@promptbook/cli 0.92.0-11 → 0.92.0-12

package/esm/index.es.js

@@ -47,7 +47,7 @@ const BOOK_LANGUAGE_VERSION = '1.0.0';
  * @generated
  * @see https://github.com/webgptorg/promptbook
  */
-const PROMPTBOOK_ENGINE_VERSION = '0.92.0-11';
+const PROMPTBOOK_ENGINE_VERSION = '0.92.0-12';
 /**
  * TODO: string_promptbook_version should be constrained to the all versions of Promptbook engine
  * Note: [💞] Ignore a discrepancy between file name and entity name
@@ -4944,6 +4944,35 @@ function isPipelinePrepared(pipeline) {
  *     - [♨] Are tasks prepared
  */
 
+/**
+ * Serializes an error into a [🚉] JSON-serializable object
+ *
+ * @public exported from `@promptbook/utils`
+ */
+function serializeError(error) {
+    const { name, message, stack } = error;
+    const { id } = error;
+    if (!Object.keys(ALL_ERRORS).includes(name)) {
+        console.error(spaceTrim((block) => `
+
+                    Cannot serialize error with name "${name}"
+
+                    Authors of Promptbook probably forgot to add this error into the list of errors:
+                    https://github.com/webgptorg/promptbook/blob/main/src/errors/0-index.ts
+
+
+                    ${block(stack || message)}
+
+                `));
+    }
+    return {
+        name: name,
+        message,
+        stack,
+        id, // Include id in the serialized object
+    };
+}
+
 /**
  * Recursively converts JSON strings to JSON objects
 
@@ -5113,35 +5142,6 @@ function createTask(options) {
  * TODO: [🐚] Split into more files and make `PrepareTask` & `RemoteTask` + split the function
  */
 
-/**
- * Serializes an error into a [🚉] JSON-serializable object
- *
- * @public exported from `@promptbook/utils`
- */
-function serializeError(error) {
-    const { name, message, stack } = error;
-    const { id } = error;
-    if (!Object.keys(ALL_ERRORS).includes(name)) {
-        console.error(spaceTrim((block) => `
-
-                    Cannot serialize error with name "${name}"
-
-                    Authors of Promptbook probably forgot to add this error into the list of errors:
-                    https://github.com/webgptorg/promptbook/blob/main/src/errors/0-index.ts
-
-
-                    ${block(stack || message)}
-
-                `));
-    }
-    return {
-        name: name,
-        message,
-        stack,
-        id, // Include id in the serialized object
-    };
-}
-
 /**
  * Format either small or big number
  *
@@ -5393,6 +5393,24 @@ const MANDATORY_CSV_SETTINGS = Object.freeze({
     // encoding: 'utf-8',
 });
 
+/**
+ * Converts a CSV string into an object
+ *
+ * Note: This is wrapper around `papaparse.parse()` with better autohealing
+ *
+ * @private - for now until `@promptbook/csv` is released
+ */
+function csvParse(value /* <- TODO: string_csv */, settings, schema /* <- TODO: Make CSV Schemas */) {
+    settings = { ...settings, ...MANDATORY_CSV_SETTINGS };
+    // Note: Autoheal invalid '\n' characters
+    if (settings.newline && !settings.newline.includes('\r') && value.includes('\r')) {
+        console.warn('CSV string contains carriage return characters, but in the CSV settings the `newline` setting does not include them. Autohealing the CSV string.');
+        value = value.replace(/\r\n/g, '\n').replace(/\r/g, '\n');
+    }
+    const csv = parse(value, settings);
+    return csv;
+}
+
 /**
  * Function to check if a string is valid CSV
  *
@@ -5415,31 +5433,13 @@ function isValidCsvString(value) {
     }
 }
 
