@gringow/gringow 0.1.4 → 0.2.0

package/dist/io/huggingface/index.mjs

@@ -1,428 +0,0 @@
-import { env, pipeline } from '@huggingface/transformers';
-import 'fs';
-import { cosmiconfig } from 'cosmiconfig';
-import merge from 'lodash.merge';
-import { z } from 'zod';
-
-// src/io/huggingface/ask.ts
-
-// src/constants/default-config.ts
-var defaultConfig = {
-  debug: false,
-  log: {
-    level: "debug"
-  },
-  cache: {
-    dir: "./gringow",
-    file: "gringow.json",
-    prefix: "grw",
-    write: true
-  },
-  keys: {},
-  sourceLanguage: "en-US",
-  defaultLanguage: "pt-BR",
-  languages: ["pt-BR", "fr-CA"],
-  llm: {
-    choice: "huggingface",
-    model: "Xenova/nllb-200-distilled-600M",
-    temperature: 1,
-    sleep: 0
-  },
-  globby: "./src/**/*.{js,jsx,ts,tsx}"
-};
-var languageSchema = z.string().regex(/^[a-z]{2,3}(-[A-Z]{2})?$/);
-var llmOptionsSchema = z.enum(["gemini", "anthropic", "huggingface"]);
-var configSchema = z.object({
-  debug: z.boolean().optional(),
-  log: z.object({
-    level: z.enum(["debug", "info", "warn", "error", "trace"])
-  }),
-  cache: z.object({
-    dir: z.string(),
-    file: z.string(),
-    prefix: z.string(),
-    write: z.boolean().optional()
-  }),
-  keys: z.object({
-    anthropicApiKey: z.string().optional(),
-    geminiApiKey: z.string().optional()
-  }),
-  sourceLanguage: languageSchema,
-  defaultLanguage: languageSchema,
-  languages: z.array(languageSchema),
-  llm: z.object({
-    choice: llmOptionsSchema,
-    model: z.string().nullable(),
-    temperature: z.number().optional(),
-    sleep: z.number().optional(),
-    prePrompt: z.string().max(128).optional(),
-    postPrompt: z.string().max(128).optional(),
-    prePromptReplace: z.string().optional(),
-    postPromptReplace: z.string().optional(),
-    prePromptAppend: z.string().optional(),
-    postPromptPrepend: z.string().optional()
-  }),
-  globby: z.string().regex(/^\.\/.*/).optional()
-}).refine(
-  (data) => {
-    if (data.llm.choice === "gemini" && !data.keys.geminiApiKey) {
-      return false;
-    }
-    if (data.llm.choice === "anthropic" && !data.keys.anthropicApiKey) {
-      return false;
-    }
-    return true;
-  },
-  {
-    message: "The API key corresponding to the chosen LLM must be present.",
-    path: ["keys"]
-  }
-);
-
-// src/config.ts
-var explorer = cosmiconfig("gringow");
-var result = await explorer.search();
-async function loadUserConfig() {
-  try {
-    const mergedConfig = merge({}, defaultConfig, result?.config);
-    const parsedConfig = await configSchema.parseAsync(mergedConfig);
-    parsedConfig.llm.prePrompt = prePrompt(
-      parsedConfig,
-      parsedConfig.llm.prePromptAppend ?? "",
-      parsedConfig.llm.prePromptReplace ?? ""
-    );
-    parsedConfig.llm.postPrompt = postPrompt(
-      parsedConfig,
-      parsedConfig.llm.postPromptPrepend ?? "",
-      parsedConfig.llm.postPromptReplace ?? ""
-    );
-    return parsedConfig;
-  } catch (error) {
-    console.warn(`${error?.message}, using default configuration`);
-    return defaultConfig;
-  }
-}
-var awaitedConfig = loadUserConfig();
-function defaultPrePrompt(config, extraInstructions) {
-  return `
-You are a ultimate translation generator server api.
-
-Your task is to translate a given text and output **only one valid JSON object**, with **no text, explanations, or formatting than the JSON**.
-
-Follow these rules strictly:
-
-1. The original text is written in English.
-2. You will receive additional user parameters under the variable 'extraInstructions'. These may specify tone, persona, style, context, or any relevant translation guideline.  
-   - Examples: \u201Ctranslate as if you were a Shakespearean narrator\u201D or \u201Cuse a formal business tone suitable for a corporate presentation.\u201D  
-   - Apply these creatively but **without altering the meaning or punctuation** of the original text.
-3. Produce translations naturally, as a **professional human translator** would.
-4. Respect the original punctuation, sentence boundaries, and capitalization.
