AI Action: Generate personalized messages

[belmai]

User description

TP-2510: "Create personalized message" action

---

PR Type

New AI Action added

---

Description

• Implemented the generateMessage action to create personalized messages using LLM, with error handling and integration with OpenAI.
• Defined fields and data points for message generation, including validation schemas using Zod.
• Added extensive tests for both mocked and real OpenAI calls to ensure message generation accuracy and reliability.
• Developed constants and prompt templates to guide LLM in generating structured and compliant messages.

---

Changes walkthrough 📝

	Relevant files
Enhancement	8 files datapoints.ts Define data points for message generation                                extensions/shelly/actions/generateMessage/config/datapoints.ts Defined data points for subject and message. Set value types for data points as strings. +12/-0    fields.ts Define fields and validation schema for message form          extensions/shelly/actions/generateMessage/config/fields.ts Defined fields for message form including communicationObjective, personalizationInput, and others. Implemented validation schema using Zod. +72/-0    index.ts Export fields, validation schema, and data points                extensions/shelly/actions/generateMessage/config/index.ts Exported fields and validation schema. Exported data points. +2/-0      generateMessage.ts Implement generateMessage action for personalized messages extensions/shelly/actions/generateMessage/generateMessage.ts Implemented generateMessage action for personalized message generation. Integrated with LLM for message generation. Added error handling for message generation process. +56/-0    index.ts Export generateMessage function                                                    extensions/shelly/actions/generateMessage/index.ts Exported generateMessage function. +1/-0      constants.ts Define constants and prompt templates for LLM                        extensions/shelly/actions/generateMessage/lib/generateMessageWithLLM/constants.ts Defined constants and prompt templates for LLM message generation. Created structured output parser using Zod schema. +132/-0  generateMessageWithLLM.ts Implement generateMessageWithLLM function with retry mechanism extensions/shelly/actions/generateMessage/lib/generateMessageWithLLM/generateMessageWithLLM.ts Implemented generateMessageWithLLM function for message generation. Added retry mechanism for handling non-JSON responses. +82/-0    index.ts Export generateMessageWithLLM function                                      extensions/shelly/actions/generateMessage/lib/generateMessageWithLLM/index.ts Exported generateMessageWithLLM function. +1/-0
Tests	4 files generateMessage.test.ts Add tests for message generation with mocked LLM                  extensions/shelly/actions/generateMessage/generateMessage.test.ts Added tests for message generation with mocked LLM calls. Verified message generation with and without additional instructions. +128/-0  generateMessageRealOpenAI.test.ts Add tests for real OpenAI message generation                          extensions/shelly/actions/generateMessage/generateMessageRealOpenAI.test.ts Added tests for real OpenAI calls in message generation. Verified message content and structure. +95/-0    generateMessageWithLLM.test.ts Add unit tests for generateMessageWithLLM function              extensions/shelly/actions/generateMessage/lib/generateMessageWithLLM/generateMessageWithLLM.test.ts Added unit tests for generateMessageWithLLM function. Tested message generation for different scenarios and languages. +88/-0    generateMessageWithLLMRealOpenAI.test.ts Add integration tests for generateMessageWithLLM with real OpenAI extensions/shelly/actions/generateMessage/lib/generateMessageWithLLM/generateMessageWithLLMRealOpenAI.test.ts Added integration tests for generateMessageWithLLM using real OpenAI. Verified message content for different scenarios. +72/-0

---

💡 PR-Agent usage: Comment /help "your question" on any pull request to receive relevant information

[github-actions]

PR Reviewer Guide 🔍

Here are some key observations to aid the review process:

⏱️ Estimated effort to review: 3 🔵🔵🔵⚪⚪

🧪 PR contains tests

🔒 No security concerns identified

⚡ Recommended focus areas for review Error Handling The error handling in the 'onEvent' function could be improved by providing more specific error messages based on the type of error encountered. This would help in debugging and maintaining the code. Parsing Logic The logic to parse the AIMessageChunk seems overly complex and could potentially be simplified or better handled within the structured output chain.

[github-actions]

PR Code Suggestions ✨

No code suggestions found for the PR.