@promptbook/legacy-documents 0.71.0-16 → 0.71.0-18

package/umd/index.umd.js

@@ -15,7 +15,7 @@
     /**
      * The version of the Promptbook library
      */
-    var PROMPTBOOK_VERSION = '0.71.0-15';
+    var PROMPTBOOK_VERSION = '0.71.0-17';
     // TODO: [main] !!!! List here all the versions and annotate + put into script
 
     /*! *****************************************************************************
@@ -2555,8 +2555,7 @@
      * Multiple LLM Execution Tools is a proxy server that uses multiple execution tools internally and exposes the executor interface externally.
      *
      * Note: Internal utility of `joinLlmExecutionTools` but exposed type
-     * @public exported from `@promptbook/types`
-     * TODO: !!!!!! Export as runtime class not just type
+     * @public exported from `@promptbook/core`
      */
     var MultipleLlmExecutionTools = /** @class */ (function () {
         /**
@@ -3094,7 +3093,7 @@
      * TODO: [®] DRY Register logic
      */
 
-    // TODO: !!!!!! Maybe delete this function
+    // TODO: !!!!!!last - Maybe delete this function
     /**
      * Creates a message with all registered scrapers
      *
@@ -3202,7 +3201,6 @@
      * @private within the repository
      */
     function sourceContentToName(sourceContent) {
-        // TODO: !!!!!! Better name for source than gibberish hash
         var hash = cryptoJs.SHA256(hexEncoder__default["default"].parse(JSON.stringify(sourceContent)))
             //    <- TODO: [🥬] Encapsulate sha256 to some private utility function
             .toString( /* hex */)
@@ -3333,7 +3331,7 @@
                                                     content = _a.sent();
                                                     return [2 /*return*/, new Blob([
                                                             content,
-                                                            // <- TODO: !!!!!! Maybe not working
+                                                            // <- TODO: !!!!!! Test that this is working
                                                         ], { type: mimeType_1 })];
                                             }
                                         });
@@ -3346,7 +3344,7 @@
                                             switch (_c.label) {
                                                 case 0:
                                                     _b = (_a = JSON).parse;
-                                                    return [4 /*yield*/, promises.readFile(filename_1, 'utf-8')];
+                                                    return [4 /*yield*/, tools.fs.readFile(filename_1, 'utf-8')];
                                                 case 1: return [2 /*return*/, _b.apply(_a, [_c.sent()])];
                                             }
                                         });
@@ -3356,7 +3354,7 @@
                                     return __awaiter(this, void 0, void 0, function () {
                                         return __generator(this, function (_a) {
                                             switch (_a.label) {
-                                                case 0: return [4 /*yield*/, promises.readFile(filename_1, 'utf-8')];
+                                                case 0: return [4 /*yield*/, tools.fs.readFile(filename_1, 'utf-8')];
                                                 case 1: return [2 /*return*/, _a.sent()];
                                             }
                                         });
@@ -5962,12 +5960,13 @@
          * Note: `$` is used to indicate that this function is not a pure function - it leaves files on the disk and you are responsible for cleaning them by calling `destroy` method of returned object
          */
         DocumentScraper.prototype.$convert = function (source) {
+            var _a;
             return __awaiter(this, void 0, void 0, function () {
-                var _a, _b, externalProgramsPaths, _c, rootDirname, _d, cacheDirname, _e, isCacheCleaned, _f, isVerbose, extension, cacheFilehandler, command_1;
+                var _b, _c, rootDirname, _d, cacheDirname, _e, isCacheCleaned, _f, isVerbose, extension, cacheFilehandler, command_1;
                 return __generator(this, function (_g) {
                     switch (_g.label) {
                         case 0:
-                            _a = this.options, _b = _a.externalProgramsPaths, externalProgramsPaths = _b === void 0 ? {} : _b, _c = _a.rootDirname, rootDirname = _c === void 0 ? process.cwd() : _c, _d = _a.cacheDirname, cacheDirname = _d === void 0 ? SCRAPE_CACHE_DIRNAME : _d, _e = _a.isCacheCleaned, isCacheCleaned = _e === void 0 ? false : _e, _f = _a.isVerbose, isVerbose = _f === void 0 ? IS_VERBOSE : _f;
+                            _b = this.options, _c = _b.rootDirname, rootDirname = _c === void 0 ? process.cwd() : _c, _d = _b.cacheDirname, cacheDirname = _d === void 0 ? SCRAPE_CACHE_DIRNAME : _d, _e = _b.isCacheCleaned, isCacheCleaned = _e === void 0 ? false : _e, _f = _b.isVerbose, isVerbose = _f === void 0 ? IS_VERBOSE : _f;
                             if (!$isRunningInNode()) {
                                 throw new KnowledgeScrapeError('Scraping .docx files is only supported in Node environment');
                             }
@@ -5975,7 +5974,7 @@
                                 throw new EnvironmentMismatchError('Can not scrape documents without filesystem tools');
                                 //          <- TODO: [🧠] What is the best error type here`
                             }
-                            if (externalProgramsPaths.pandocPath === undefined) {
+                            if (((_a = this.tools.executables) === null || _a === void 0 ? void 0 : _a.pandocPath) === undefined) {
                                 throw new MissingToolsError('Pandoc is required for scraping .docx files');
                             }
                             if (source.filename === null) {
@@ -5995,7 +5994,7 @@
                             return [4 /*yield*/, isFileExisting(cacheFilehandler.filename, this.tools.fs)];
                         case 2:
                             if (!!(_g.sent())) return [3 /*break*/, 5];
-                            command_1 = "\"".concat(externalProgramsPaths.pandocPath, "\" -f ").concat(extension, " -t markdown \"").concat(source.filename, "\" -o \"").concat(cacheFilehandler.filename, "\"");
+                            command_1 = "\"".concat(this.tools.executables.pandocPath, "\" -f ").concat(extension, " -t markdown \"").concat(source.filename, "\" -o \"").concat(cacheFilehandler.filename, "\"");
                             // TODO: !!!!!! [🕊] Make execCommand standard (?node-)util of the promptbook
                             return [4 /*yield*/, $execCommand(command_1)];
                         case 3:
@@ -6091,7 +6090,7 @@
     $scrapersMetadataRegister.register(legacyDocumentScraperMetadata);
 
