@babelize/engine 0.1.0 → 0.1.3

package/dist/modules/locale-builder/po.d.ts

@@ -1,33 +0,0 @@
-import type { TranslationTable } from 'gettext-parser';
-/**
- * Read a .po file content and return key-value pairs (msgid → msgstr).
- * For untranslated entries, msgstr will be an empty string.
- */
-export declare function readPoFile(content: string): Record<string, string>;
-/**
- * Read a .po file and return the full gettext-parser translation table
- * (preserves comments, msgctxt, plural forms, etc.)
- */
-export declare function readPoFileRaw(content: string): TranslationTable;
-/**
- * Write a .po file from key-value pairs.
- * Uses a gettext-parser translation table as the base for preserving
- * comments, msgctxt, and header metadata.
- */
-export declare function writePoFile(translations: Record<string, string>, options?: {
-    context?: string;
-    pluralForms?: string;
-    baseContent?: string;
-}): string;
-/**
- * Compile a .po file to .mo binary format.
- */
-export declare function compileMo(poContent: string): Buffer;
-/**
- * Build a gettext-parser table from flat translation entries.
- * Useful when generating a fresh .po file from extracted keys.
- */
-export declare function buildPoTable(translations: Record<string, string>, options?: {
-    sourceLanguage?: string;
-    pluralForms?: string;
-}): TranslationTable;

package/dist/modules/locale-builder/po.js

@@ -1,131 +0,0 @@
-"use strict";
-var __importDefault = (this && this.__importDefault) || function (mod) {
-    return (mod && mod.__esModule) ? mod : { "default": mod };
-};
-Object.defineProperty(exports, "__esModule", { value: true });
-exports.readPoFile = readPoFile;
-exports.readPoFileRaw = readPoFileRaw;
-exports.writePoFile = writePoFile;
-exports.compileMo = compileMo;
-exports.buildPoTable = buildPoTable;
-const gettext_parser_1 = __importDefault(require("gettext-parser"));
-/**
- * Read a .po file content and return key-value pairs (msgid → msgstr).
- * For untranslated entries, msgstr will be an empty string.
- */
-function readPoFile(content) {
-    const parsed = gettext_parser_1.default.po.parse(content);
-    const result = {};
-    for (const context of Object.values(parsed.translations)) {
-        for (const [msgid, entry] of Object.entries(context)) {
-            if (!msgid)
-                continue; // skip header entry
-            const msgstr = entry.msgstr?.[0] || '';
-            result[msgid] = msgstr;
-        }
-    }
-    return result;
-}
-/**
- * Read a .po file and return the full gettext-parser translation table
- * (preserves comments, msgctxt, plural forms, etc.)
- */
-function readPoFileRaw(content) {
-    return gettext_parser_1.default.po.parse(content);
-}
-/**
- * Write a .po file from key-value pairs.
- * Uses a gettext-parser translation table as the base for preserving
- * comments, msgctxt, and header metadata.
- */
-function writePoFile(translations, options) {
-    const ctx = options?.context || '';
-    // Build the gettext-parser table structure
-    const table = {
-        charset: 'utf-8',
-        headers: {
-            'content-type': 'text/plain; charset=utf-8',
-            'plural-forms': options?.pluralForms || 'nplurals=2; plural=(n!=1);',
-        },
-        translations: {
-            [ctx]: {},
-        },
-    };
-    // If base content is provided, merge existing metadata
-    if (options?.baseContent) {
-        const base = gettext_parser_1.default.po.parse(options.baseContent);
-        table.headers = { ...base.headers, ...table.headers };
-        if (base.charset)
-            table.charset = base.charset;
-        // Preserve existing entries for context
-        if (base.translations[ctx]) {
-            for (const [msgid, entry] of Object.entries(base.translations[ctx])) {
-                if (!msgid)
-                    continue;
-                table.translations[ctx][msgid] = entry;
-            }
-        }
-    }
-    // Apply translations
-    for (const [msgid, msgstr] of Object.entries(translations)) {
-        if (!msgid)
-            continue;
-        if (table.translations[ctx][msgid]) {
-            table.translations[ctx][msgid].msgstr = [msgstr];
-        }
-        else {
-            table.translations[ctx][msgid] = {
-                msgid,
-                msgstr: [msgstr],
-            };
-        }
-    }
-    const compiled = gettext_parser_1.default.po.compile(table, {
-        foldLength: 0,
-        sort: false,
-        escapeCharacters: true,
-    });
-    return compiled.toString('utf-8');
-}
-/**
- * Compile a .po file to .mo binary format.
- */
-function compileMo(poContent) {
-    const parsed = gettext_parser_1.default.po.parse(poContent);
-    return gettext_parser_1.default.mo.compile(parsed);
-}
-/**
- * Build a gettext-parser table from flat translation entries.
- * Useful when generating a fresh .po file from extracted keys.
- */
-function buildPoTable(translations, options) {
-    const ctx = '';
-    const headers = {
-        'content-type': 'text/plain; charset=utf-8',
-        'plural-forms': options?.pluralForms || 'nplurals=2; plural=(n!=1);',
-    };
-    if (options?.sourceLanguage) {
-        headers['language'] = options.sourceLanguage;
-    }
-    const table = {
-        charset: 'utf-8',
-        headers,
-        translations: {
-            [ctx]: {},
-        },
-    };
-    // Add header entry
-    table.translations[ctx][''] = {
-        msgid: '',
-        msgstr: [Object.entries(headers).map(([k, v]) => `${k}: ${v}`).join('\n')],
-    };
-    for (const [msgid, msgstr] of Object.entries(translations)) {
-        if (!msgid)
-            continue;
-        table.translations[ctx][msgid] = {
-            msgid,
-            msgstr: [msgstr],
-        };
-    }
-    return table;
-}

