@promptbook/cli 0.92.0-7 → 0.92.0-9

package/esm/index.es.js

@@ -27,6 +27,7 @@ import * as OpenApiValidator from 'express-openapi-validator';
 import swaggerUi from 'swagger-ui-express';
 import Anthropic from '@anthropic-ai/sdk';
 import { OpenAIClient, AzureKeyCredential } from '@azure/openai';
+import Bottleneck from 'bottleneck';
 import OpenAI from 'openai';
 import { Readability } from '@mozilla/readability';
 import { JSDOM } from 'jsdom';
@@ -46,7 +47,7 @@ const BOOK_LANGUAGE_VERSION = '1.0.0';
  * @generated
  * @see https://github.com/webgptorg/promptbook
  */
-const PROMPTBOOK_ENGINE_VERSION = '0.92.0-7';
+const PROMPTBOOK_ENGINE_VERSION = '0.92.0-9';
 /**
  * TODO: string_promptbook_version should be constrained to the all versions of Promptbook engine
  * Note: [💞] Ignore a discrepancy between file name and entity name
@@ -15737,6 +15738,9 @@ const OPENAI_MODELS = exportJson({
  * Note: [💞] Ignore a discrepancy between file name and entity name
  */
 
+// Default rate limits (requests per minute) - adjust as needed based on Azure OpenAI tier
+const DEFAULT_RPM$1 = 60;
+// <- TODO: !!! Put in some better place
 /**
  * Execution Tools for calling Azure OpenAI API.
  *
@@ -15754,6 +15758,10 @@ class AzureOpenAiExecutionTools {
          * OpenAI Azure API client.
          */
         this.client = null;
+        // TODO: Allow configuring rate limits via options
+        this.limiter = new Bottleneck({
+            minTime: 60000 / (this.options.maxRequestsPerMinute || DEFAULT_RPM$1),
+        });
     }
     get title() {
         return 'Azure OpenAI';
@@ -15831,7 +15839,9 @@ class AzureOpenAiExecutionTools {
                 console.info(colors.bgWhite('messages'), JSON.stringify(messages, null, 4));
             }
             const rawRequest = [modelName, messages, modelSettings];
-            const rawResponse = await this.withTimeout(client.getChatCompletions(...rawRequest)).catch((error) => {
+            const rawResponse = await this.limiter
+                .schedule(() => this.withTimeout(client.getChatCompletions(...rawRequest)))
+                .catch((error) => {
                 if (this.options.isVerbose) {
                     console.info(colors.bgRed('error'), error);
                 }
@@ -15927,7 +15937,9 @@ class AzureOpenAiExecutionTools {
                 [rawPromptContent],
                 modelSettings,
             ];
-            const rawResponse = await this.withTimeout(client.getCompletions(...rawRequest)).catch((error) => {
+            const rawResponse = await this.limiter
+                .schedule(() => this.withTimeout(client.getCompletions(...rawRequest)))
+                .catch((error) => {
                 if (this.options.isVerbose) {
                     console.info(colors.bgRed('error'), error);
                 }
@@ -16777,6 +16789,9 @@ resultContent, rawResponse) {
  * TODO: [🤝] DRY Maybe some common abstraction between `computeOpenAiUsage` and `computeAnthropicClaudeUsage`
  */
 
+// Default rate limits (requests per minute) - adjust as needed based on OpenAI tier
+const DEFAULT_RPM = 60;
+// <- TODO: !!! Put in some better place
 /**
  * Execution Tools for calling OpenAI API
  *
@@ -16794,6 +16809,10 @@ class OpenAiExecutionTools {
          * OpenAI API client.
          */
         this.client = null;
+        // TODO: Allow configuring rate limits via options
+        this.limiter = new Bottleneck({
+            minTime: 60000 / (this.options.maxRequestsPerMinute || DEFAULT_RPM),
+        });
     }
     get title() {
         return 'OpenAI';
@@ -16897,7 +16916,9 @@ class OpenAiExecutionTools {
         if (this.options.isVerbose) {
             console.info(colors.bgWhite('rawRequest'), JSON.stringify(rawRequest, null, 4));
         }
-        const rawResponse = await client.chat.completions.create(rawRequest).catch((error) => {
+        const rawResponse = await this.limiter
+            .schedule(() => client.chat.completions.create(rawRequest))
+            .catch((error) => {
             assertsError(error);
             if (this.options.isVerbose) {
                 console.info(colors.bgRed('error'), error);
@@ -16974,7 +16995,9 @@ class OpenAiExecutionTools {
         if (this.options.isVerbose) {
             console.info(colors.bgWhite('rawRequest'), JSON.stringify(rawRequest, null, 4));
         }
-        const rawResponse = await client.completions.create(rawRequest).catch((error) => {
+        const rawResponse = await this.limiter
+            .schedule(() => client.completions.create(rawRequest))
+            .catch((error) => {
             assertsError(error);
             if (this.options.isVerbose) {
                 console.info(colors.bgRed('error'), error);
@@ -17038,7 +17061,9 @@ class OpenAiExecutionTools {
         if (this.options.isVerbose) {
             console.info(colors.bgWhite('rawRequest'), JSON.stringify(rawRequest, null, 4));
         }
-        const rawResponse = await client.embeddings.create(rawRequest).catch((error) => {
+        const rawResponse = await this.limiter
+            .schedule(() => client.embeddings.create(rawRequest))
+            .catch((error) => {
             assertsError(error);
             if (this.options.isVerbose) {
                 console.info(colors.bgRed('error'), error);