table2llm

Parse complex tables from CSV, XLSX, and DOCX into LLM-friendly formats such as JSON, Markdown, and HTML.

Features

• Supports:
  • CSV
  • XLSX (Excel)
  • DOCX (Word tables)
• Handles:
  • Merged cells (rowspan / colspan)
  • Nested tables (DOCX)
• Outputs:
  • JSON (structured records)
  • Markdown
  • HTML

Installation

pip install table2llm

Usage

Unified API (Recommended)

from table2llm import unpack

result = unpack("data.csv", output_format="json")
print(result)

result = unpack("data.xlsx", output_format="md")
print(result)

result = unpack("data.docx", output_format="html")
print(result)

Direct Parser API

from table2llm import CsvParser, DocxParser, XlsxParser

csv_result = CsvParser().parse_to_json("data.csv")
xlsx_result = XlsxParser().parse_to_md("data.xlsx")
docx_result = DocxParser().parse_to_html("data.docx")

print(csv_result)
print(xlsx_result)
print(docx_result)

Output Formats

JSON

• Structured as list of records
• Suitable for LLM ingestion

Markdown

• Clean table format
• Useful for prompts or documentation

HTML

• Preserves structure (rowspan / colspan)
• Best for rendering

Notes

• Choose parser based on file type
• Empty rows are skipped
• Text is normalized and cleaned internally