TBL Metadata Parser - Data Structure Extraction

Instructions for MigrationArchitect Agent

When tasked with analyzing a .tbl file (e.g., ships.tbl), you MUST use the scripts/parse_tbl.py skill.

The skill will return a structured JSON schema (e.g., ships_schema.json). This JSON is the required input for the generate-resource skill.

When to Use

Use this skill whenever you encounter a .tbl file during migration planning. This is a Plan phase skill that extracts structured data from legacy plain-text database files.

What This Skill Does

This skill deconstructs the legacy plain-text database files into a structured, engine-agnostic JSON schema, forming the first step of the data-driven pipeline.

Extracts These Data Patterns

• Object Delimiters: Uses $Name: as the delimiter to identify the start of a new object
• Key-Value Pairs: Parses all subsequent lines as key-value pairs (splitting on :)
• Data Type Inference: Performs intelligent data type "guessing" to convert text values to:
  • float (for numeric values with decimals)
  • int (for integer values)
  • array (for comma-separated values)
  • string (for text values)

Skips Non-Data Content

• Comments and blank lines
• Section headers that don't contain key-value data
• Malformed entries (in non-strict mode)

Expected Output Format

The skill returns structured JSON data like this:

{
  "schema_type": "ship_stats",
  "entries": {
    "GTF Apollo": {
      "$Name": "GTF Apollo",
      "$Mass": 20000.0,
      "$Max Velocity": 70.0,
      "$Shields": 150.0,
      "$Armor": 100.0,
      "Weapons": [
        "Subach HL-7",
        "Tempest"
      ]
    },
    "GTF Artemis": {
      "$Name": "GTF Artemis",
      "$Mass": 18500.0,
      "$Max Velocity": 75.0,
      "$Shields": 125.0,
      "$Armor": 85.0
    }
  }
}

Usage Instructions

Basic Usage

python scripts/parse_tbl.py input/ships.tbl > output/ships_schema.json

The skill handles:

1. Reading the .tbl file in text mode ('r')
2. Using $Name: as the delimiter to identify the start of a new object
3. Parsing all subsequent lines as key-value pairs (splitting on :)
4. Performing intelligent data type "guessing" to convert text values to appropriate types
5. Building a nested dictionary structure (e.g., {"GTF Apollo": {"$Mass": 20000.0, ...}})
6. Returning the data as JSON via print(json.dumps(schema_dict))

Data Type Mapping

• Text with decimals → float
• Text with only digits → int
• Comma-separated values → Array
• Text in quotes → string
• Other text → string

Next Steps in Pipeline

After using this skill:

1. Save the JSON output to a file
2. Pass the JSON file path to GDScriptEngineer
3. GDScriptEngineer will use generate-resource skill with this data
4. The generate-resource skill will create Custom Resource scripts with @export properties

Critical: Do NOT Use For

• Parsing binary files (use parse-pof instead)
• Converting 3D geometry (not relevant for table data)
• Extracting metadata from model files (handled separately)

This skill focuses solely on extracting structured data from plain-text table files, providing the clean data foundation for the resource generation pipeline.