package/dist/modules/translator/translator.d.ts

@@ -1,5 +0,0 @@
-import type { BabelizeConfig, TranslationResult, TranslationUnit, TranslatorClient } from '../../types';
-export interface TranslatorOptions {
-    client: TranslatorClient;
-}
-export declare function translateUnits(config: BabelizeConfig, units: TranslationUnit[], opts: TranslatorOptions): Promise<TranslationResult[]>;

package/dist/modules/translator/translator.js

@@ -1,46 +0,0 @@
-"use strict";
-Object.defineProperty(exports, "__esModule", { value: true });
-exports.translateUnits = translateUnits;
-const glossary_1 = require("../glossary");
-const language_mapper_1 = require("./language-codes/language-mapper");
-const batch_optimizer_1 = require("./batch-optimizer");
-async function translateUnits(config, units, opts) {
-    if (!units.length)
-        return [];
-    const sourceFlores = (0, language_mapper_1.toFlores)(config.sourceLocale);
-    const results = [];
-    const glossaryTerms = config.glossary?.neverTranslate ?? [];
-    const batches = (0, batch_optimizer_1.createBatches)(units);
-    // Sequential batches — NEVER parallel on HF free tier.
-    for (const batch of batches) {
-        const payload = batch.units.map((u) => {
-            const { masked, terms } = (0, glossary_1.applyGlossaryMask)(u.sourceText, glossaryTerms);
-            return { unit: u, masked, terms };
-        });
-        for (const { unit, masked, terms } of payload) {
-            const targetFlores = (0, language_mapper_1.toFlores)(unit.targetLocale);
-            const res = await opts.client.translate({
-                entries: [{ key: 'main', value: masked }],
-                sourceLang: sourceFlores,
-                targetLang: targetFlores,
-            });
-            if (!Array.isArray(res.translations) || res.translations.length === 0) {
-                throw new Error('HF translator returned no translations');
-            }
-            const translatedText = res.translations[0].value;
-            if (typeof translatedText !== 'string') {
-                throw new Error('HF translator returned an invalid non-string translation value');
-            }
-            const restored = (0, glossary_1.restoreGlossaryMask)(translatedText, terms);
-            results.push({
-                sourceText: unit.sourceText,
-                sourceLocale: unit.sourceLocale,
-                targetLocale: unit.targetLocale,
-                translatedText: restored,
-                provider: 'hf',
-                model: res.model,
-            });
-        }
-    }
-    return results;
-}