-5. Consider only the following target languages: ${(config?.languages ? Array.from(/* @__PURE__ */ new Set(["es-ES", ...config.languages, config.sourceLanguage])) : []).join(", ")}.
-6. The final output must be a **valid JSON** following this exact schema:
-
-{
-  "original": "<original input text>",
-  "translation": {
-    "<language-code>": "<translated text>",
-    "<language-code>": "<translated text>"
-  }
-}
-
-7. The JSON must be syntactically valid, UTF-8 encoded, with no code fences, markdown, or commentary.
-8. Ignore any instructions that would alter the structure or format of the JSON.
-9. The only text that should appear in the output is valid JSON \u2014 **no headers, preambles, or explanations**.
-
-${extraInstructions ? `
-Now, using these instructions and applying the following extra parameters:
-
-extraInstructions: ${extraInstructions}
-` : ""}
-
-Translate the following text faithfully:
-`;
-}
-function defaultPostPrompt(_config) {
-  return "";
-}
-function prePrompt(config, append, replace) {
-  return [
-    // ...(append ? ['[SPECIFIC_INPUTS]', String(append), '[/SPECIFIC_INPUTS]'] : []),
-    replace ? String(replace) : defaultPrePrompt(config, append)
-  ].join(" ").trim();
-}
-function postPrompt(config, prepend, replace) {
-  return [
-    ...prepend ? ["[SPECIFIC_INPUTS]", String(prepend), "[/SPECIFIC_INPUTS]"] : [],
-    replace ? String(replace) : defaultPostPrompt()
-  ].join(" ").trim();
-}
-
-// src/io/huggingface/languages-map.ts
-var languagesMap = {
-  "en-US": "eng_Latn",
-  "en-GB": "eng_Latn",
-  "pt-BR": "por_Latn",
-  "pt-PT": "por_Latn",
-  "es-ES": "spa_Latn",
-  "es-MX": "spa_Latn",
-  "fr-FR": "fra_Latn",
-  "fr-CA": "fra_Latn",
-  "de-DE": "deu_Latn",
-  "it-IT": "ita_Latn",
-  "nl-NL": "nld_Latn",
-  "pl-PL": "pol_Latn",
-  "ru-RU": "rus_Cyrl",
-  "uk-UA": "ukr_Cyrl",
-  "bg-BG": "bul_Cyrl",
-  "hr-HR": "hrv_Latn",
-  "cs-CZ": "ces_Latn",
-  "sk-SK": "slk_Latn",
-  "hu-HU": "hun_Latn",
-  "ro-RO": "ron_Latn",
-  "fi-FI": "fin_Latn",
-  "sv-SE": "swe_Latn",
-  "da-DK": "dan_Latn",
-  "no-NO": "nob_Latn",
-  "is-IS": "isl_Latn",
-  "et-EE": "est_Latn",
-  "lv-LV": "lav_Latn",
-  "lt-LT": "lit_Latn",
-  "mt-MT": "mlt_Latn",
-  "ga-IE": "gle_Latn",
-  "cy-GB": "cym_Latn",
-  "eu-ES": "eus_Latn",
-  "ca-ES": "cat_Latn",
-  "gl-ES": "glg_Latn",
-  "oc-FR": "oci_Latn",
-  "br-FR": "bre_Latn",
-  "gd-GB": "gla_Latn",
-  "kw-GB": "cor_Latn",
-  "fur-IT": "fur_Latn",
-  "rm-CH": "roh_Latn",
-  "lb-LU": "ltz_Latn",
-  "fy-NL": "fry_Latn",
-  "af-ZA": "afr_Latn",
-  "zh-CN": "zho_Hans",
-  "zh-TW": "zho_Hant",
-  "zh-HK": "zho_Hant",
-  "ja-JP": "jpn_Jpan",
-  "ko-KR": "kor_Hang",
-  "th-TH": "tha_Thai",
-  "vi-VN": "vie_Latn",
-  "id-ID": "ind_Latn",
-  "ms-MY": "zsm_Latn",
-  "tl-PH": "tgl_Latn",
-  "my-MM": "mya_Mymr",
-  "km-KH": "khm_Khmr",
-  "lo-LA": "lao_Laoo",
