# Vector Text Fixer

> Fix garbled text in PDF/SVG vector graphics for final editing in AI.

- Skill: `dvcrn/vector-text-fixer` (Agent Skill, multi-file: 3 files)
- Install (CLI): `npx skillmds@latest add dvcrn/vector-text-fixer`
- Raw SKILL.md: https://api.skillmd.com/api/skills/dvcrn/vector-text-fixer/raw
- Safety review: pending (external: skill-scanner PASS, skillspector PASS)
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Docs & Writing
- Author: dvcrn (https://skillmd.com/u/dvcrn)
- Updated: 2026-09-08
- Page: https://skillmd.com/skills/dvcrn/vector-text-fixer

---


# Vector Text Fixer

Fixes garbled text in PDF/SVG vector graphics to make them editable in AI tools.

## Features

- **Garbled Text Detection**: Automatically identifies garbled text in PDF/SVG files
- **Smart Repair**: Infers original text content based on context
- **Batch Processing**: Supports batch processing of multiple files in a folder
- **Format Preservation**: Repaired files maintain original vector format and layout
- **AI-assisted Editing**: Outputs intermediate format that can be imported into AI editors

## Supported Scenarios

### 1. PDF Garbled Text Repair
- Box/question mark issues caused by font embedding problems
- Garbled text caused by encoding conversion errors
- Abnormal characters generated by missing font substitution
- Multi-language mixed encoding issues

### 2. SVG Garbled Text Repair
- Text entity encoding errors
- Special character escaping issues
- Display abnormalities caused by invalid font references
- XML encoding declaration errors

## Usage

### Command Line

```bash
# Fix a single PDF file
python scripts/main.py --input document.pdf --output fixed.pdf

# Fix a single SVG file
python scripts/main.py --input diagram.svg --output fixed.svg

# Batch process folder
python scripts/main.py --batch ./input_folder --output ./output_folder

# Interactive repair (manually specify replacement content)
python scripts/main.py --input doc.pdf --interactive

# Export as editable format (JSON)
python scripts/main.py --input doc.pdf --export-json editable.json
```

### Python API

```python
from scripts.main import VectorTextFixer

# Create fixer instance
fixer = VectorTextFixer()

# Fix PDF
result = fixer.fix_pdf("input.pdf", "output.pdf")

# Fix SVG
result = fixer.fix_svg("input.svg", "output.svg")

# Batch processing
results = fixer.batch_fix("./input_folder", "./output_folder")

# Get text map (for AI editing)
text_map = fixer.extract_text_map("input.pdf")
```

## Input Parameters

| Parameter | Type | Required | Description |
|------|------|------|------|
| `--input` | str | Yes* | Input file path (PDF or SVG) |
| `--batch` | str | No | Batch processing input folder |
| `--output` | str | Yes* | Output file/folder path |
| `--interactive` | bool | No | Enable interactive repair mode |
| `--export-json` | str | No | Export editable JSON format |
| `--encoding` | str | No | Specify source file encoding (default: auto-detect) |
| `--font-substitution` | dict | No | Font replacement mapping |
| `--repair-level` | str | No | Repair level: minimal, standard, aggressive (default: standard) |

*At least one of --input and --batch is required

## Output Format

### Repaired PDF/SVG
- Maintains original vector format
- Garbled text replaced with readable content
- Fonts and layout remain unchanged

### JSON Export Format
```json
{
  "file_type": "pdf",
  "pages": [
    {
      "page_num": 1,
      "text_blocks": [
        {
          "id": "tb_001",
          "bbox": [100, 200, 300, 220],
          "original_text": "�����",
          "detected_encoding": "UTF-8",
          "confidence": 0.3,
          "suggested_fix": "Sample Text"
        }
      ]
    }
  ],
  "fonts_used": ["Arial", "SimSun"],
  "repair_summary": {
    "total_blocks": 15,
    "fixed_blocks": 12,
    "skipped_blocks": 3
  }
}
```

## Garbled Text Detection Rules

The tool uses the following rules to detect garbled text:

