# Estonian source identities

The following publisher files were downloaded directly and verified by byte length and SHA-256 before generation. Raw source files are not redistributed here.

| Role | Source | Bytes | SHA-256 |
| --- | --- | ---: | --- |
| Sole positive admission | [EKI ühendsõnastik 2024 headwords](https://arhiiv.eki.ee/tarkvara/wordlist/ys_public_est.json) | 6,994,820 | `5359dc9df5621fc54760727b0b8cfd7a77919d29fa891987976c101c6aa92aba` |
| Frequency threshold and rank | [ENC 2021 lemma frequencies](https://arhiiv.eki.ee/tarkvara/enc21/21.lemmad.frq.txt) | 17,109,379 | `ed29906db44fefffe609defccb4a6d4d864e9b773956b29fe415645258c5431e` |
| POS classification and exclusion | [ENC 2021 tagged lemma frequencies](https://arhiiv.eki.ee/tarkvara/enc21/21.lemmad.sl.frq.txt) | 19,353,011 | `128c94eb7b3199b458268273946cc9d875f852c3a4f858d4d8af000b83e3109f` |

The ENC resource is attributed to EKI and Indrek Hein: [DOI 10.15155/3-00-0000-0000-0000-08D1FL](https://doi.org/10.15155/3-00-0000-0000-0000-08D1FL). It supplies signals only; a corpus occurrence cannot admit a word.