CSV & XLSX

Map real-world localization tables without hard-coding a single spreadsheet layout.

Map columns explicitly

CSV and XLSX share column controls. Selectors accept auto, a header name, or spreadsheet letters. XLSX additionally accepts sheet_name and sheet_index. Extra columns can be retained in Data.extensions.

tabular.py

import lokit
document = lokit.parse.csv(    "messages.csv",    source_locale="en-US",    target_locale="fr-FR",    source_column="source",    target_column="translation",    id_column="key",    status_column="status",)

Handle multilingual tables

A table with several language columns can become one multilingual document. csv_targets() and xlsx_targets() instead return a dictionary of independent single-target documents.

multilingual_xlsx.py

documents = lokit.parse.xlsx_targets(    "messages.xlsx",    source_locale="en-US",    target_columns={"fr-FR": "French", "de-DE": "German"},    id_column="Key",)

Output behavior

Tabular formats have materialized and async parsers, but no synchronous document stream. Writers support header style, column names, inclusion flags, and output order. Regeneration reuses the original column mapping and sheet selection.