pve-vm-setup/TASKS.md

TASKS

Codex working agreement

Use these rules for every implementation task in this repository:

• Write tests first.
• Use uv for all Python commands and dependency management.
• Use ruff for linting and formatting.
• Prefer small vertical slices that can be implemented and verified independently.
• Keep Proxmox API and business logic out of Textual widgets where possible.
• Field-like widgets should use compact 3-row layouts by default; bordered fields must not include an empty inner spacer row between the content and the border.
• Every screen must have explicit default, loading, success, empty, and error states where applicable.
• Every interactive screen must be covered by Textual interaction tests using run_test() and Pilot.
• Important visual states should be covered by snapshot tests.
• At the end of each task, summarize what changed and list unresolved UX concerns.

Repository commands

Codex should use these commands:

• Install dependencies: uv sync
• Run app: uv run python -m pve_vm_setup
• Run tests: uv run pytest
• Run lint checks: uv run ruff check .
• Format code: uv run ruff format .

Shared visual validation

• Use snapshot tests and visual review for form-heavy screens to confirm that single-line field widgets stay 3 rows high by default.
• Verify that bordered field-like controls do not render an empty spacer line above or below the value row.
• Allow taller controls only when a widget is intentionally multi-line or needs additional vertical content, and call that out in the task summary as a UX choice.

Implementation backlog

Task 1: Bootstrap the Textual app and project structure

Create the initial Textual application structure and make the repository runnable and testable.

Requirements:

• Create the application entrypoint used by uv run python -m pve_vm_setup.
• Set up a project structure that separates app shell, screens, widgets, models, and services.
• Add the initial test setup for unit tests, Textual interaction tests, and snapshot tests.
• Add a central state or domain module for the VM configuration workflow.
• Add a service interface for Proxmox API access so UI code can be tested with fakes or mocks.

Definition of done:

• Add or update tests first.
• uv run pytest passes.
• uv run ruff check . passes.
• uv run ruff format . has been run.
• The app opens a minimal Textual shell screen.
• Summarize what changed and list unresolved UX concerns.

Task 2: Implement authentication screen and Proxmox login flow

Build the login screen shown in SPEC.md and authenticate before the VM creation workflow starts.

Requirements:

• Ask the user for username and password.
• Load available authentication realms from the Proxmox API.
• Let the user choose the authentication realm.
• Default the authentication realm to Linux PAM standard authentication.
• Handle authentication failure and loading states.
• Prevent the wizard from continuing until authentication succeeds.

Definition of done:

• Add or update tests first.
• Add service tests for realm loading and login handling.
• Add Textual interaction tests using run_test() and Pilot for the login flow.
• Add snapshot coverage for default, loading, and authentication-error states.
• Summarize what changed and list unresolved UX concerns.

Task 3: Implement the general VM configuration screen

Build the general configuration screen from SPEC.md.

Fields:

• Node
• VM ID
  • Default: next free VM ID above 100
• Name
• Resource Pool
  • Default: empty
  • Empty is allowed
  • Load available resource pools from the server
  • The pool list may be empty
• Tags
  • List existing tags
  • Allow adding tags one by one
  • Allow removing tags
• High Availability (HA)
  • Default: enabled
• Start at boot
  • Default: disabled
• Start/shutdown order
  • Optional
• Startup delay
  • Optional
• Shutdown timeout
  • Optional

Definition of done:

• Add or update tests first.
• Cover default values, validation, optional fields, and empty pool lists.
• Add service tests for loading nodes, next VM ID, pools, and existing tags.
• Add Textual interaction tests using run_test() and Pilot.
• Add snapshot coverage for the default general screen and an empty-pool state.
• Summarize what changed and list unresolved UX concerns.

Task 4: Implement the OS selection screen

Build the OS configuration screen from SPEC.md.

Requirements:

• Ask whether the user wants to use installation media.
• Valid choices:
  • ISO
  • Physical disc drive
  • No
• If the user selects ISO:
  • Load available storages from the Proxmox API
  • Default storage to cephfs
  • When the storage changes, reload the available ISOs
  • The ISO list may be empty
  • Let the user choose the ISO
  • Default ISO to the latest available NixOS minimal ISO matching this pattern:
    • nixos-minimal-<two digit year>-<two digit month>.<some alphanumeric id or hash>.<architecture>-linux.iso
• Let the user select guest type and version
  • Default type: Linux
  • Default version: 6.x - 2.6 Kernel

Definition of done:

• Add or update tests first.
• Add service tests for storage and ISO loading and NixOS ISO default selection.
• Add Textual interaction tests for switching media type, storage, and ISO selection.
• Add snapshot coverage for:
  • default OS screen
  • ISO selected with results
  • ISO selected with no results
• Summarize what changed and list unresolved UX concerns.

Task 5: Implement the system configuration screen

Build the system configuration screen from SPEC.md.

Fields:

• Graphic card
  • Default: Default
• Machine
  • Default: q35
• Firmware BIOS
  • Default: OVMF (UEFI)
• Add EFI Disk
  • Default: enabled
• EFI storage
  • Default: ceph-pool
• Pre-Enroll keys
  • Default: disabled
• SCSI Controller
  • Default: VirtIO SCSI single
• Qemu Agent
  • Default: enabled
• TPM
  • Default: disabled

Definition of done:

• Add or update tests first.
• Model system settings separately from widgets.
• Add Textual interaction tests using run_test() and Pilot.
• Add snapshot coverage for the default system screen.
• Summarize what changed and list unresolved UX concerns.

Task 6: Implement the disks configuration screen

Build the disks configuration screen from SPEC.md.