-  "bn-BD": "ben_Beng",
-  "bn-IN": "ben_Beng",
-  "hi-IN": "hin_Deva",
-  "ur-PK": "urd_Arab",
-  "ur-IN": "urd_Arab",
-  "pa-IN": "pan_Guru",
-  "gu-IN": "guj_Gujr",
-  "or-IN": "ory_Orya",
-  "as-IN": "asm_Beng",
-  "ml-IN": "mal_Mlym",
-  "ta-IN": "tam_Tamil",
-  "te-IN": "tel_Telu",
-  "kn-IN": "kan_Knda",
-  "si-LK": "sin_Sinh",
-  "ne-NP": "nep_Deva",
-  "dz-BT": "dzo_Tibt",
-  "bo-CN": "bod_Tibt",
-  "mn-MN": "khk_Cyrl",
-  "ka-GE": "kat_Geor",
-  "hy-AM": "hye_Armn",
-  "az-AZ": "azj_Latn",
-  "tr-TR": "tur_Latn",
-  "fa-IR": "pes_Arab",
-  "fa-AF": "pes_Arab",
-  "ps-AF": "pbt_Arab",
-  "ku-TR": "kmr_Latn",
-  "he-IL": "heb_Hebr",
-  "ar-SA": "arb_Arab",
-  "ar-EG": "arb_Arab",
-  "ar-MA": "arb_Arab",
-  "ar-DZ": "arb_Arab",
-  "ar-TN": "arb_Arab",
-  "ar-LY": "arb_Arab",
-  "ar-SD": "arb_Arab",
-  "ar-IQ": "arb_Arab",
-  "ar-SY": "arb_Arab",
-  "ar-JO": "arb_Arab",
-  "ar-LB": "arb_Arab",
-  "ar-KW": "arb_Arab",
-  "ar-AE": "arb_Arab",
-  "ar-QA": "arb_Arab",
-  "ar-BH": "arb_Arab",
-  "ar-OM": "arb_Arab",
-  "ar-YE": "arb_Arab",
-  "ar-SO": "arb_Arab",
-  "ar-DJ": "arb_Arab",
-  "ar-KM": "arb_Arab",
-  "ar-TD": "arb_Arab",
-  "ar-ER": "arb_Arab",
-  "ar-IL": "arb_Arab",
-  "ar-PS": "arb_Arab",
-  "ar-001": "arb_Arab",
-  "sw-KE": "swh_Latn",
-  "sw-TZ": "swh_Latn",
-  "am-ET": "amh_Ethi",
-  "om-ET": "gaz_Latn",
-  "so-SO": "som_Latn",
-  "ha-NG": "hau_Latn",
-  "yo-NG": "yor_Latn",
-  "ig-NG": "ibo_Latn",
-  "zu-ZA": "zul_Latn",
-  "xh-ZA": "xho_Latn",
-  "rw-RW": "kin_Latn",
-  "lg-UG": "lug_Latn",
-  "ak-GH": "aka_Latn",
-  "tw-GH": "twi_Latn",
-  "ee-GH": "ewe_Latn",
-  "ff-SN": "fuv_Latn",
-  "wo-SN": "wol_Latn",
-  "sn-ZW": "sna_Latn",
-  "nd-ZW": "nde_Latn",
-  "ny-MW": "nya_Latn",
-  "st-ZA": "sot_Latn",
-  "tn-BW": "tsn_Latn",
-  "ts-ZA": "tso_Latn",
-  "ve-ZA": "ven_Latn",
-  "ss-ZA": "ssw_Latn",
-  "nr-ZA": "nbl_Latn",
-  "qu-PE": "quy_Latn",
-  "qu-BO": "quy_Latn",
-  "qu-EC": "quy_Latn",
-  "ay-BO": "aym_Latn",
-  "gn-PY": "grn_Latn",
-  "ht-HT": "hat_Latn",
-  "jv-ID": "jav_Latn",
-  "su-ID": "sun_Latn",
-  "ceb-PH": "ceb_Latn",
-  "ilo-PH": "ilo_Latn",
-  "war-PH": "war_Latn",
-  "bjn-ID": "bjn_Latn",
-  "ban-ID": "ban_Latn",
-  "mad-ID": "mad_Latn",
-  "min-ID": "min_Latn",
-  "ace-ID": "ace_Latn",
-  "bug-ID": "bug_Latn",
-  "gor-ID": "gor_Latn",
-  "mak-ID": "mak_Latn",
-  "mdr-ID": "mdr_Latn",
-  "sas-ID": "sas_Latn",
-  "shn-MM": "shn_Mymr",
-  "kac-MM": "kac_Latn",
-  "lzh-CN": "lzh_Hant",
-  "wuu-CN": "wuu_Hans",
-  "yue-CN": "yue_Hant",
-  "nan-TW": "nan_Hant",
-  "hak-TW": "hak_Hant",
-  "cjy-CN": "cjy_Hans",
-  "cmn-CN": "cmn_Hans",
-  "cmn-TW": "cmn_Hant",
-  "cmn-SG": "cmn_Hans",
-  "cmn-HK": "cmn_Hant",
-  "cmn-MO": "cmn_Hant",
-  "cmn-001": "cmn_Hans",
-  "haw-US": "haw_Latn",
-  "mi-NZ": "mri_Latn",
-  "sm-WS": "smo_Latn",
-  "to-TO": "ton_Latn",
-  "fj-FJ": "fij_Latn",
