feat: enhanced AI code generator with multi-provider support and optimized prompts (#2146)

Multi-provider API support:
- Per-provider test buttons in config (OpenAI, Anthropic, Gemini, DeepSeek, Custom)
- Optional API key for custom base URL (Ollama, LM Studio)
- Provider icons in config and code generator
- Human-readable HTTP error messages with API response details
- Disable reasoning/thinking for local models (reasoning_effort: none)

Two-step plan-then-code generation:
- Step 1: Model analyzes task and creates implementation plan with device IDs
- Step 2: Model generates code based on plan + API examples
- Collapsible plan view in UI for debugging
- Status display ("Planning..." / "Generating code...")
- 600s timeout for local models

Optimized prompts for small local models (tested with qwen2.5-coder:14b):
- Concrete code examples showing correct on(), setState(), getState() syntax
- WRONG/CORRECT pairs for common mistakes (adapter.set, console.log, etc.)
- Compact function signature list covering all 80+ API functions
- FORBIDDEN list preventing wrong patterns
- Task placed after device list (recency bias)
- TODO_DEVICE_ID placeholder for missing devices

Build fixes:
- Node 25 compatibility: rmSync instead of rmdirSync({recursive:true})
- admin/img/ preserved during clean builds
- Flexible result area height (flex:1 instead of fixed calc)

New documentation files:
- docs-compact.md: full API in 12.5KB (used for code generation)
- docs-essential.md: core functions with examples in 2.6KB
- Original docs.md preserved for reference

Co-authored-by: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Co-authored-by: Bluefox <dogafox@gmail.com>

Author: 3 people

CLAUDE.md

@@ -26,7 +26,21 @@ Executes Javascript, Typescript Scripts.
 <!--
     ### **WORK IN PROGRESS**
 -->
-### **WORK IN PROGRESS**
+### WORK IN PROGRESS
+* Per-provider test buttons in adapter config (OpenAI, Anthropic, Gemini, DeepSeek, Custom API)
+* Optional API key field for custom base URL providers (e.g. Ollama without auth)
+* Provider icons on test buttons and in model dropdown
+* Human-readable HTTP error messages with API response details
+* Two-step AI code generation: plan first, then generate code
+* Collapsible plan view in AI code generator UI
+* Status display during generation ("Planning..." / "Generating code...")
+* Optimized prompts with code examples for better results with small local models
+* Compact API reference (docs-compact.md) for reduced context usage
+* Disable reasoning/thinking for local models (reasoning_effort: none)
+* TODO_DEVICE_ID placeholder when a required device is not in the device list
+* Node 25 compatibility: replaced deprecated rmdirSync with rmSync in build tasks
+* Flexible result area height in AI code generator (no extra scrollbar)
+* Added translations for all 11 languages
 * (@GermanBluefox) Added support for plain import/export
 * (@GermanBluefox) Correcting error in configuration
 * (@GermanBluefox) disallow writing into node_modules folder by scripts