carloforte 0.1.0

Extract structured data from Excel files with minimal token usage

carloforte

Extract structured data from Excel files with minimal token usage.

carloforte uses an island-detection algorithm to convert Excel sheets into a compact intermediate representation (CSV, Markdown, or JSON), making it efficient to pass spreadsheet data to LLMs.

Installation

TBC

Usage

import carloforte

# Extract all sheets as CSV (default)
text = carloforte.extract("data.xlsx")

# Extract specific sheets as Markdown
text = carloforte.extract("data.xlsx", sheets=["Revenue", "Costs"], fmt="markdown")

# Extract as JSON
text = carloforte.extract("data.xlsx", fmt="json")

Formats

Format	Best for
csv	Compact, low token count
markdown	Readable, good for LLM prompts
json	Structured output, programmatic use

CLI

carloforte data.xlsx --fmt markdown
carloforte data.xlsx --sheets Revenue Costs --fmt json

How it works

Excel sheets often contain multiple disconnected tables, empty rows, and metadata scattered around. carloforte detects each contiguous block of data ("island") independently and serialises only what matters — reducing token usage by 60–75% compared to passing raw Excel content to an LLM.

Architecture

flowchart LR
    A["📄 .xlsx file"] --> B["_reader\nload sheets"]
    B --> C["dict[sheet → grid]"]
    C --> D["_islands\nBFS detection"]
    D --> E["dict[sheet → islands]"]
    E --> F{"fmt?"}
    F -->|csv| G["CSV"]
    F -->|markdown| H["Markdown"]
    F -->|json| I["JSON"]

License

MIT