index field when connected.

Outputs

Port	Type	Description
element	Any	sequence[index].

Fields

Field	Type	Default	Choices
Index	int	0

📥 Load .npy

id numpy_load_npy · Data · Python export: yes

Load a NumPy .npy file into an array (np.load). Pickle support is off by default — only enable it for files you trust.

Outputs

Port	Type	Description
array	ndarray	The loaded array.

Fields

Field	Type	Default	Choices
File (.npy)	file
Allow pickle (only enable for trusted files)	checkbox	false
Memory-map (large files; read-only)	checkbox	false

📦 Load .npz

id numpy_load_npz · Data · Python export: yes

Load a NumPy .npz archive as a dict of {name: array}. Wire it into Dict Get to pull out one array.

Outputs

Port	Type	Description
arrays	dict	{array name: ndarray} for every array in the archive.

Fields

Field	Type	Default	Choices
File (.npz)	file

🧮 NumPy Array Table

id array_variable · Data · Python export: yes

An editable table of values (choose float / int / bool / string) that outputs a NumPy array (ndarray). Use the +R / +C buttons to grow it. For a pandas DataFrame instead, use DataFrame Table.

Outputs

Port	Type	Description
array variable	ndarray	The table as a 2-D NumPy array.

🖼️ Read Image + Preview

id image_read · Data · Python export: yes

Load an image file as an RGB NumPy array (H x W x 3, uint8) and show a thumbnail on the node. For a preview-less reader (or to read greyscale / keep the original mode) see Read Image File in I/O; to look at an array you already have, use Image Show.

Outputs

Port	Type	Description
array	ndarray	RGB pixel array, shape (H, W, 3), dtype uint8.

Fields

Field	Type	Default	Choices
Image	image

📤 Save .npy

id numpy_save_npy · Data · Python export: yes

Write an array to a .npy file and output the path it wrote to, so later nodes can chain off it.

Inputs

Port	Type	Description
array	ndarray	Array to save.

Outputs

Port	Type	Description
filepath	str	Path of the written file.

Fields

Field	Type	Default	Choices
Output path	text	out.npy

🔤 String

id string · Data · Python export: yes

A piece of text. Wire it into any port that takes a string (column names, file paths, keys).

Outputs

Port	Type	Description
string var	string	The text.

Fields

Field	Type	Default	Choices
Text	string	abc

🪢 Tuple

id tuple_obj · Data · Python export: yes

A literal tuple, typed as comma-separated values (1, 2, 3 or "a", "b").

Outputs

Port	Type	Description
tuple	tuple	The parsed tuple.

Fields

Field	Type	Default	Choices
Tuple (CSV)	text	1, 2, 3

I/O

📄 CSV Read

id csv_read · I/O · Python export: yes

Load a CSV file into a DataFrame (pandas.read_csv) and show its shape and first rows.

Outputs

Port	Type	Description
dataframe	DataFrame	The parsed CSV.

Fields

Field	Type	Default	Choices
File	file

🖼️ Read Image File

id img_read · I/O · Python export: yes

Read an image file into a NumPy array using imageio, keeping the file's own mode and bit depth (no preview). To see the picture, wire the array into Image Show, or use Read Image + Preview in Data.

Inputs

Port	Type	Description
filepath	str	Image path (overrides the Image field). Overrides the Image field when connected.

Outputs

Port	Type	Description
image	ndarray	Pixel array, shape (H, W[, C]).

Fields

Field	Type	Default	Choices
Image	file

📗 Spreadsheet Read

id spreadsheet_read · I/O · Python export: yes

Read an Excel (.xlsx / .xlsm) or CSV file into a DataFrame. Excel files are previewed as a spreadsheet with their real formatting (fills, bold, borders, number formats) and sheet tabs; right-click the preview for a full-screen view. Legacy .xls is not supported.

Outputs

Port	Type	Description
dataframe	DataFrame	The chosen sheet's data (first row as header if ticked).

Fields

Field	Type	Default	Choices
File (.xlsx, .xlsm, .csv)	file
Sheet name (blank = first)	text
First row is header	checkbox	true
Preview rows	int	50

📤 Spreadsheet Write

id spreadsheet_write · I/O · Python export: yes

Write a DataFrame to an .xlsx file (one sheet, optionally with the index) and pass the table through. Don't wire it into a Spreadsheet Read of the same file in the same graph — they may run at the same time.

Inputs

Port	Type	Description
dataframe	DataFrame	Table to write.

Outputs

Port	Type	Description
dataframe	DataFrame	The input table, unchanged.

Fields

Field	Type	Default	Choices
Output file (.xlsx)	text	output.xlsx
Sheet name	text	Sheet1
Include index column	checkbox	false

💾 Write CSV

id csv_write · I/O · Python export: yes

Write an array / list / DataFrame to a CSV file (no index column) and pass the data through. Wire a filename into the port to override the field.

Inputs

Port	Type	Description
data	DataFrame | ndarray | list	Data to write.
filename	str	Output path (overrides the Filename field). Overrides the Filename field when connected.

Outputs

Port	Type	Description
data	Any	The input data, unchanged.

Fields

Field	Type	Default	Choices
Filename	text	output.csv

💾 Write Image File

id img_write · I/O · Python export: yes

Write an image array to a file (format from the extension, e.g. .png) using imageio, and pass the array through.

Inputs

Port	Type	Description
image	ndarray	Pixel array to write, shape (H, W[, C]).
filename	str	Output path (overrides the Filename field). Overrides the Filename field when connected.

Outputs

Port	Type	Description
image	ndarray	The input array, unchanged.

Fields

Field	Type	Default	Choices
Filename	text	out.png