1. **Replacement Character Detection**: Identifies U+FFFD (�) and box characters
2. **Control Character Filtering**: Excludes non-printing control characters
3. **Encoding Consistency**: Detects anomalies caused by mixed encodings
4. **Font Fallback Detection**: Identifies substitution characters generated due to missing fonts
5. **Probability Model**: Garbled text probability assessment based on character frequency

## Repair Strategies

### Minimal
- Only repairs obvious errors (replacement characters, null bytes)
- Maintains maximum integrity of original text
- Suitable for minor garbled text issues

### Standard
- Repairs common encoding issues
- Smart font replacement
- Balances repair rate and accuracy

### Aggressive
- Comprehensive text re-encoding
- Uses OCR-assisted recognition
- Suitable for severely garbled documents

## Examples

### Fix Single Page PDF

**Input**:
```bash
python scripts/main.py --input report.pdf --output fixed_report.pdf
```

**Output**:
```
✓ Processing: report.pdf
✓ Detected 5 garbled text blocks
✓ Fixed 4 blocks automatically
⚠ 1 block requires manual review
✓ Output saved: fixed_report.pdf
✓ Report saved: fixed_report_repair_log.json
```

### Export Editable JSON

**Input**:
```bash
python scripts/main.py --input diagram.svg --export-json editable.json
```

**Output JSON Structure**:
```json
{
  "file_type": "svg",
  "svg_info": {
    "width": 800,
    "height": 600,
    "viewBox": "0 0 800 600"
  },
  "text_elements": [
    {
      "id": "text_1",
      "x": 100,
      "y": 200,
      "font_family": "Arial",
      "font_size": 14,
      "original": "�����",
      "user_editable": "",
      "confidence": 0.25
    }
  ]
}
```

## Dependencies

```
pdfplumber>=0.10.0      # PDF parsing
PyMuPDF>=1.23.0         # PDF processing (fitz)
cairosvg>=2.7.0         # SVG conversion
beautifulsoup4>=4.12.0  # SVG parsing
fonttools>=4.40.0       # Font processing
chardet>=5.0.0          # Encoding detection
Pillow>=10.0.0          # Image processing
```

## Limitations

- Encrypted PDFs require password unlock before processing
- Severely damaged vector files may not be fully repairable
- Some rare fonts may not map correctly
- Scanned PDFs require OCR recognition first

## Version Information

- **Version**: 1.0.0
- **Last Updated**: 2026-02-06
- **Status**: Ready for use

## Risk Assessment

| Risk Indicator | Assessment | Level |
|----------------|------------|-------|
| Code Execution | Python/R scripts executed locally | Medium |
| Network Access | No external API calls | Low |
| File System Access | Read input files, write output files | Medium |
| Instruction Tampering | Standard prompt guidelines | Low |
| Data Exposure | Output files saved to workspace | Low |

## Security Checklist

- [ ] No hardcoded credentials or API keys
- [ ] No unauthorized file system access (../)
- [ ] Output does not expose sensitive information
- [ ] Prompt injection protections in place
- [ ] Input file paths validated (no ../ traversal)
- [ ] Output directory restricted to workspace
- [ ] Script execution in sandboxed environment
- [ ] Error messages sanitized (no stack traces exposed)
- [ ] Dependencies audited
## Prerequisites

```bash
# Python dependencies
pip install -r requirements.txt
```

## Evaluation Criteria

### Success Metrics
- [ ] Successfully executes main functionality
- [ ] Output meets quality standards
- [ ] Handles edge cases gracefully
- [ ] Performance is acceptable

### Test Cases
1. **Basic Functionality**: Standard input → Expected output
2. **Edge Case**: Invalid input → Graceful error handling
3. **Performance**: Large dataset → Acceptable processing time

## Lifecycle Status

- **Current Stage**: Draft
- **Next Review Date**: 2026-03-06
- **Known Issues**: None
- **Planned Improvements**: 
  - Performance optimization
  - Additional feature support