A powerful VS Code extension designed specifically for data scientists and researchers who iterate heavily on code. It provides a structured, file-scoped iterative coding environment that lives alongside your scripts and Jupyter notebooks.
The Problem
When working in data science or algorithm development, you often duplicate cells to test a variation of the code, comment out old blocks "just in case", or restart your kernel entirely because your scope got polluted. This results in messy notebooks, duplicated logic, and lost trains of thought.
The Solution
Jupyter Workflow Manager acts as a contextual memory and iteration layer directly mapped to your active file.
Key Features
File-Scoped Workflows: Every .py or .ipynb file automatically gets its own isolated .workflow.json graph safely tracking your logic.
Micro-Versioning: Instead of duplicating lines or cells, you create "Steps" (e.g., Load Data, Clean Data). Inside each Step, you create "Versions". You can swap the active version of a step instantaneously using the sidebar.
True Code Reuse: Your logic lives inside the workflow graph. Your editor simply acts as a scratchpad. Apply steps directly into your editor context, iterate natively, save, and export.
Compare & Diff: Curious what changed? Select "Compare with Active Version" in the sidebar to open VS Code's native diff viewer to see exactly how your experimental iteration differs from your baseline.
Clean Export: When your experiment is complete, hit "Export Clean Workflow" to generate a pristine, non-polluted python script from the exact versions currently marked active in your workflow tree.
Usage
Open any Python script or Jupyter Notebook.
Open the Workflow view in the Activity Bar.
Click the + icon to add a Step (e.g. 1_train_model).
Click the + inside the step to add a version (it will capture selected text, or inject an #add your code here: placeholder).
Open the version, add your code natively, and hit Ctrl+S to save it back into the graph.
Auto-apply, branch, and experiment safely without fear of losing your original code.
Requirements
No external dependencies are required. Powered purely by VS Code's robust Node system.
