IO & Display¶

Table Reader¶

Reads a tabular file (CSV, TSV) using pandas and outputs a DataFrame.

Details

file_path — path to the input file (widget or upstream port).
separator — column delimiter (default: ,).

Direction	Port	Type
Input	`file_path`	path
Output	`out`	table

Folder Iterator¶

Selects a folder and file pattern for batch processing.

Details

The actual looping is managed by the Batch Runner in main.py.

folder_path — directory to iterate over.
pattern — glob pattern for matching files (default: *.csv).
iterate_mode — iterate over Files or Subdirectories.

Direction	Port	Type
Output	`file_path`	path

Properties: Iterate

Video Iterator¶

Browses and iterates over frames of a video file.

Details

Preview any frame with the browse slider. Select a start/end range with the dual-handle range slider, then use Batch Run to process each frame through the downstream graph.

video_path — path to the video file.

Direction	Port	Type
Output	`file_path`	path

Batch Accumulator¶

Collects the output of each batch iteration and merges them after the batch finishes.

Details

Connect upstream data to the in port; the out port emits the merged result only after the entire batch is complete.

Batch context stamping:

Automatically adds frame and file metadata to each collected value.
For TableData, this lets downstream nodes identify which frame each row came from.

Direction	Port	Type
Input	`in`	any
Output	`out`	any

Image Reader¶

Reads an image file and outputs it as a float32 [0,1] numpy array.

Details

Supported formats:

Standard — JPEG, PNG, BMP, and other PIL-supported formats (8-bit)
TIFF — 8/12/14/16-bit microscopy TIFFs (bit depth preserved). Multi-page TIFFs output a CollectionData with one ImageData per page.
OIR — Olympus .oir files (Rust accelerated, with Python fallback)

The original bit depth is stored as metadata for downstream nodes (threshold sliders, histogram, save). All processing uses float32 [0,1] internally.

Options:

channels — comma-separated channel numbers (0-4, where 0 = black/pad). 2 for single grayscale channel, 1,2,3 for RGB, 2,3,4 to map channels 2/3/4 as R/G/B, 1,0,3 to map ch1 as red, black as green, ch3 as blue.

Direction	Port	Type
Input	`file_path`	path
Output	`out`	image

Data Saver¶

Saves incoming data to a file on disk.

Details

Supported output types:

DataFrame — saved as CSV, TSV, or .pzfx (GraphPad Prism)
Figure — saved as an image at the figure's native DPI
Image — saved via PIL in any format matching the file extension
file_path — destination path (widget or upstream port).

Direction	Port	Type
Input	`in`	any
Input	`file_path_in`	path

Batch Gate¶

Pass-through gate that pauses the batch pipeline for user review.

Details

Wire between any two nodes using the single any-typed input/output. Blocking happens inside evaluate(), so multiple gates pause independently at their own step in the topological evaluation order.

Controls:

Next — let this iteration continue past the gate
Refresh — re-evaluate upstream nodes and update previews
Pass All — stop pausing for the rest of this batch run

Direction	Port	Type
Input	`in`	any
Output	`out`	any

Pop-up Display¶

Takes any input and pops up a preview window to inspect it.

Details

Supported data types:

DataFrame — shown as an editable table dialog
Figure — rendered to PNG and shown in a scrollable image dialog
Image — displayed as a scrollable PIL image dialog
Other — shown as a plain text message box

Direction	Port	Type
Input	`in`	any

Data Table Node¶

Displays incoming DataFrame data directly on the node surface.

Direction	Port	Type
Input	`in`	table

Data Figure Node¶

Displays incoming Image or Figure data directly on the node surface.

Details

Accepts FigureData (with optional SVG override) or raw matplotlib figures.

Direction	Port	Type
Input	`in`	figure

Image Viewer¶

Displays a PIL Image directly on the node surface for quick inline inspection.

Details

Accepted input types:

ImageData — unwraps the payload
LabelData — uses the pre-generated colored visualization
Raw PIL Image — displayed as-is

Direction	Port	Type
Input	`in`	image

Universal Node¶

Executes arbitrary Python code to process multiple inputs and push results to outputs.

Details

Available variables in user code:

inputs — list of upstream data values
output — assign the result here (auto-wrapped into TableData, ImageData, or FigureData)
pd, np, plt, sns — pre-imported libraries

Direction	Port	Type
Input	`in`	any
Output	`out`	any

Path Modifier¶

Takes a file path and modifies it by adding a suffix, changing the extension, or overriding the folder.

Details

suffix — string appended to the file stem (default: _analyzed).
ext — replacement file extension (leave empty to keep original).
folder — optional folder override for the output path.

Direction	Port	Type
Input	`path`	path
Output	`path`	path

Collect¶

Pack multiple data items into a named collection.

Details

Connect any number of items to the multi-input port. Each connection gets a name (auto-populated from the upstream port name, editable). The output is a single CollectionData that flows as one wire.

Downstream nodes that expect a single item will automatically loop over all items in the collection and repack the results.

Direction	Port	Type
Input	`in`	any
Output	`collection`	collection

Select Collection¶

Extract a single item from a collection by name.

Details

Type a name or pick from the dropdown. The dropdown auto-populates with available item names when the collection is connected.

Direction	Port	Type
Input	`collection`	collection
Output	`out`	any

Pop Collection¶

Extract one item from a collection and output the rest separately.

Details

Two outputs: the extracted item on item, and a new collection without that item on rest. Type a name or pick from the dropdown.

Direction	Port	Type
Input	`collection`	collection
Output	`item`	any
Output	`rest`	collection

Split Collection¶

Split a collection into two groups by selecting which items go to each output.

Details

Type item names separated by ' | ' or pick from the dropdown to add. Selected items go to selected, the rest go to rest.

Direction	Port	Type
Input	`collection`	collection
Output	`selected`	collection
Output	`rest`	collection

Save Collection¶

Saves all items in a collection to disk.

Details

Each item is saved as a separate file using the item name as a suffix. Supports images (TIFF, PNG), tables (CSV, TSV), and figures.

If a path is connected, it is used as the base — the item name is inserted before the extension. Otherwise the folder + extension fields are used.

Direction	Port	Type
Input	`file_path`	path
Output	`status`	table

Rename Collection¶

Rename items in a collection using a visual mapping table.

Details

When a collection is connected, the table auto-populates with original names. Edit the 'New Name' column to rename items. Leave blank to keep the original name.

Direction	Port	Type
Input	`collection`	collection
Output	`collection`	collection

Collection Info¶

Outputs a table listing item names, types, shapes, and metadata.

Details

All number and string valued metadata fields are included as extra columns.

Direction	Port	Type
Input	`collection`	collection
Output	`info`	table

Filter Collection¶

Keep or remove items by pattern matching on names.

Details

Supports simple wildcards (* and ?) or exact names. Multiple patterns separated by | (pipe).

Mode:

Keep — only matching items pass through
Remove — matching items are excluded

Direction	Port	Type
Input	`collection`	collection
Output	`matched`	collection
Output	`rest`	collection

Properties: Mode

Map Names¶

Batch rename collection items using find/replace, prefix, or suffix.

Details

Operations (applied in order): 1. Find/Replace — replace substring in all names 2. Prefix — add text before each name 3. Suffix — add text after each name

Direction	Port	Type
Input	`collection`	collection
Output	`collection`	collection