nlqk.defaults

defaults.py

(C) 2025 Damir Cavar and NLP Lab

Defaults for the NLQK library.

 1#!/usr/bin/env python
 2# -*- coding: utf-8 -*-
 3
 4"""
 5defaults.py
 6
 7(C) 2025 [Damir Cavar](https://damir.cavar.me/) and [NLP Lab](https://nlp-lab.org/)
 8
 9Defaults for the NLQK library.
10
11"""
12
13
14from pathlib import Path
15
16
17# Data main path
18DATA_FOLDER_NAME = "nlqk_data"
19DATA_FOLDER_WIN = Path("AppData") / "Roaming"
20DATA_FOLDER_CORPORA = "corpora"
21
22# Data paths and URLs
23SIMLEX_999_ZIP_FILE = "SimLex-999.zip"
24SIMLEX_999_FOLDER = "SimLex-999"
25SIMLEX_999_FILE = "SimLex-999.txt"
26SIMLEX_999_URL = "https://fh295.github.io/SimLex-999.zip"
27
28
29# OpenAI 3rd generation embedding models (name : # dimensions)
30OPEN_AI_EMBEDDING_MODELS = {
31    'text-embedding-3-small': 1536,
32    'text-embedding-3-large': 3072,
33    'text-embedding-ada-002': 1535
34}
DATA_FOLDER_NAME = 'nlqk_data'
DATA_FOLDER_WIN = PosixPath('AppData/Roaming')
DATA_FOLDER_CORPORA = 'corpora'
SIMLEX_999_ZIP_FILE = 'SimLex-999.zip'
SIMLEX_999_FOLDER = 'SimLex-999'
SIMLEX_999_FILE = 'SimLex-999.txt'
SIMLEX_999_URL = 'https://fh295.github.io/SimLex-999.zip'
OPEN_AI_EMBEDDING_MODELS = {'text-embedding-3-small': 1536, 'text-embedding-3-large': 3072, 'text-embedding-ada-002': 1535}