SynthOCR
Copyright (c) 2026 Ardalan Jalal

The SynthOCR source code is licensed under either of the MIT License
(LICENSE-MIT) or the Apache License, Version 2.0 (LICENSE-APACHE), at your
option.

The code license does not cover the following, which keep their own terms:

- Fonts in fonts/open/: SIL Open Font License 1.1. Each family includes its
  OFL.txt; synthocr/font_catalog.json pins the upstream commit and checksums.
- Tatoeba sentences in data/raw/tatoeba/: CC BY 2.0 France. Contributor
  attribution is kept with every record and in each dataset's origins.jsonl.
- Synthetic template text in data/raw/synthetic-v2/ and the Kurdish spelling
  cases in data/raw/kurdish-spelling-v1/: CC0 1.0, as recorded per record.
- data/raw/examples.jsonl: example text written for this project, licensed
  with the code.
- data/raw/kurdish_corpus.txt, data/raw/legacy_text.txt, data/raw/symbols.txt
  and data/raw/char_map.json: original inputs whose upstream source and
  redistribution terms were not recorded. No license is granted for them.
- The Kurdish reference characters in synthocr/kurdish_characters.json come
  from https://github.com/ardalan-jalal/font-checking.

Datasets you generate inherit the terms of the text and fonts used to make them.
See data/SOURCES.md and fonts/README.md for details.
