onelake-tui 0.5.0

Unofficial TUI and client library for Microsoft Fabric OneLake

OneLake TUI (Unofficial)

A terminal UI for browsing Microsoft Fabric workspaces, lakehouses, and Delta tables — right from your terminal. No portal, no notebooks, no Spark cluster required.

pip install onelake-tui
az login
onelake-tui

onelake-tui is the published PyPI package and currently bundles both the terminal UI and the onelake_client library.

What is this?

OneLake TUI gives you a keyboard-driven file-manager experience for Microsoft Fabric's OneLake storage. Browse workspaces, inspect lakehouses and warehouses, preview files, and explore Delta table metadata — all without leaving the terminal.

Built with Textual and an async Python client for the Fabric REST and OneLake DFS APIs.

Features

• Three-panel layout — workspace picker → item list → DFS file tree + detail/preview
• Rich file preview — press Enter on any file:
  • Markdown, JSON (NDJSON), CSV, Parquet, Avro, syntax-highlighted code
  • All previews are selectable and copyable
• Delta table inspector — tabbed detail view:
  • Schema, data preview (first 100 rows), transaction history, Change Data Feed
• Schema-aware table detection — handles both Tables/schema/table (mirrored DBs) and Tables/table (lakehouses)
• Live search — press / to filter workspaces
• Multi-environment — PROD, MSIT, DXT, DAILY via --env flag
• Copy menu — press y to choose HTTPS/ABFSS + named/GUID URI formats
• Vim-friendly navigation — j/k/g/G + h/l, with Tab/Shift+Tab still supported
• Toggleable footer — Ctrl+F hides/shows the status bar
• Zero config — uses az login, no service keys or config files required

Roadmap (linked issues)

Roadmap source of truth: ../README.md#roadmap

Area	Status
OneLake TUI (Unofficial)	✅ Working (browse, preview, inspect, copy path)
File preview (MD/JSON/CSV/Parquet/Avro)	✅ Done
Delta table detail (schema/data/history/CDF)	✅ Done
Workspace search/filter	✅ Done
OneLake CLI	🔲 Planned (onelake ls, onelake cat, onelake cp)
Download/upload	🔲 Planned
Shortcuts discoverability/manageability	🔲 Planned
ADLS direct connections	🔲 Planned
OneLake security policy management	🔲 Planned
Bulk operations	🔲 Planned
Copying data + soft-deleted data handling	🔲 Planned

Installation

Prerequisites: Python 3.11+, Azure CLI (az login)

pip install onelake-tui    # from PyPI
onelake-tui                # launch (PROD)
onelake-tui --env msit     # Microsoft internal testing ring

From source (for development):

cd TUI
uv sync
uv run onelake-tui

Quick Start

1. Authenticate — az login (if you haven't already)
2. Launch — onelake-tui
3. Select workspace — pick from the left panel
4. Select item — choose a lakehouse, warehouse, or mirrored DB
5. Browse files — expand folders in the tree
6. Preview — press Enter on any file
7. Explore tables — expand Tables/ for Delta metadata
8. Search — press / to filter workspaces

Keybindings

Key	Action
↑ / ↓	Navigate
j / k	Navigate (vim-style)
g / G	Jump to top / bottom
← / →	Collapse/expand tree nodes
Enter	Preview file / Expand folder
/	Search/filter workspaces
Escape	Close search / go back
h / l	Previous/next panel
Tab / Shift+Tab	Switch panels
y	Open copy menu (HTTPS/ABFSS, named/GUID)
Ctrl+F	Toggle footer visibility
S	Save SVG screenshot to current directory
r	Refresh
?	Show help
q	Quit

Architecture

TUI/src/
├── onelake_client/            # Standalone async Python client library
│   ├── auth.py                #   Dual-scope token management
│   ├── environment.py         #   Environment ring config (PROD/MSIT/DXT/DAILY)
│   ├── _http.py               #   httpx retry + pagination
│   ├── fabric/client.py       #   Fabric REST API (control plane)
│   ├── dfs/client.py          #   OneLake DFS API (data plane)
│   ├── tables/delta.py        #   Delta table metadata reader
│   ├── tables/iceberg.py      #   Iceberg table metadata reader
│   └── models/                #   Pydantic data models
└── onelake_tui/               # Textual-based terminal UI
    ├── app.py                 #   Main app, keybindings, event wiring
    ├── app.tcss               #   Layout CSS (3-panel + footer)
    ├── workspace_picker.py    #   Flat filterable workspace list
    ├── item_list.py           #   Item list for selected workspace
    ├── tree.py                #   DFS file tree (single item)
    ├── detail.py              #   Detail/preview panel with rich rendering
    ├── sprite.py              #   OneLake-inspired splash art + shimmer animation
    ├── status_bar.py          #   3-line footer (path, shortcuts, auth)
    ├── nodes.py               #   Node dataclasses
    └── banner.py              #   Welcome screen delegate

Navigation model: Workspace → Item → DFS files/folders/tables. Each stage is a separate widget:

• WorkspacePicker (flat OptionList, filterable)
• ItemList (flat OptionList per workspace)
• OneLakeTree (lazy-loading tree per item)
• DetailPanel (context-aware: metadata on highlight, preview on Enter)

Auth Configuration

Authentication uses DefaultAzureCredential, which supports (in order):

Method	Use case
az login	Local development
Service principal	CI/CD pipelines
Managed identity	Azure-hosted environments

Environment Configuration

uv run onelake-tui              # PROD (default)
uv run onelake-tui --env msit   # Microsoft internal testing
uv run onelake-tui --env dxt    # Developer testing
uv run onelake-tui --env daily  # Daily builds

Each environment maps to the correct Fabric REST and OneLake DFS hostnames automatically.

Debugging

tail -f ~/.onelake-tui/debug.log

Development

cd TUI
uv sync --all-extras
uv run pytest           # Unit + integration tests
uv run ruff check src/  # Lint
uv run ruff format src/ # Format

License

MIT