package/dist/utils/dtl-mask.d.ts

@@ -1,25 +0,0 @@
-/**
- * Mask Django Template Language tags with private-use sentinels
- * so node-html-parser treats them as opaque text.
- *
- * Handles: {% %}, {{ }}, {# #}
- *
- * After masking, the HTML can be safely parsed by node-html-parser
- * and run through the standard HTML injection pipeline.
- */
-/**
- * Mask DTL template tags in content.
- * Returns the masked string and a restore function to undo masking.
- */
-export declare function maskTemplateTags(content: string): {
-    masked: string;
-    restore: (masked: string) => string;
-};
-/**
- * Check if content contains any Django template tags.
- */
-export declare function hasTemplateTags(content: string): boolean;
-/**
- * Extract all DTL template tags from content for inspection.
- */
-export declare function extractTemplateTags(content: string): string[];

package/dist/utils/dtl-mask.js

@@ -1,58 +0,0 @@
-"use strict";
-/**
- * Mask Django Template Language tags with private-use sentinels
- * so node-html-parser treats them as opaque text.
- *
- * Handles: {% %}, {{ }}, {# #}
- *
- * After masking, the HTML can be safely parsed by node-html-parser
- * and run through the standard HTML injection pipeline.
- */
-Object.defineProperty(exports, "__esModule", { value: true });
-exports.maskTemplateTags = maskTemplateTags;
-exports.hasTemplateTags = hasTemplateTags;
-exports.extractTemplateTags = extractTemplateTags;
-// Unicode Private Use Area block tag characters used as sentinels
-const BLOCK_START = '\uE000';
-const BLOCK_END = '\uE001';
-/**
- * Mask DTL template tags in content.
- * Returns the masked string and a restore function to undo masking.
- */
-function maskTemplateTags(content) {
-    const replacements = [];
-    let counter = 0;
-    // Unified pattern for {% %}, {{ }}, {# #}
-    // Match opening braces with space-or-tag or closing braces with space-or-tag
-    const tagPattern = /(\{%.*?%\}|\{\{.*?\}\}|\{#.*?#\})/gs;
-    const masked = content.replace(tagPattern, (match) => {
-        const sentinel = `${BLOCK_START}${counter}${BLOCK_END}`;
-        replacements.push({ sentinel, original: match });
-        counter++;
-        return sentinel;
-    });
-    return {
-        masked,
-        restore: (maskedContent) => {
-            let restored = maskedContent;
-            for (const { sentinel, original } of replacements) {
-                // Escape sentinel for safe use in regex replacement
-                restored = restored.split(sentinel).join(original);
-            }
-            return restored;
-        },
-    };
-}
-/**
- * Check if content contains any Django template tags.
- */
-function hasTemplateTags(content) {
-    return /\{%.*?%\}|\{\{.*?\}\}|\{#.*?#\}/s.test(content);
-}
-/**
- * Extract all DTL template tags from content for inspection.
- */
-function extractTemplateTags(content) {
-    const tagPattern = /\{%.*?%\}|\{\{.*?\}\}|\{#.*?#\}/gs;
-    return content.match(tagPattern) || [];
-}

package/dist/utils/py-extractor.d.ts

@@ -1,14 +0,0 @@
-import type { ExtractedString, Warning } from '../runtime/types';
-/**
- * Extract translatable strings from Python source code.
- *
- * Matches: _(), gettext(), gettext_lazy(), ugettext*(), ngettext()
- * Also opt-in: verbose_name=, help_text=, label= kwargs
- *
- * Uses a character-position lexer approach (not a full AST parser)
- * to handle single/double/triple quotes, raw strings, f-strings, and escape sequences.
- */
-export declare function extractStringsFromPython(content: string, filePath: string): {
-    strings: ExtractedString[];
-    warnings: Warning[];
-};