-/**
- * Converts a CSV string into an object
- *
- * Note: This is wrapper around `papaparse.parse()` with better autohealing
- *
- * @private - for now until `@promptbook/csv` is released
- */
-function csvParse(value /* <- TODO: string_csv */, settings, schema /* <- TODO: Make CSV Schemas */) {
-    settings = { ...settings, ...MANDATORY_CSV_SETTINGS };
-    // Note: Autoheal invalid '\n' characters
-    if (settings.newline && !settings.newline.includes('\r') && value.includes('\r')) {
-        console.warn('CSV string contains carriage return characters, but in the CSV settings the `newline` setting does not include them. Autohealing the CSV string.');
-        value = value.replace(/\r\n/g, '\n').replace(/\r/g, '\n');
-    }
-    const csv = parse(value, settings);
-    return csv;
-}
-
 /**
  * Definition for CSV spreadsheet
  *
  * @public exported from `@promptbook/core`
  *                          <- TODO: [🏢] Export from package `@promptbook/csv`
  */
-const CsvFormatDefinition = {
+const CsvFormatParser = {
     formatName: 'CSV',
     aliases: ['SPREADSHEET', 'TABLE'],
     isValid(value, settings, schema) {
@@ -5451,7 +5451,7 @@ const CsvFormatDefinition = {
     heal(value, settings, schema) {
         throw new Error('Not implemented');
     },
-    subvalueDefinitions: [
+    subvalueParsers: [
         {
             subvalueName: 'ROW',
             async mapValues(value, outputParameterName, settings, mapCallback) {
@@ -5512,10 +5512,10 @@ const CsvFormatDefinition = {
     ],
 };
 /**
- * TODO: [🍓] In `CsvFormatDefinition` implement simple `isValid`
- * TODO: [🍓] In `CsvFormatDefinition` implement partial `canBeValid`
- * TODO: [🍓] In `CsvFormatDefinition` implement `heal
- * TODO: [🍓] In `CsvFormatDefinition` implement `subvalueDefinitions`
+ * TODO: [🍓] In `CsvFormatParser` implement simple `isValid`
+ * TODO: [🍓] In `CsvFormatParser` implement partial `canBeValid`
+ * TODO: [🍓] In `CsvFormatParser` implement `heal
+ * TODO: [🍓] In `CsvFormatParser` implement `subvalueParsers`
  * TODO: [🏢] Allow to expect something inside CSV objects and other formats
  */
 
@@ -5524,7 +5524,7 @@ const CsvFormatDefinition = {
  *
  * @private still in development [🏢]
  */
-const JsonFormatDefinition = {
+const JsonFormatParser = {
     formatName: 'JSON',
     mimeType: 'application/json',
     isValid(value, settings, schema) {
@@ -5536,28 +5536,28 @@ const JsonFormatDefinition = {
     heal(value, settings, schema) {
         throw new Error('Not implemented');
     },
-    subvalueDefinitions: [],
+    subvalueParsers: [],
 };
 /**
  * TODO: [🧠] Maybe propper instance of object
  * TODO: [0] Make string_serialized_json
  * TODO: [1] Make type for JSON Settings and Schema
  * TODO: [🧠] What to use for validating JSONs - JSON Schema, ZoD, typescript types/interfaces,...?
- * TODO: [🍓] In `JsonFormatDefinition` implement simple `isValid`
- * TODO: [🍓] In `JsonFormatDefinition` implement partial `canBeValid`
- * TODO: [🍓] In `JsonFormatDefinition` implement `heal
- * TODO: [🍓] In `JsonFormatDefinition` implement `subvalueDefinitions`
+ * TODO: [🍓] In `JsonFormatParser` implement simple `isValid`
+ * TODO: [🍓] In `JsonFormatParser` implement partial `canBeValid`
+ * TODO: [🍓] In `JsonFormatParser` implement `heal
+ * TODO: [🍓] In `JsonFormatParser` implement `subvalueParsers`
  * TODO: [🏢] Allow to expect something inside JSON objects and other formats
  */
 
