// Models
import { AllModels, ModelEnum, NonModelEnum, ModelInfoMap } from 'llm-info';

console.log(AllModels);
/*
[
  'gpt-4',
  'gpt-4-turbo',
  'gpt-4o',
  'gpt-4o-64k-output-alpha',
  'gpt-4o-mini',
  'gpt-4o-2024-08-06',
  'o1-preview',
  'o1-mini',
  'o1',
  'claude-3-5-sonnet-20240620',
  'claude-3-5-sonnet-20241022',
  'claude-3-5-haiku-20241022',
  'o3-mini',
  'claude-3-7-sonnet-20250219',
  'deepseek-chat',
  'deepseek-reasoner',
  'gemini-2.5-pro-exp-03-25'
]
*/

// Model Info
const modelInfo = ModelInfoMap['gpt-4o'];
console.log(modelInfo);
/*
{
  name: 'GPT-4o',
  provider: 'openai',
  contextWindowTokenLimit: 128000,
  outputTokenLimit: 4096,
  pricePerMillionInputTokens: 5,
  pricePerMillionOutputTokens: 15,
  tokenizerId: 'Xenova/gpt-4o'
}
*/

// Providers
import { AI_PROVIDER_NAME_MAP } from 'llm-info';

console.log(AI_PROVIDER_NAME_MAP);
/*
{
  openai: 'OpenAI',
  anthropic: 'Anthropic',
  'azure-openai': 'Azure OpenAI',
  deepseek: 'DeepSeek',
  openrouter: 'OpenRouter',
  google: 'Google'
}
*/

// Tokenizer
import { AutoTokenizer } from '@xenova/transformers';
const testSentence =
  "Many words map to one token, but some don't: indivisible.";
const results: string[] = [];
for (let i = 0; i < AllModels.length; i++) {
  const model = AllModels[i];
  if (ModelInfoMap[model].tokenizerId) {
    const tokenizer = await AutoTokenizer.from_pretrained(
      ModelInfoMap[model].tokenizerId
    );
    const tokens = tokenizer.encode(testSentence);
    results.push(`${model}: ${tokens.length}`);
  }
}
console.log(`Test sentence: ${testSentence}\n${results.join('\n')}`);
// Test sentence: Many words map to one token, but some don't: indivisible.
// gpt-4: 15
// gpt-4o: 14
// gpt-4o-mini: 14
// claude-3-5-sonnet-20240620: 16