package/dist/utils/py-extractor.js

@@ -1,285 +0,0 @@
-"use strict";
-Object.defineProperty(exports, "__esModule", { value: true });
-exports.extractStringsFromPython = extractStringsFromPython;
-const string_utils_1 = require("./string-utils");
-// Function calls that indicate translatable strings
-const GETTEXT_CALLS = [
-    '_',
-    'gettext',
-    'gettext_lazy',
-    'ngettext',
-    'ngettext_lazy',
-    'ugettext',
-    'ugettext_lazy',
-    'pgettext',
-    'pgettext_lazy',
-    'npgettext',
-    'npgettext_lazy',
-];
-// Keyword argument names that Django/Python treat as translatable
-const TRANSlatable_KWARGS = [
-    'verbose_name',
-    'verbose_name_plural',
-    'help_text',
-    'label',
-];
-/**
- * Extract translatable strings from Python source code.
- *
- * Matches: _(), gettext(), gettext_lazy(), ugettext*(), ngettext()
- * Also opt-in: verbose_name=, help_text=, label= kwargs
- *
- * Uses a character-position lexer approach (not a full AST parser)
- * to handle single/double/triple quotes, raw strings, f-strings, and escape sequences.
- */
-function extractStringsFromPython(content, filePath) {
-    const ctx = {
-        filePath,
-        content,
-        strings: [],
-        seenTexts: new Set(),
-        warnings: [],
-    };
-    let i = 0;
-    const len = content.length;
-    while (i < len) {
-        // Skip lines that are comments
-        if (content[i] === '#') {
-            i = skipToNextLine(content, i);
-            continue;
-        }
-        // Skip string literals on their own (not preceded by a function call or kwarg)
-        // We only extract strings that are arguments to gettext functions or Django kwargs
-        // Check for gettext function calls: _( "..." ), gettext( "..." ), etc.
-        for (const callName of GETTEXT_CALLS) {
-            const callPattern = new RegExp(`\\b${callName.replace(/_/g, '\\_')}\\s*\\(`);
-            const match = callPattern.exec(content.slice(i));
-            if (match && match.index === 0) {
-                const parenStart = i + match[0].length;
-                const args = collectCallArgs(content, parenStart);
-                if (args) {
-                    for (const arg of args) {
-                        const str = extractStringFromArg(content, arg);
-                        if (str !== null) {
-                            addString(ctx, str.text, arg.startLine, 0, str.context);
-                        }
-                    }
-                }
-                i = args ? (args[args.length - 1]?.end || parenStart) : parenStart;
-                break;
-            }
-        }
-        // Check for Django keyword arguments: verbose_name="...", help_text="..."
-        for (const kwarg of TRANSlatable_KWARGS) {
-            const kwargPattern = new RegExp(`\\b${kwarg}\\s*=\\s*`);
-            const match = kwargPattern.exec(content.slice(i));
-            if (match && match.index === 0) {
-                const valueStart = i + match[0].length;
-                const str = extractStringLiteral(content, valueStart);
-                if (str !== null) {
-                    addString(ctx, str.text, findLineNumber(content, valueStart), 0, str.context);
-                }
-                i = valueStart + (str ? str.rawLength : 0);
-                break;
-            }
-        }
-        i++;
-    }
-    return { strings: ctx.strings, warnings: ctx.warnings };
-}
-/**
- * Collect argument ranges for a function call starting after the opening paren.
- * Returns an array of {start, end, startLine} for each argument position.
- */
-function collectCallArgs(content, start) {
-    let depth = 1;
-    let i = start;
-    const args = [];
-    let argStart = -1;
-    let parenDepth = 0;
-    while (i < content.length && depth > 0) {
-        const ch = content[i];
-        if (ch === '(') {
-            parenDepth++;
-        }
-        else if (ch === ')') {
-            parenDepth--;
-            if (parenDepth === 0 && depth === 1) {
-                if (argStart !== -1) {
-                    args.push({
-                        start: argStart,
-                        end: i,
-                        startLine: findLineNumber(content, argStart),
-                    });
-                }
-                depth--;
-                i++;
-                continue;
-            }
-        }
-        else if (ch === ',' && parenDepth === 1) {
-            if (argStart !== -1) {
-                args.push({
-                    start: argStart,
-                    end: i,
-                    startLine: findLineNumber(content, argStart),
-                });
-            }
-            argStart = -1;
-        }
-        else if (ch === '"' || ch === "'" || ch === 'f') {
-            // String literal — skip over it
-            const strEnd = skipString(content, i);
