incucyteai 0.1.0

CLI and Tkinter GUI for downloading Incucyte live-cell imaging TIFFs.

IncucyteAI — Auto-Download Agent

Automatically download TIF images from the Incucyte live-cell imaging system via its SOAP API. Replaces the manual click-to-download workflow with a script that polls for new scans and saves them to a folder.

Requirements

• Python 3.10+
• Must be on the Imperial College London network (the Incucyte device is not internet-accessible)
• Windows with the Incucyte desktop client installed is required for first-time password encryption/login

Install

From PyPI, once published:

pip install incucyteai

From this source checkout:

pip install -e .

Quick Start (GUI)

incucyteai-gui

1. Enter the Incucyte host IP and your credentials, click Login
2. Select one or more vessels from the list
3. (Optional) Click Select Wells to choose specific wells
4. Choose channel filters (Phase, Green/Color 1, Red/Color 2)
5. Set the output folder
6. Set Start from: "First scan" to get all historical images, "Today" for new ones only, or a custom date
7. Click Download Now to review the export dialog, or Start Watching to poll automatically

Quick Start (CLI)

# Test connectivity
incucyteai probe --host 129.31.116.189

# Login (saves credentials)
incucyteai login --host 129.31.116.189

# List experiments and vessels
incucyteai vessels

# Download all images from a vessel, starting from its very first scan
incucyteai download -v 38 -o ./images --start-from first

# Download from a specific date
incucyteai download -v 38 -o ./images --start-from 2026-03-01

# Download Phase + green + red fluorescence channels as ImageJ hyperstacks
incucyteai download -v 38 -o ./images --channels phase,green,red --hyperstack

# Download each selected well/channel as a time stack across scans
incucyteai download -v 38 -o ./images --wells D2-D5 --channels phase,green --time-stack

# Download one channel+time hyperstack per selected well
incucyteai download -v 38 -o ./images --channels phase,green,red --time-stack --hyperstack

# Watch mode: poll every 10 minutes for new scans
incucyteai watch -v 38 -o ./images -i 10 --start-from first

The original script entry points still work from a source checkout:

python incucyte_gui.py
python incucyte_downloader.py vessels
python -m incucyteai vessels

Features

• Parallel scan-time checks - the export-list building step also uses the worker count

• Parallel downloads — configurable number of worker threads (default 4)

• Optional green LUT for Phase — Phase contrast images are kept as returned by the API by default; use --green-lut to save green-channel RGB TIFs

• Named fluorescence channels — Incucyte Color 1 is shown as Green and Color 2 as Red; CLI aliases color1 and color2 still work

• ImageJ hyperstack export — optional one-file-per-well/time TIFF stacks containing all selected channels

• ImageJ time stack export — optional one-file-per-well/channel stacks across selected scans; combine with --hyperstack for channel+time stacks

• Pre-flight export dialog — one-shot downloads show source/calibration details, stack axes, channels, wells, output folder, and filename examples before starting

• Start from any date — download from the vessel's first scan, today, or any custom date

• Progress dialog (GUI) — file count, progress bar, speed estimate, ETA, cancel button

• Source-image progress for stacks — time-stack exports show inner source-image progress plus output-file completion

• Window title updates — shows download progress and next-poll countdown when minimized

• Retry with backoff — transient network failures retry up to 3 times

• State tracking — already-downloaded images are skipped on subsequent runs

• Well & channel filters — download only specific wells and/or image channels

• Elapsed-time filenames — files named with elapsed time from experiment start (e.g., VID38_A1_1_024h30m.tif)

• Graceful shutdown — closing the window during watch mode stops cleanly

CLI Options

download

Flag	Description
-v, --vessel	Vessel ID to download
-o, --output	Output directory
-s, --start-from	first (all history), YYYY-MM-DD, or omit for today
-d, --date	Single date to download (YYYY-MM-DD)
-w, --wells	Well filter (e.g., A1, A1,B3, A1-D4)
-c, --channels	Channel filter (phase, green/color1, red/color2, all)
--hyperstack	Save selected channels as one ImageJ hyperstack TIFF per well/site/time
--time-stack	Save individual wells as ImageJ time stacks across selected scans
--workers	Parallel download threads (default: 4)
--green-lut	Apply green LUT to Phase images
--no-green-lut	Keep Phase images as returned by the API (default)

watch

Same as download, plus:

Flag	Description
-i, --interval	Poll interval in minutes (default: 10)

File Structure

IncucyteAI/
  pyproject.toml           pip/build metadata and command entry points
  LICENSE                  BSD-3-Clause license
  incucyteai/              package namespace for python -m incucyteai
  incucyte_downloader.py   CLI tool and core download logic
  incucyte_gui.py          Tkinter GUI
  CLAUDE.md                Agent context (for Claude Code)
  SETUP_PLAN.md            On-site deployment guide
  README.md                This file
  .gitignore               Excludes state files and WSDL
  .tmp/                    Config, state, captures (gitignored)

Runtime config is saved outside the package when installed with pip:

• Windows: %APPDATA%\IncucyteAI
• macOS/Linux: $XDG_CONFIG_HOME/incucyteai or ~/.config/incucyteai
• Override: set INCUCYTEAI_HOME to a folder path

Existing source checkouts keep using .tmp/ if local state already exists there.

License

IncucyteAI is distributed under the BSD 3-Clause License.

Copyright (c) 2026, Jamie Malcolm. Ownership is personal rather than lab or institute owned.

Build and Publish

python -m pip install --upgrade build twine
python -m build
python -m twine check dist/*
python -m twine upload dist/*

Setup at Imperial

See SETUP_PLAN.md for step-by-step instructions.