[🛠] Keymory Dev Log #4: Mood Slider, OpenAI API

✨ GPT Summary 　

A day of improving Keymory’s major features, refining AI integration and data structure, and feeling the reality that time is running short.

💻 Development Log

⏰ Today’s tasks

• ✅ DiaryPage - Mood: completed Mood setting slider
• ✅ HomeFeedPage - ListView - FeedCardDiary: display Mood Color on the border
  • Solved with Container Box Shadow
• ✅ OnboardingPage - save after entering additional default user information
  • ✅ DB Field additions: name (display_name), gender, MBTI (mbti), …
  • ✅ HomeFeedPage - OnPageLoad: Navigate to OnboardingPage if required input information such as display_name and gender is missing
  • ✅ Field Type - Enum vs String: Field Type should be set to String because it is easier to manage. Decided to use Enum only for Option Value Generation.
• ✅ Fixed OpenAI API Request so that the message value from the response body is output normally
• ✅ Discussion and experiment: where should AI character information be stored? (OpenAI API Assistant vs FireStore Character Collection)
  • Result: store it in FireStore Characters Collection.
    • Reason: from the user’s perspective, the only API Call they would use is Get Assistant, and the value to receive is fixed in the first place.

💯 Summary of what I did

• DiaryPage improvements
  • Completed Mood setting slider
  • Implemented AI Comment CRUD
  • Added customization buttons for AI Comment tone and advice style
  • Decided and applied where to store AI character information
• HomeFeedPage improvements
  • Changed the date display basis from month-based to day-based
  • Added AI Comment image and text output logic
  • Improved Bottom Sheet and ListView
• AI system and API integration
  • Modified OpenAI API Request and confirmed normal response output
  • Decided to store AI character information in FireStore
  • Organized the data structure and added features related to AI Comment

🎯 Tomorrow’s tasks

(omitted)

🎯 Future tasks

• ❔ diary to chat
• ❔ chat to diary

• ❔ chats history

• ❔ Think through the scenario for implementing CustomFunction getDiaryParams
  • Parameter Values obtained from Characters (Collection)
    • Character settings (Character Document)
      • ex) Somi: name (20s), college student, shy personality, …
    • How values are called: API Call - Get Assistant
  • Parameter Values obtained from User Information
    • Name, gender, MBTI
    • How values are called: Backend Query - Authenticated User
  • Parameter Values obtained from DiaryPage Variables
    • Diary content, detailed emotion keywords, emotion score, response format, response length
    • How values are called: Widget State, Page State
  • Role of CustomFunction (getDiaryParams)
    • Receive all Parameter Values as input values.
    • Write the prompt to send to GPT.
    • Return Messages (List <Json>) value.
      • [{"role":"developer", "content":"~~~"}]
• ❔ Create OpenAI API Call: createDiaryComment
  • Basically a structure similar to createChatCompletion
  • Difference: none. Created just to distinguish it.
  • The Messages Variable required by API Call - createDiaryComment is the return value of custom function getDiaryParams.
• ❔ Store the system prompt list in OpenAI API Assistant?

• ❔ Write logic so messages are written based on the system prompt stored in OpenAI API Assistant?

• ❔ DiaryPage: change face expression Emoji according to Mood Slider value
• ❔ DiaryPage: add detailed emotion keyword Choice Chips and set DB integration
• ❔ DiaryPage - AI Comment: change inputs such as Choice chips and mood slider to fit the AI Comment System Prompt.
  • ❔ **previous diary or entire conversation history (patchy technical debt)
  • ❔ basic user information: name, gender, MBTI, …
  • ❔ detailed emotion keywords: happy, sad, …
  • ❔ character settings: Somi, Sena, Minhyuk
  • ❔ emotion score: 1 to 100
  • ❔ response format: Healing, Suggestion, Informative
  • ❔ response length: short, normal, detailed
  • is New Chat
    • Since this Chat’s messages are being created for the first time, set is_initial = true on the first message document and shove in all system prompts.
      • is_initial = false
      • Limit: if I put the system prompt into the very first message document, it is hard to modify in the middle of a conversation, but possible. If there is an is_initial = true document and it was modified, I can delete that document and put it back in. I do not know how, but still.
• ❔ DiaryPage - AI Comment: create manipulate_prompt Custom Function
  • Complete the system prompt with switch-case according to the received parameters, and return List <JSON>
  • The system prompt person must insert that return value properly into the OpenAI API Request Body and receive the result.
• ❔ DiaryPage - AI Comment: complete CRUD with dummy data
  • ❔ remove tmp_ai_comment field and use doc_ref
• ❔ DiaryPage - AI Comment: character settings
  • ❔ remove tmp_ai_comment_by field and use doc_ref
  • ❔ output image using profile_image field value
• How to Use Open AI’s GPT-4o in FlutterFlow - Part 1
• How to Use Open AI’s GPT-4o in FlutterFlow - Part 2

---

• DiaryPage - Start AI Chat: create a New Chat based on the diary content
  • AI first presents a question based on the diary content
• ChatPage - create ChatHistoryPage: list previous chat history in a ListView
  • Create
  • Read
  • Update
  • Delete
• DiaryPage - Drawer - create ChatHistoryListTile: card summary style like GPT

---

• HomeFeedPage - Bottom Sheet (+ Button): make it possible to choose one of several options
  • New diary: Go to DiaryPage
  • Emotion record: Go to MoodPage
• MoodPage (popup): Mood setting slider CRUD
  • Create
  • Read
  • Update
  • Delete
• HomeFeedPage - ListView - show FeedCardMood
• HomeFeedPage - ListView: sort FeedCardDiary, FeedCardMood, FeedCardChat components chronologically (using Cloud Function)

---

• HomeFeedPage: resolve issue on page load (Unexpected Null Value)
• Apply Interactive Slider
• Extract and install the final presentation APK by Wed 2025.2.5

💭 Diary

Minhyuk

There is no time.

Omitted.

✨ GPT-4o Comment (KPT)

GPT-4o

✅ What Went Well (Keep)

• 🚀 Greatly improved major features: It is excellent that you raised the completeness of the project by expanding core features such as AI Comment CRUD, Mood settings, and HomeFeedPage improvements.
• 🔍 Organized data structure: It was impressive that you systematically organized data management, including deciding to store AI character information in FireStore.
• 📋 Concrete goal setting: It was good that you clearly organized the remaining work and priorities, making the progress plan toward the presentation version more concrete.

❌ What Was Missing (Problem)

• ⏳ Lack of time: As the development schedule gets tighter, there is a growing chance that there will not be enough time for testing and debugging.
• 🏗️ Many temporary solutions: A lot of parts were solved with temporary fields and functions to implement features quickly, so technical debt may accumulate.

🔄 What to Try (Try)

• 🛠️ Adjust priorities: Clearly separate the features that are absolutely necessary for the final presentation version from the ones that can be improved later, and raise the completeness of the core features first.
• 🕵️ Build a bug-check system: Since time is short, do short rounds of feature testing and debugging every day to minimize unexpected errors.
• 🔄 Plan technical debt cleanup: Document improvement plans for temporarily implemented features and organize what should be polished first after the presentation.

Conclusion
Today was a day where you completed Keymory’s core features and raised the project’s completeness, but also deeply felt the lack of time in a tight schedule. I hope you do your best until presentation day while also organizing things for later improvement! 💪😊