-            if (strEnd !== i) {
-                if (argStart === -1)
-                    argStart = i;
-                i = strEnd;
-                continue;
-            }
-        }
-        // Track first non-whitespace as arg start
-        if (argStart === -1 && !/\s/.test(ch) && ch !== ',') {
-            // Check if this is a keyword arg or string
-            if (ch !== ')') {
-                argStart = i;
-            }
-        }
-        i++;
-    }
-    return args.length > 0 ? args : null;
-}
-/**
- * Extract the string value from an argument position.
- * Handles named arguments (_("text"), _("text", count)).
- * Only extracts the first string argument (singular form).
- */
-function extractStringFromArg(content, arg) {
-    // Skip keyword name if present (e.g., msgid="text")
-    let i = arg.start;
-    const end = arg.end;
-    while (i < end && content[i] !== '"' && content[i] !== "'" && content[i] !== 'f') {
-        i++;
-    }
-    if (i >= end)
-        return null;
-    const str = extractStringLiteral(content, i);
-    if (!str)
-        return null;
-    return {
-        text: str.text,
-        rawLength: str.rawLength,
-        context: str.context,
-    };
-}
-/**
- * Extract a string literal from content starting at the given position.
- * Handles: "...", '...', """...""", '''...''', f"...", r"...", b"..."
- */
-function extractStringLiteral(content, start) {
-    let i = start;
-    const len = content.length;
-    // Skip prefix characters: r, f, b, rb, fr, etc.
-    while (i < len && /[a-zA-Z]/.test(content[i]) && content[i] !== '"' && content[i] !== "'") {
-        i++;
-    }
-    if (i >= len)
-        return null;
-    const quote = content[i];
-    if (quote !== '"' && quote !== "'")
-        return null;
-    // Check for triple-quote
-    const isTriple = content.slice(i, i + 3) === quote.repeat(3);
-    const openQuote = isTriple ? quote.repeat(3) : quote;
-    const closeQuote = openQuote;
-    const valueStart = i + openQuote.length;
-    let j = valueStart;
-    let text = '';
-    let context = 'string-literal';
-    // Detect f-string
-    const prefixStart = start;
-    const prefix = content.slice(prefixStart, i).toLowerCase();
-    if (prefix.includes('f')) {
-        context = 'template-literal';
-    }
-    while (j < len) {
-        if (content[j] === '\\') {
-            // Escape sequence — skip next char
-            j += 2;
-            continue;
-        }
-        if (content.slice(j, j + closeQuote.length) === closeQuote) {
-            // Found closing quote
-            const rawLength = j + closeQuote.length - start;
-            return { text, rawLength, context };
-        }
-        text += content[j];
-        j++;
-    }
-    return null;
-}
-/**
- * Skip over a string literal starting at the given position.
- * Returns the index after the closing quote.
- */
-function skipString(content, start) {
-    let i = start;
-    const len = content.length;
-    // Skip prefix characters
-    while (i < len && /[a-zA-Z]/.test(content[i]) && content[i] !== '"' && content[i] !== "'") {
-        i++;
-    }
-    if (i >= len)
-        return start;
-    const quote = content[i];
-    if (quote !== '"' && quote !== "'")
-        return start;
-    const isTriple = content.slice(i, i + 3) === quote.repeat(3);
-    const closeQuote = isTriple ? quote.repeat(3) : quote;
-    i += closeQuote.length;
-    while (i < len) {
-        if (content[i] === '\\') {
-            i += 2;
-            continue;
-        }
-        if (content.slice(i, i + closeQuote.length) === closeQuote) {
-            return i + closeQuote.length;
-        }
-        i++;
-    }
-    return start; // Unterminated string
-}
-/**
- * Add a string to the extraction context, deduplicating by content hash.
- */
-function addString(ctx, text, line, column, context) {
-    const trimmed = text.trim();
-    if (!trimmed || trimmed.length < 2)
-        return;
-    if (!(0, string_utils_1.isTranslatableText)(trimmed))
-        return;
-    const contentKey = `${ctx.filePath}:${(0, string_utils_1.hashString)(trimmed)}`;
-    if (ctx.seenTexts.has(contentKey))
-        return;
-    const id = `${ctx.filePath}:${line}:${(0, string_utils_1.hashString)(trimmed)}`;
-    ctx.seenTexts.add(contentKey);
-    ctx.strings.push({
-        text: trimmed,
-        file: ctx.filePath,
-        line,
-        column,
-        context,
-        id,
-    });
-}
-function skipToNextLine(content, pos) {
-    const newline = content.indexOf('\n', pos);
-    return newline === -1 ? content.length : newline + 1;
-}
-function findLineNumber(content, pos) {
-    let line = 1;
-    for (let i = 0; i < pos && i < content.length; i++) {
-        if (content[i] === '\n')
-            line++;
-    }
-    return line;
-}