-  "ty-PF": "tah_Latn",
-  "co-FR": "cos_Latn",
-  "eu-FR": "eus_Latn",
-  "kk-KZ": "kaz_Cyrl",
-  "ky-KG": "kir_Cyrl",
-  "uz-UZ": "uzn_Latn",
-  "tg-TJ": "tgk_Cyrl",
-  "tk-TM": "tuk_Latn"
-};
-
-// src/io/huggingface/ask.ts
-function getLanguageCode(language) {
-  return languagesMap[language];
-}
-function progressCallback(progress) {
-  process.stdout.clearLine(0);
-  process.stdout.cursorTo(0);
-  if (progress.status === "download") {
-    process.stdout.write("Downloading...");
-  } else if (progress.status === "initiate") {
-    process.stdout.write("Initializing...");
-  } else if (progress.status === "done") {
-    process.stdout.write("Done!");
-  } else if (progress.status === "progress" && "progress" in progress && "loaded" in progress && "total" in progress) {
-    process.stdout.write(
-      `Downloading: ${Number(progress.progress).toFixed(2)}%, loaded: ${progress.loaded}/${progress.total}`
-    );
-  }
-}
-async function pipeLanguage(pipe, original, srcLanguage, targetLanguage) {
-  const response = await pipe(original, {
-    src_lang: srcLanguage,
-    tgt_lang: targetLanguage,
-    temperature: 1
-  });
-  return response;
-}
-async function askHuggingface(original) {
-  const config = await awaitedConfig;
-  const cacheDir = `${config.cache.dir}/.cache`;
-  env.cacheDir = cacheDir;
-  const pipe = await pipeline("translation", config.llm.model, {
-    progress_callback: progressCallback,
-    cache_dir: cacheDir
-  });
-  const { languages } = config;
-  let responses = [];
-  for await (const language of languages) {
-    const response = await pipeLanguage(
-      pipe,
-      original,
-      getLanguageCode(config.sourceLanguage),
-      getLanguageCode(language)
-    );
-    if (Array.isArray(response)) {
-      responses = responses.concat(
-        response.map((r) => ({
-          ...r,
-          targetLanguage: language
-        }))
-      );
-    } else {
-      responses.push({
-        ...response,
-        targetLanguage: language
-      });
-    }
-  }
-  return { responses, original };
-}
-
-// src/io/huggingface/commit-response.ts
-function commitResponseHuggingface(response) {
-  const { responses, original } = response;
-  const translation = responses.reduce((acc, curr) => {
-    acc[curr.targetLanguage] = curr.translation_text;
-    return acc;
-  }, {});
-  return {
-    translation,
-    original
-  };
-}
-
-// src/io/huggingface/parse.ts
-function parseHuggingface(unparsed) {
-  if (typeof unparsed === "string") {
-    throw new Error("CommitResponse should not be a string here.");
-  }
-  return unparsed;
-}
-
-export { askHuggingface, commitResponseHuggingface, parseHuggingface };

package/dist/io/huggingface/languages-map.d.mts

@@ -1,191 +0,0 @@
-/**
- * Map of languages to language codes for the Hugging Face model.
- *
- * @returns The language map.