 /**
  * Definition for any text - this will be always valid
  *
- * Note: This is not useful for validation, but for splitting and mapping with `subvalueDefinitions`
+ * Note: This is not useful for validation, but for splitting and mapping with `subvalueParsers`
  *
  * @public exported from `@promptbook/core`
  */
-const TextFormatDefinition = {
+const TextFormatParser = {
     formatName: 'TEXT',
     isValid(value) {
         return typeof value === 'string';
@@ -5566,9 +5566,9 @@ const TextFormatDefinition = {
         return typeof partialValue === 'string';
     },
     heal() {
-        throw new UnexpectedError('It does not make sense to call `TextFormatDefinition.heal`');
+        throw new UnexpectedError('It does not make sense to call `TextFormatParser.heal`');
     },
-    subvalueDefinitions: [
+    subvalueParsers: [
         {
             subvalueName: 'LINE',
             async mapValues(value, outputParameterName, settings, mapCallback) {
@@ -5588,10 +5588,10 @@ const TextFormatDefinition = {
 /**
  * TODO: [1] Make type for XML Text and Schema
  * TODO: [🧠][🤠] Here should be all words, characters, lines, paragraphs, pages available as subvalues
- * TODO: [🍓] In `TextFormatDefinition` implement simple `isValid`
- * TODO: [🍓] In `TextFormatDefinition` implement partial `canBeValid`
- * TODO: [🍓] In `TextFormatDefinition` implement `heal
- * TODO: [🍓] In `TextFormatDefinition` implement `subvalueDefinitions`
+ * TODO: [🍓] In `TextFormatParser` implement simple `isValid`
+ * TODO: [🍓] In `TextFormatParser` implement partial `canBeValid`
+ * TODO: [🍓] In `TextFormatParser` implement `heal
+ * TODO: [🍓] In `TextFormatParser` implement `subvalueParsers`
  * TODO: [🏢] Allow to expect something inside each item of list and other formats
  */
 
@@ -5624,7 +5624,7 @@ function isValidXmlString(value) {
  *
  * @private still in development [🏢]
  */
-const XmlFormatDefinition = {
+const XmlFormatParser = {
     formatName: 'XML',
     mimeType: 'application/xml',
     isValid(value, settings, schema) {
@@ -5636,17 +5636,17 @@ const XmlFormatDefinition = {
     heal(value, settings, schema) {
         throw new Error('Not implemented');
     },
-    subvalueDefinitions: [],
+    subvalueParsers: [],
 };
 /**
  * TODO: [🧠] Maybe propper instance of object
  * TODO: [0] Make string_serialized_xml
  * TODO: [1] Make type for XML Settings and Schema
  * TODO: [🧠] What to use for validating XMLs - XSD,...
- * TODO: [🍓] In `XmlFormatDefinition` implement simple `isValid`
- * TODO: [🍓] In `XmlFormatDefinition` implement partial `canBeValid`
- * TODO: [🍓] In `XmlFormatDefinition` implement `heal
- * TODO: [🍓] In `XmlFormatDefinition` implement `subvalueDefinitions`
+ * TODO: [🍓] In `XmlFormatParser` implement simple `isValid`
+ * TODO: [🍓] In `XmlFormatParser` implement partial `canBeValid`
+ * TODO: [🍓] In `XmlFormatParser` implement `heal
+ * TODO: [🍓] In `XmlFormatParser` implement `subvalueParsers`
  * TODO: [🏢] Allow to expect something inside XML and other formats
  */
 
@@ -5655,12 +5655,7 @@ const XmlFormatDefinition = {
  *
  * @private internal index of `...` <- TODO [🏢]
  */
-const FORMAT_DEFINITIONS = [
-    JsonFormatDefinition,
-    XmlFormatDefinition,
-    TextFormatDefinition,
-    CsvFormatDefinition,
-];
+const FORMAT_DEFINITIONS = [JsonFormatParser, XmlFormatParser, TextFormatParser, CsvFormatParser];
 /**
  * Note: [💞] Ignore a discrepancy between file name and entity name
  */
@@ -5830,7 +5825,7 @@ function extractJsonBlock(markdown) {
 }
 /**
  * TODO: Add some auto-healing logic + extract YAML, JSON5, TOML, etc.
- * TODO: [🏢] Make this logic part of `JsonFormatDefinition` or `isValidJsonString`
+ * TODO: [🏢] Make this logic part of `JsonFormatParser` or `isValidJsonString`
  */
 
 /**
@@ -6051,7 +6046,7 @@ const CountUtils = {
     PAGES: countPages,
 };
 /**
- * TODO: [🧠][🤠] This should be probbably as part of `TextFormatDefinition`
+ * TODO: [🧠][🤠] This should be probbably as part of `TextFormatParser`
  * Note: [💞] Ignore a discrepancy between file name and entity name
  */
 