     /**
-     * Scraper for .docx files
+     * Scraper for old document files (like .doc and .rtf)
      *
      * @see `documentationUrl` for more details
      * @public exported from `@promptbook/legacy-documents`
@@ -6118,12 +6117,13 @@
          * Note: `$` is used to indicate that this function is not a pure function - it leaves files on the disk and you are responsible for cleaning them by calling `destroy` method of returned object
          */
         LegacyDocumentScraper.prototype.$convert = function (source) {
+            var _a;
             return __awaiter(this, void 0, void 0, function () {
-                var _a, _b, externalProgramsPaths, _c, rootDirname, _d, cacheDirname, _e, isCacheCleaned, _f, isVerbose, extension, cacheFilehandler, documentSourceOutdirPathForLibreOffice_1, command_1, files_1, file;
+                var _b, _c, rootDirname, _d, cacheDirname, _e, isCacheCleaned, _f, isVerbose, extension, cacheFilehandler, documentSourceOutdirPathForLibreOffice_1, command_1, files_1, file;
                 return __generator(this, function (_g) {
                     switch (_g.label) {
                         case 0:
-                            _a = this.options, _b = _a.externalProgramsPaths, externalProgramsPaths = _b === void 0 ? {} : _b, _c = _a.rootDirname, rootDirname = _c === void 0 ? process.cwd() : _c, _d = _a.cacheDirname, cacheDirname = _d === void 0 ? SCRAPE_CACHE_DIRNAME : _d, _e = _a.isCacheCleaned, isCacheCleaned = _e === void 0 ? false : _e, _f = _a.isVerbose, isVerbose = _f === void 0 ? IS_VERBOSE : _f;
+                            _b = this.options, _c = _b.rootDirname, rootDirname = _c === void 0 ? process.cwd() : _c, _d = _b.cacheDirname, cacheDirname = _d === void 0 ? SCRAPE_CACHE_DIRNAME : _d, _e = _b.isCacheCleaned, isCacheCleaned = _e === void 0 ? false : _e, _f = _b.isVerbose, isVerbose = _f === void 0 ? IS_VERBOSE : _f;
                             if (!$isRunningInNode()) {
                                 throw new KnowledgeScrapeError('Scraping .doc files is only supported in Node environment');
                             }
@@ -6131,7 +6131,7 @@
                                 throw new EnvironmentMismatchError('Can not scrape (legacy) documents without filesystem tools');
                                 //          <- TODO: [🧠] What is the best error type here`
                             }
-                            if (externalProgramsPaths.libreOfficePath === undefined) {
+                            if (((_a = this.tools.executables) === null || _a === void 0 ? void 0 : _a.libreOfficePath) === undefined) {
                                 throw new MissingToolsError('LibreOffice is required for scraping .doc and .rtf files');
                             }
                             if (source.filename === null) {
@@ -6157,7 +6157,7 @@
                             documentSourceOutdirPathForLibreOffice_1 = path.join(path.dirname(cacheFilehandler.filename), 'libreoffice')
                                 .split('\\')
                                 .join('/');
-                            command_1 = "\"".concat(externalProgramsPaths.libreOfficePath, "\" --headless --convert-to docx \"").concat(source.filename, "\"  --outdir \"").concat(documentSourceOutdirPathForLibreOffice_1, "\"");
+                            command_1 = "\"".concat(this.tools.executables.libreOfficePath, "\" --headless --convert-to docx \"").concat(source.filename, "\"  --outdir \"").concat(documentSourceOutdirPathForLibreOffice_1, "\"");
                             // TODO: !!!!!! [🕊] Make execCommand standard (?node-)util of the promptbook - this should trigger build polution error
                             return [4 /*yield*/, $execCommand(command_1)];
                         case 3: