The format display of the content in the code block is incorrect.

[funtime1]

issue 1

source code：

1. test

# datetime对象 (结构化属性)
dt.year       # 年 (如2023)
dt.month      # 月 (1-12)
dt.day        # 日 (1-31)
dt.hour       # 时 (0-23)
dt.minute     # 分 (0-59)
dt.second     # 秒 (0-59)
dt.microsecond # 微秒 (0-999999)

# time.struct_time (9元组)
st.tm_year   # 年
st.tm_mon    # 月 (1-12)
st.tm_mday   # 日 (1-31)
st.tm_hour   # 时 (0-23)
st.tm_min    # 分 (0-59)
st.tm_sec    # 秒 (0-61, 60/61是闰秒)
st.tm_wday   # 周几 (0-6)
st.tm_yday   # 年中第几天 (1-366)
st.tm_isdst  # 夏令时标志 (-1/0/1)

When I exit this card again and then open the card once more, there is an error in the format display:

issue 2

The font color and the background color in the code block are the same. Is it possible to increase the contrast, that is, set them to different colors?

issue 3

When neither AI_MODEL nor AI_API_KEY is set, an error will occur when adding a collection. The error report appears in the backend logs, not on the front-end page.

# AI
AI_MODEL=
AI_API_KEY=

When I set it as a random string, for example:

# AI
AI_MODEL=111
AI_API_KEY=111

there will be no error.

issue 4

Can the API address of an AI model be customized, since it is not possible to access Google's large-scale models in China? We use deepseek.

issue 5

Is it necessary to add the function of sub-categorization?

[0010aor]

Hi funtime1,

Thank you so much for taking the time to report these issues, I truly appreciate your detailed feedback, it really helps highlight pain points and improve the experience for everyone.

Issue 1: Formatting error when reopening the card

I’ve tried to reproduce this behavior on my end, but the formatting seems consistent when I reopen the card. Could you please share more details about your environment (browser type/version, desktop or mobile)? That would really help in pinpointing the root cause.

Issue 2: Font color and background color are the same

Thanks for pointing this out. This is already tracked here:
GitHub Issue #40

It’s a simple fix, do you have any suggestions for background and font color combinations that you’d prefer? I’d be happy to take that into account.

Issue 3: No error when AI keys are unset, only backend log shows it

Great catch. I agree we should handle this more gracefully. Here’s what I’m thinking:

• Add clear front-end feedback when AI_MODEL or AI_API_KEY are missing.
• Improve tooltips and messaging, especially in Guest Mode.
• Possibly integrate it with this ticket:
  Clarify AI Unavailability in Guest Mode and Improve Disabled UI Feedback #53

Issue 4: Customizable AI model API address

Yes, I completely agree, having a customizable API endpoint would be a big improvement, especially for regions where access is limited (like China). Currently, we use Gemini because it's cost-effective. I'm also researching options that work with models like DeepSeek or Qwen.

Would accessing models via openrouter.ai or https://groq.com/ be a potential solution for your use case? Can you access it from your location? Do you perhaps know any good API alternative cost/performance efficient?

Issue 5: Sub-categorization feature

Could you elaborate a bit more on what you’d like to see here? Are you thinking of nested folders, tags within categories, or something else?

Thanks again for all this input

[funtime1]

Issue 1:

I'm using the Edge browser on my PC, and the version is 135.0.3179.98.

Issue 2:

1. It would be fine if it's like the code blocks on the current GitHub issues page.
   
2. Highlight.js or Prism.js or SyntaxHighlighter

Issue 3:

When only using "Add New Collection", there is no need to use the AI function, and in fact, there is no need to fill in information such as AI_MODEL.

Issue 4:

It is also acceptable to use openrouter.ai or https://groq.com/. I have tested and found that I can access api.groq.com and openrouter.ai.

Issue 5:

It means nested folders. Just like the multi-layered folders in Anki:
https://docs.ankiweb.net/getting-started.html#decks:~:text=Decks%20can%20contain,the%20Hanzi%20cards.