@@ -6079,7 +6074,7 @@ function checkExpectations(expectations, value) {
 }
 /**
  * TODO: [💝] Unite object for expecting amount and format
- * TODO: [🧠][🤠] This should be part of `TextFormatDefinition`
+ * TODO: [🧠][🤠] This should be part of `TextFormatParser`
  * Note: [💝] and [🤠] are interconnected together
  */
 
@@ -6307,7 +6302,7 @@ async function executeAttempts(options) {
             if (task.format) {
                 if (task.format === 'JSON') {
                     if (!isValidJsonString($ongoingTaskResult.$resultString || '')) {
-                        // TODO: [🏢] Do more universally via `FormatDefinition`
+                        // TODO: [🏢] Do more universally via `FormatParser`
                         try {
                             $ongoingTaskResult.$resultString = extractJsonBlock($ongoingTaskResult.$resultString || '');
                         }
@@ -6445,16 +6440,16 @@ async function executeFormatSubvalues(options) {
                     ${block(pipelineIdentification)}
                 `));
     }
-    const subvalueDefinition = formatDefinition.subvalueDefinitions.find((subvalueDefinition) => [subvalueDefinition.subvalueName, ...(subvalueDefinition.aliases || [])].includes(task.foreach.subformatName));
-    if (subvalueDefinition === undefined) {
+    const subvalueParser = formatDefinition.subvalueParsers.find((subvalueParser) => [subvalueParser.subvalueName, ...(subvalueParser.aliases || [])].includes(task.foreach.subformatName));
+    if (subvalueParser === undefined) {
         throw new UnexpectedError(
         // <- TODO: [🧠][🧐] Should be formats fixed per promptbook version or behave as plugins (=> change UnexpectedError)
         spaceTrim((block) => `
                     Unsupported subformat name "${task.foreach.subformatName}" for format "${task.foreach.formatName}"
 
                     Available subformat names for format "${formatDefinition.formatName}":
-                    ${block(formatDefinition.subvalueDefinitions
-            .map((subvalueDefinition) => subvalueDefinition.subvalueName)
+                    ${block(formatDefinition.subvalueParsers
+            .map((subvalueParser) => subvalueParser.subvalueName)
             .map((subvalueName) => `- ${subvalueName}`)
             .join('\n'))}
 
@@ -6468,7 +6463,7 @@ async function executeFormatSubvalues(options) {
         formatSettings = csvSettings;
         // <- TODO: [🤹‍♂️] More universal, make simmilar pattern for other formats for example \n vs \r\n in text
     }
-    const resultString = await subvalueDefinition.mapValues(parameterValue, task.foreach.outputSubparameterName, formatSettings, async (subparameters, index) => {
+    const resultString = await subvalueParser.mapValues(parameterValue, task.foreach.outputSubparameterName, formatSettings, async (subparameters, index) => {
         let mappedParameters;
         // TODO: [🤹‍♂️][🪂] Limit to N concurrent executions
         // TODO: When done [🐚] Report progress also for each subvalue here
@@ -6530,6 +6525,27 @@ async function getExamplesForTask(task) {
     return RESERVED_PARAMETER_MISSING_VALUE /* <- TODO: [♨] Implement */;
 }
 
