TableReader

Automatically extract clean, typed data from messy Excel and CSV files using LLM-powered table detection.

Installation

pip install git+https://github.com/mehdigmira/tablereader.git

Usage

Set environment variables:

export OPENAI_API_KEY="your-deepseek-api-key"
export OPENAI_BASE_URL="https://api.deepseek.com"

Then

from tablereader import read

# Uses OpenAI by default (requires OPENAI_API_KEY env var)
result = read("data.xlsx", sheet_name="Sheet1")

# Process a CSV file
result = read("data.csv", is_csv=True)

# Iterate through extracted tables
for table in result.tables:
for row in table:
print(row)  # Each row is a dict with typed values

Features

The read() function automatically:

• Detects table boundaries (skips headers, footers, totals)
• Identifies column types (numbers, dates, times, strings)
• Handles messy formatting (currency symbols, percentages, thousand separators)
• Streams data efficiently for large files

Requirements

• Python 3.12+
• OpenAI API key or any OpenAI-compatible API (DeepSeek, etc.)