python - Telegram Bot with MarkdownV2: "Can't parse entities: can't find end of code entity" e

admin管理员组

文章数量:1295884

I'm developing a Telegram bot using the latest version of the python-telegram-bot library (v20.x) and need to send messages in MarkdownV2 format. My goal is to include static formatting (such as a code block using triple backticks) alongside dynamic text generated by an AI (or received from another source).

Here's a simplified version of my code:

from telegram.constants import ParseMode
from telegram.helpers import escape_markdown

async def send_ai_message(update, context):
    prompt = update.message.text
    await update.message.reply_text("Thinking...")
    
    # Get dynamic text from the AI
    response_text = ai.chat_ai(prompt)
    # Escape dynamic text for MarkdownV2
    safe_response = escape_markdown(response_text, version=2)
    
    # Combine the safe dynamic text with a static code block
    message_text = f"```python\n{safe_response}\n```"
    
    await context.bot.send_message(
        chat_id=update.effective_chat.id,
        text=message_text,
        parse_mode=ParseMode.MARKDOWN_V2
    )

However, when I run this code, I get the following error:

Bad Request: Can't parse entities: can't find end of code entity at byte offset ...

I understand that the issue likely stems from MarkdownV2’s strict formatting rules, where certain special characters (e.g., _, *, `, [, ], etc.) must be properly escaped. In my case, I need to preserve the static Markdown formatting (the code block markers) while ensuring that the dynamic text is fully escaped.

My questions are:

What could be causing this error, and why does it occur even after using escape_markdown on the dynamic text? How can I safely combine static Markdown formatting with dynamic text so that the message parses correctly using MarkdownV2? Are there any additional considerations or best practices when working with MarkdownV2 in the latest version of python-telegram-bot (noting that ParseMode is now in telegram.constants)?

What did I try?

I attempted to combine a statically formatted code block with dynamic text generated by my AI. I used the escape_markdown function (with version 2 for MarkdownV2) to process the dynamic content and then inserted that result into a code block using triple backticks. For example:

safe_response = escape_markdown(response_text, version=2)
message_text = f"```python\n{safe_response}\n```"

I then sent the message with parse_mode=ParseMode.MARKDOWN_V2.

What was I expecting?

I expected that the resulting message would be correctly parsed by Telegram—displaying a nicely formatted code block containing the dynamic text with all special MarkdownV2 characters properly escaped. However, instead of the expected output, I received the error:

Bad Request: Can't parse entities: can't find end of code entity at byte offset ...

This indicates that despite escaping the dynamic text, there is still a formatting issue (likely due to unbalanced or improperly escaped markdown characters) when combined with the static code block markers.

本文标签：