AgentSkillsCN

data-type-converter

在 JSON、CSV、XML、YAML、TOML 等数据格式之间转换。支持嵌套结构、数组,并尽可能保留数据类型。

SKILL.md
--- frontmatter
name: data-type-converter
description: Convert between data formats (JSON, CSV, XML, YAML, TOML). Handles nested structures, arrays, and preserves data types where possible.

Data Type Converter

Convert data between JSON, CSV, XML, YAML, and TOML formats. Handles nested structures, arrays, and complex data with intelligent flattening options.

Quick Start

python
from scripts.data_converter import DataTypeConverter

# JSON to CSV
converter = DataTypeConverter()
converter.convert("data.json", "data.csv")

# YAML to JSON
converter.convert("config.yaml", "config.json")

# With options
converter.convert("data.json", "data.csv", flatten=True)

Features

  • 5 Formats: JSON, CSV, XML, YAML, TOML
  • Nested Data: Flatten or preserve nested structures
  • Arrays: Handle array data intelligently
  • Type Preservation: Maintain data types where possible
  • Pretty Output: Formatted, human-readable output
  • Batch Processing: Convert multiple files

API Reference

Basic Conversion

python
converter = DataTypeConverter()

# Auto-detect format from extension
converter.convert("input.json", "output.csv")
converter.convert("input.xml", "output.json")
converter.convert("input.yaml", "output.toml")

With Options

python
# Flatten nested structures for CSV
converter.convert("nested.json", "flat.csv", flatten=True)

# Pretty print output
converter.convert("data.json", "pretty.json", indent=4)

# Specify root element for XML
converter.convert("data.json", "data.xml", root="records")

Programmatic Access

python
# Load and convert in memory
data = converter.load("data.json")
converter.save(data, "data.yaml")

# String conversion
json_str = '{"name": "John", "age": 30}'
yaml_str = converter.convert_string(json_str, "json", "yaml")

Batch Processing

python
# Convert all JSON files to CSV
converter.batch_convert(
    input_dir="./json_files",
    output_dir="./csv_files",
    output_format="csv"
)

CLI Usage

bash
# Basic conversion
python data_converter.py --input data.json --output data.csv

# With flattening
python data_converter.py --input nested.json --output flat.csv --flatten

# Batch convert
python data_converter.py --input-dir ./json --output-dir ./csv --format csv

# Pretty print
python data_converter.py --input data.json --output pretty.json --indent 4

CLI Arguments

ArgumentDescriptionDefault
--inputInput fileRequired
--outputOutput fileRequired
--input-dirInput directory for batch-
--output-dirOutput directory-
--formatOutput formatFrom extension
--flattenFlatten nested dataFalse
--indentIndentation spaces2
--rootXML root elementroot

Conversion Matrix

From/ToJSONCSVXMLYAMLTOML
JSON-YesYesYesYes
CSVYes-YesYesYes
XMLYesYes-YesYes
YAMLYesYesYes-Yes
TOMLYesYesYesYes-

Examples

JSON to CSV (Flat)

python
converter = DataTypeConverter()

# Input: data.json
# [{"name": "John", "age": 30}, {"name": "Jane", "age": 25}]

converter.convert("data.json", "data.csv")

# Output: data.csv
# name,age
# John,30
# Jane,25

Nested JSON to Flat CSV

python
# Input: nested.json
# [{"user": {"name": "John", "email": "j@test.com"}, "orders": 5}]

converter.convert("nested.json", "flat.csv", flatten=True)

# Output: flat.csv
# user.name,user.email,orders
# John,j@test.com,5

YAML Config to JSON

python
# Input: config.yaml
# database:
#   host: localhost
#   port: 5432
# debug: true

converter.convert("config.yaml", "config.json")

# Output: config.json
# {"database": {"host": "localhost", "port": 5432}, "debug": true}

XML to JSON

python
# Input: data.xml
# <users>
#   <user><name>John</name><age>30</age></user>
# </users>

converter.convert("data.xml", "data.json")

# Output: data.json
# {"users": {"user": {"name": "John", "age": "30"}}}

Dependencies

code
pyyaml>=6.0
toml>=0.10.0
xmltodict>=0.13.0
pandas>=2.0.0

Limitations

  • CSV doesn't support nested data (requires flattening)
  • XML attribute handling is basic
  • TOML doesn't support null values
  • Very deep nesting may cause issues with some formats
  • Array handling varies by format