Requirements:

• Allow the user to configure zero to multiple disks.
• Allow the user to add, modify, and remove disks one by one.
• For each disk, support:
  • Bus/Device
    • Default bus: SCSI
    • Default device: increasing integer starting from 0 such as scsi0, scsi1, ...
  • Storage
    • Default: ceph-pool
  • Disk size in GiB
    • Default: 32
  • Format
    • Default: RAW
    • Not changeable
  • Cache
    • Default: no cache
  • Discard
    • Default: disabled
  • IO Thread
    • Default: enabled
  • SSD emulation
    • Default: enabled
  • Backup
    • Default: enabled
  • Skip replication
    • Default: disabled
  • Async IO
    • Default: io_uring

Definition of done:

• Add or update tests first.
• Model disk configuration separately from widgets.
• Add service tests for available storage loading if required by the UI.
• Add Textual interaction tests for add, edit, and remove disk flows.
• Add snapshot coverage for the default disk state and a multi-disk state.
• Summarize what changed and list unresolved UX concerns.

Task 7: Implement the CPU configuration screen

Build the CPU configuration screen from SPEC.md.

Fields:

• Cores
  • Default: 2
• Sockets
  • Default: 1
• CPU Type
  • Default: host

Definition of done:

• Add or update tests first.
• Cover default values and validation.
• Add Textual interaction tests using run_test() and Pilot.
• Add snapshot coverage for the default CPU screen.
• Summarize what changed and list unresolved UX concerns.

Task 8: Implement the memory configuration screen

Build the memory configuration screen from SPEC.md.

Fields:

• Memory size in MiB
  • Default: 2048
• Min Memory in MiB
  • Default: same as Memory size
• Ballooning
  • Default: enabled
• Allow KSM
  • Default: enabled

Definition of done:

• Add or update tests first.
• Cover default values, derived defaults, and validation.
• Add Textual interaction tests using run_test() and Pilot.
• Add snapshot coverage for the default memory screen.
• Summarize what changed and list unresolved UX concerns.

Task 9: Implement the network configuration screen

Build the network configuration screen from SPEC.md.

Fields:

• No Network device
  • Default: disabled
• Bridge
  • Default: vmbr9
• VLAN Tag
  • Default: none
• Model
  • Default: virtio
• MAC Address
  • Default: auto-generated by the API
• Firewall
  • Default: enabled
• Disconnected
  • Default: disabled
• MTU
  • Default: none
• Rate Limit in MB/s
  • Default: none
• Multiqueue
  • Default: none

Definition of done:

• Add or update tests first.
• Cover defaults, validation, and no-network behavior.
• Add Textual interaction tests using run_test() and Pilot.
• Add snapshot coverage for the default network screen and the no-network state.
• Summarize what changed and list unresolved UX concerns.

Task 10: Implement the confirmation screen

Build the confirmation screen from SPEC.md.

Requirements:

• Display a summary of the full VM configuration collected in previous steps.
• Clearly show all relevant settings before submission.
• Show validation issues or missing required inputs.
• Provide a Create VM button.

Definition of done:

• Add or update tests first.
• Add Textual interaction tests for reaching the confirmation step.
• Add snapshot coverage for a fully populated confirmation screen and a validation-error state.
• Summarize what changed and list unresolved UX concerns.

Task 11: Implement VM creation against the Proxmox API

Submit the VM creation request after confirmation.

Requirements:

• Translate the collected workflow state into the correct Proxmox API request payload.
• Use the Proxmox VE API to create the VM.
• After VM creation succeeds, send the required follow-up configuration request(s) to:
  • add serial0: socket
  • set vga: serial0
• Treat serial-console configuration as part of the overall success criteria because it is required for using the Proxmox VE xterm.js serial console.
• Handle API request errors cleanly.
• Distinguish between:
  • VM creation failure before the VM exists
  • post-creation serial-console configuration failure after the VM already exists
• Show progress, success, and failure states.
• Preserve enough information in the UI so the user can understand what failed.

Definition of done:

• Add or update tests first.
• Add service tests for payload building, request sequencing, and error handling.
• Add Textual interaction tests for submission, full success, VM-create failure, and post-create serial-console failure flows.
• Add snapshot coverage for submission, success, VM-create error, and post-create serial-console error states.
• Summarize what changed and list unresolved UX concerns.

Task 12: Polish navigation, error handling, and documentation

Improve the overall wizard experience and repository documentation.

Requirements:

• Ensure step-to-step navigation is clear and keyboard-friendly.
• Make back/next/confirm actions predictable across screens.
• Standardize loading, empty, success, and error messaging.
• Update README.md if the real module name or run command differs from the placeholder.
• Document any remaining constraints, assumptions, or known gaps.

Definition of done:

• Add or update tests first where behavior changes.
• Add or update interaction tests for navigation flows.
• Add snapshot coverage for any changed visual states.
• Summarize what changed and list unresolved UX concerns.

API references

• Proxmox API documentation: https://pve.proxmox.com/wiki/Proxmox_VE_API
• Proxmox API schema / viewer: https://pve.proxmox.com/pve-docs/api-viewer/index.html

Suggested prompt template for Codex

Use this prompt pattern for each task:

Implement TASKS.md task N. First add or update tests. Use Textual interaction tests with run_test() and Pilot. Add snapshot coverage for relevant default, loading, empty, success, and error states. Use uv for commands and ruff for linting / formatting. Keep Proxmox API logic in a service layer. Run uv run pytest, uv run ruff check ., and uv run ruff format .. Summarize what changed and list unresolved UX concerns.