[FEAT] Support architecture presets in dynamic project generation

[bnbong]

Requesting Merging

Description

Part of #41
Closes #45

This PR connects the architecture preset selected in fastkit init --interactive to the actual project generation flow.

Instead of always deploying the same base scaffold, interactive init now chooses a preset-specific template, writes generated config files into preset-appropriate paths, preserves richer template main.py files when needed, and surfaces compatibility warnings for feature combinations that still require manual wiring.

Type of Change

• BUG FIX
• ADDING NEW TEMPLATE
• FEATURE ADDED/UPDATED
• HOTFIX
• DELETING UNNECESSARY FEATURES
• DOCUMENTATION & DEVOPS
• Etc..

Test Environment

• Local macOS (Apple Silicon / M1), zsh
• Python via uv
• Verified commands:
  • uv run pytest tests/test_backends/test_preset_layout.py tests/test_cli_operations/test_cli_interactive_integration.py -q
    • Result: 32 passed
  • uv run mkdocs build
    • Result: Documentation built in 14.87 seconds

Major Changes

• Added preset-aware generation strategy

  • Introduced PresetLayoutStrategist and PresetProfile
  • Centralized preset-specific decisions for:
    • base template selection
    • main.py regeneration vs preservation
    • database config output path
    • auth config output path
    • compatibility warnings for preserve-main presets

• Connected interactive init to architecture presets

  • fastkit init --interactive now uses the selected preset to choose the base template instead of always deploying fastapi-empty
  • Unknown or missing preset ids safely fall back to minimal to preserve backward compatibility

• Implemented preset-specific layout behavior

  • minimal -> fastapi-empty
  • single-module -> fastapi-single-module
  • classic-layered -> fastapi-default
  • domain-starter -> fastapi-domain-starter

• Added selective main.py overlay behavior

  • minimal and single-module regenerate main.py from selected features
  • classic-layered and domain-starter preserve the template-shipped main.py
  • This avoids clobbering router-aware template entrypoints for richer presets

• Added preset-specific config file targets

  • Database and auth config files are now written into locations that match each preset’s structure
  • Examples:
    • minimal / single-module: src/config/...
    • classic-layered: src/core/...
    • domain-starter: src/app/core/...

• Added compatibility warnings for preserve-main presets

  • The CLI now warns when selected features depend on dynamic main.py wiring that is intentionally skipped for preserved-main presets
  • Current warning targets include:
    • Rate-Limiting
    • Prometheus
  • CORS is documented separately as already wired in shipped templates and activated via .env configuration

• Added reference documentation for the supported matrix

  • New docs page: docs/en/reference/preset-feature-matrix.md
  • Documents:
    • preset -> base template mapping
    • generated file locations
    • supported feature behavior per preset
    • manual wiring expectations for preserve-main presets
  • Added the page to MkDocs navigation
  • Linked it from the interactive-mode FAQ

• Added regression coverage

  • New strategist unit tests for:
    • supported preset ids
    • base template mapping
    • regeneration policy
    • output paths
    • compatibility warnings
  • Expanded CLI interactive integration tests for:
    • per-preset layout generation
    • database config placement
    • preserved vs regenerated main.py
    • compatibility warning behavior

Screenshots (optional)

N/A

Etc

• This PR implements the generation-layer behavior for the architecture preset step introduced in [FEATURE] Add architecture preset selection to interactive init #44.
• The fallback behavior for callers without architecture_preset is intentionally minimal to preserve pre-[FEATURE] Support architecture presets in dynamic project generation #45 interactive init behavior.
• The preset/feature matrix docs are intended to be the source of truth for current preset support and manual-wiring expectations.

[bnbong]

@codex review

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.

📢 Thoughts on this report? Let us know!

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: b48b4db24e

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[chatgpt-codex-connector]

Derive Docker entrypoint from preset main module

Now that interactive init can select domain-starter as the base template, this line allows projects whose app entrypoint is src/app/main.py; however Docker generation still hardcodes src.main:app in DynamicConfigGenerator._generate_dockerfile. When users choose Docker deployment with domain-starter, the generated container fails to start because src.main does not exist. Please thread the preset-aware module path into Docker file generation (or derive it from the strategist) so the CMD matches the selected layout.

Useful? React with 👍 / 👎.