# Prompt Files Cleanup Plan

## Current Status

### Files:
1. **prompt_english.txt** (10K) ✅ **CURRENTLY USED** - Master prompt
2. prompt_with_custom_data.txt (10K) - Dutch duplicate with same features
3. prompt.txt (8.3K) - Old Dutch basic prompt
4. prompt_improved.txt (8.3K) - Old duplicate
5. prompt_original_backup.txt (2.6K) - Very old backup

### What's Used:
```python
# analyze_from_urls_optimized.py line 98:
with open("prompt_english.txt", "r", encoding="utf-8") as f:
    base_prompt = f.read()
```

## Recommendation: KEEP ONLY ONE

### Keep:
- **prompt_english.txt** - The unified master prompt with all improvements:
  - ✅ English language (better GPT-4o-mini results)
  - ✅ Custom criteria data placeholders
  - ✅ SHORT-STAY VACATION RENTAL emphasis
  - ✅ LIVABILITY ASSESSMENT
  - ✅ HABITABILITY requirements
  - ✅ All 6 criteria defined
  - ✅ "Major renovation = LOW SCORE" penalties

### Archive (move to _archive/):
- prompt_with_custom_data.txt - Dutch duplicate
- prompt.txt - Old basic Dutch
- prompt_improved.txt - Old version
- prompt_original_backup.txt - Very old backup

## Cleanup Actions

```bash
cd /Users/jonathan/SynologyDrive/Since\ Today/PROJECTEN/farmmatch/scraper

# Create archive directory
mkdir -p _archive/prompts

# Move old prompts
mv prompt_with_custom_data.txt _archive/prompts/
mv prompt.txt _archive/prompts/
mv prompt_improved.txt _archive/prompts/
mv prompt_original_backup.txt _archive/prompts/

# Result: Only prompt_english.txt remains
```

## Benefits

1. **Clarity** - One prompt file to maintain
2. **No Confusion** - Clear which prompt is active
3. **Easier Updates** - Only one file to edit
4. **Consistent** - All analyses use same prompt
5. **English** - Best language for GPT-4o-mini

## If Dutch Needed Later

Can always translate prompt_english.txt back to Dutch if needed, but currently:
- English gives better GPT results
- All improvements already in English version
- System configured to use English
- No reason to maintain Dutch duplicates
