Tablassist CLI

Python CLI tool for AI-assisted Tablassert table configuration generation — entity resolution, YAML validation, and Biolink documentation lookup.

Tablassist ships with two document extraction modes:

• extract-text for fast raw extraction via Textract
• extract-text-semantic for richer Docling-backed semantic extraction with Markdown output and ocr=auto by default

Installation

pip install tablassist

The base install now includes Docling, so extract-text-semantic works without a separate helper script or optional extra.

An optional extra is available for CPU compatibility:

pip install "tablassist[rtcompat]"  # Polars build for CPUs without required instructions

Requirements

• Python >= 3.13
• Environment variables TABLASSIST_USERNAME and TABLASSIST_API_KEY for API-accessing commands

Usage

# Fetch table configuration documentation
tablassist docs-table-config

Entity Resolution

# Search for entity CURIEs by term
tablassist search-curies "breast cancer"

# Search gene CURIEs within an NCBI taxon
tablassist search-gene-curies "BRCA1" --ncbi-taxon 9606

# Resolve an NCBI Taxon ID from an organism name
tablassist resolve-taxon-id "Homo sapiens"

Biolink Reference

# List all supported categories, predicates, or qualifiers
tablassist list-categories
tablassist list-predicates
tablassist list-qualifiers

# Fetch documentation for a specific Biolink element
tablassist docs-category "Gene"
tablassist docs-predicate "interacts_with"
tablassist docs-qualifier "qualified_predicate"

YAML Validation

Full config validation requires template: as the top-level key, with optional sections:. Use validate-section-str only for individual section mappings, not for whole config files.

# Validate a full config file
tablassist validate-config-file config.yaml

# Validate a single section from a YAML string
tablassist validate-section-str '<yaml>'

# Validate a full config from a YAML string
tablassist validate-config-str '<yaml>'

# Get the Section JSON schema
tablassist section-schema

Data Preview

# List sheets in an Excel file
tablassist excel-sheets data.xlsx

# Preview rows from an Excel sheet
tablassist preview-excel data.xlsx "Sheet1" 10

# Preview rows from a CSV file
tablassist preview-csv data.csv 10

# Extract text from a document (PDF, DOCX, etc.)
tablassist extract-text document.pdf

# Extract semantic Markdown from a document with Docling
tablassist extract-text-semantic document.pdf

# Extract plain text and explicitly disable OCR
tablassist extract-text-semantic document.pdf text off

extract-text is optimized for fast, low-overhead text grabs.

extract-text-semantic runs IBM Docling directly from the CLI module. It is the better choice when reading order, headings, lists, or table-aware Markdown matter more than raw speed.

Arguments for extract-text-semantic:

• file — local document path
• output_format — markdown (default) or text
• ocr — auto (default), off, or on

Use ocr=auto for the default balance. Use ocr=on for scans and image-heavy PDFs, and ocr=off when you know the source is born-digital and want the lightest path.

PMC Archive Download

# Download and extract a PMC tar archive
tablassist download-pmc-tar 12345 --dest-dir ./output

Development

uv sync                              # install dependencies
uv run ruff check .                  # lint
uv run ruff check --fix .            # lint with auto-fix
uv run ruff format .                 # format
uv run pyright                       # type check
uv run --group dev python -m pytest  # run all tests

License

Apache License 2.0