+/**
+ * Computes the cosine similarity between two embedding vectors
+ *
+ * Note: This is helping function for RAG (retrieval-augmented generation)
+ *
+ * @param embeddingVector1
+ * @param embeddingVector2
+ * @returns Cosine similarity between the two vectors
+ *
+ * @public exported from `@promptbook/core`
+ */
+function computeCosineSimilarity(embeddingVector1, embeddingVector2) {
+    if (embeddingVector1.length !== embeddingVector2.length) {
+        throw new TypeError('Embedding vectors must have the same length');
+    }
+    const dotProduct = embeddingVector1.reduce((sum, value, index) => sum + value * embeddingVector2[index], 0);
+    const magnitude1 = Math.sqrt(embeddingVector1.reduce((sum, value) => sum + value * value, 0));
+    const magnitude2 = Math.sqrt(embeddingVector2.reduce((sum, value) => sum + value * value, 0));
+    return 1 - dotProduct / (magnitude1 * magnitude2);
+}
+
 /**
  * @@@
  *
@@ -6556,7 +6572,7 @@ async function getKnowledgeForTask(options) {
         },
         content: task.content,
         parameters: {
-        /* !!!!!!!! */
+        /* !!!! */
         },
     };
     const taskEmbeddingResult = await llmTools.callEmbeddingModel(taskEmbeddingPrompt);
@@ -6591,16 +6607,6 @@ async function getKnowledgeForTask(options) {
     return knowledgePiecesLimited.map(({ content }) => `- ${content}`).join('\n');
     //                                                      <- TODO: [🧠] Some smart aggregation of knowledge pieces, single-line vs multi-line vs mixed
 }
-// TODO: !!!!!! Annotate + to new file
-function computeCosineSimilarity(embeddingVector1, embeddingVector2) {
-    if (embeddingVector1.length !== embeddingVector2.length) {
-        throw new TypeError('Embedding vectors must have the same length');
-    }
-    const dotProduct = embeddingVector1.reduce((sum, value, index) => sum + value * embeddingVector2[index], 0);
-    const magnitude1 = Math.sqrt(embeddingVector1.reduce((sum, value) => sum + value * value, 0));
-    const magnitude2 = Math.sqrt(embeddingVector2.reduce((sum, value) => sum + value * value, 0));
-    return 1 - dotProduct / (magnitude1 * magnitude2);
-}
 /**
  * TODO: !!!! Verify if this is working
  * TODO: [♨] Implement Better - use keyword search
@@ -7103,6 +7109,22 @@ function createPipelineExecutor(options) {
             cacheDirname,
             intermediateFilesStrategy,
             isAutoInstalled,
+        }).catch((error) => {
+            assertsError(error);
+            return exportJson({
+                name: 'pipelineExecutorResult',
+                message: `Unuccessful PipelineExecutorResult, last catch`,
+                order: [],
+                value: {
+                    isSuccessful: false,
+                    errors: [serializeError(error)],
+                    warnings: [],
+                    usage: UNCERTAIN_USAGE,
+                    executionReport: null,
+                    outputParameters: {},
+                    preparedPipeline,
+                },
+            });
         });
     };
     const pipelineExecutor = (inputParameters) => createTask({
@@ -8622,14 +8644,14 @@ const foreachCommandParser = {
                     `));
             // <- TODO: [🏢] List all supported format names
         }
-        const subvalueDefinition = formatDefinition.subvalueDefinitions.find((subvalueDefinition) => [subvalueDefinition.subvalueName, ...(subvalueDefinition.aliases || [])].includes(subformatName));
-        if (subvalueDefinition === undefined) {
+        const subvalueParser = formatDefinition.subvalueParsers.find((subvalueParser) => [subvalueParser.subvalueName, ...(subvalueParser.aliases || [])].includes(subformatName));
+        if (subvalueParser === undefined) {
             throw new ParseError(spaceTrim((block) => `
                         Unsupported subformat name "${subformatName}" for format "${formatName}"
 
                         Available subformat names for format "${formatDefinition.formatName}":
-                        ${block(formatDefinition.subvalueDefinitions
-                .map((subvalueDefinition) => subvalueDefinition.subvalueName)
+                        ${block(formatDefinition.subvalueParsers
+                .map((subvalueParser) => subvalueParser.subvalueName)
                 .map((subvalueName) => `- ${subvalueName}`)
                 .join('\n'))}
                     `));
@@ -13255,10 +13277,10 @@ function $initializeRunCommand(program) {
             console.info(colors.gray('--- Detailed Result ---'));
             console.info({ isSuccessful, errors, warnings, outputParameters, executionReport });
         }
-        if (saveReport && saveReport.endsWith('.json')) {
+        if (executionReport !== null && saveReport && saveReport.endsWith('.json')) {
             await writeFile(saveReport, JSON.stringify(executionReport, null, 4) + '\n', 'utf-8');
         }
-        else if (saveReport && saveReport.endsWith('.md')) {
+        else if (executionReport !== null && saveReport && saveReport.endsWith('.md')) {
             const executionReportString = executionReportJsonToString(executionReport);
             await writeFile(saveReport, executionReportString, 'utf-8');
         }