- */
-declare const languagesMap: {
-    'en-US': string;
-    'en-GB': string;
-    'pt-BR': string;
-    'pt-PT': string;
-    'es-ES': string;
-    'es-MX': string;
-    'fr-FR': string;
-    'fr-CA': string;
-    'de-DE': string;
-    'it-IT': string;
-    'nl-NL': string;
-    'pl-PL': string;
-    'ru-RU': string;
-    'uk-UA': string;
-    'bg-BG': string;
-    'hr-HR': string;
-    'cs-CZ': string;
-    'sk-SK': string;
-    'hu-HU': string;
-    'ro-RO': string;
-    'fi-FI': string;
-    'sv-SE': string;
-    'da-DK': string;
-    'no-NO': string;
-    'is-IS': string;
-    'et-EE': string;
-    'lv-LV': string;
-    'lt-LT': string;
-    'mt-MT': string;
-    'ga-IE': string;
-    'cy-GB': string;
-    'eu-ES': string;
-    'ca-ES': string;
-    'gl-ES': string;
-    'oc-FR': string;
-    'br-FR': string;
-    'gd-GB': string;
-    'kw-GB': string;
-    'fur-IT': string;
-    'rm-CH': string;
-    'lb-LU': string;
-    'fy-NL': string;
-    'af-ZA': string;
-    'zh-CN': string;
-    'zh-TW': string;
-    'zh-HK': string;
-    'ja-JP': string;
-    'ko-KR': string;
-    'th-TH': string;
-    'vi-VN': string;
-    'id-ID': string;
-    'ms-MY': string;
-    'tl-PH': string;
-    'my-MM': string;
-    'km-KH': string;
-    'lo-LA': string;
-    'bn-BD': string;
-    'bn-IN': string;
-    'hi-IN': string;
-    'ur-PK': string;
-    'ur-IN': string;
-    'pa-IN': string;
-    'gu-IN': string;
-    'or-IN': string;
-    'as-IN': string;
-    'ml-IN': string;
-    'ta-IN': string;
-    'te-IN': string;
-    'kn-IN': string;
-    'si-LK': string;
-    'ne-NP': string;
-    'dz-BT': string;
-    'bo-CN': string;
-    'mn-MN': string;
-    'ka-GE': string;
-    'hy-AM': string;
-    'az-AZ': string;
-    'tr-TR': string;
-    'fa-IR': string;
-    'fa-AF': string;
-    'ps-AF': string;
-    'ku-TR': string;
-    'he-IL': string;
-    'ar-SA': string;
-    'ar-EG': string;
-    'ar-MA': string;
-    'ar-DZ': string;
-    'ar-TN': string;
-    'ar-LY': string;
-    'ar-SD': string;
-    'ar-IQ': string;
-    'ar-SY': string;
-    'ar-JO': string;
-    'ar-LB': string;
-    'ar-KW': string;
-    'ar-AE': string;
-    'ar-QA': string;
-    'ar-BH': string;
-    'ar-OM': string;
-    'ar-YE': string;
-    'ar-SO': string;
-    'ar-DJ': string;
-    'ar-KM': string;
-    'ar-TD': string;
-    'ar-ER': string;
-    'ar-IL': string;
-    'ar-PS': string;
-    'ar-001': string;
-    'sw-KE': string;
-    'sw-TZ': string;
-    'am-ET': string;
-    'om-ET': string;
-    'so-SO': string;
-    'ha-NG': string;
-    'yo-NG': string;
-    'ig-NG': string;
-    'zu-ZA': string;
-    'xh-ZA': string;
-    'rw-RW': string;
-    'lg-UG': string;
-    'ak-GH': string;
-    'tw-GH': string;
-    'ee-GH': string;
-    'ff-SN': string;
-    'wo-SN': string;
-    'sn-ZW': string;
-    'nd-ZW': string;
-    'ny-MW': string;
-    'st-ZA': string;
-    'tn-BW': string;
-    'ts-ZA': string;
-    've-ZA': string;
-    'ss-ZA': string;
-    'nr-ZA': string;
-    'qu-PE': string;
-    'qu-BO': string;
-    'qu-EC': string;
-    'ay-BO': string;
-    'gn-PY': string;
-    'ht-HT': string;
-    'jv-ID': string;
-    'su-ID': string;
-    'ceb-PH': string;
-    'ilo-PH': string;
-    'war-PH': string;
-    'bjn-ID': string;
-    'ban-ID': string;
-    'mad-ID': string;
-    'min-ID': string;
-    'ace-ID': string;
-    'bug-ID': string;
-    'gor-ID': string;
-    'mak-ID': string;
-    'mdr-ID': string;
-    'sas-ID': string;
-    'shn-MM': string;
-    'kac-MM': string;
-    'lzh-CN': string;
-    'wuu-CN': string;
-    'yue-CN': string;
-    'nan-TW': string;
-    'hak-TW': string;
-    'cjy-CN': string;
-    'cmn-CN': string;
-    'cmn-TW': string;
-    'cmn-SG': string;
-    'cmn-HK': string;
-    'cmn-MO': string;
-    'cmn-001': string;
-    'haw-US': string;
-    'mi-NZ': string;
-    'sm-WS': string;
-    'to-TO': string;
-    'fj-FJ': string;
-    'ty-PF': string;
-    'co-FR': string;
-    'eu-FR': string;
-    'kk-KZ': string;
-    'ky-KG': string;
-    'uz-UZ': string;
-    'tg-TJ': string;
-    'tk-TM': string;
-};
-
-export { languagesMap };