ödəyici turns into ödÉ™yici

In UTF-8, ə is two bytes: C9 99. Windows-1252 reads those bytes as letters: C9 is É, and 99 is ™. So ə shows as É™. The same path turns ö into ö. ödəyici is then written ödəyici.

Turkish pages give a table for ş, ğ, and ı. É™ is not in that table, because Turkish has no ə.

If the letter is already ?, this code cannot bring it back. The byte is gone.

• Use the code as it is: metn.encode("cp1252").decode("utf-8"). ödəyici becomes ödəyici again.
• When you see É™, do not look for a replacement table. That is ə read wrongly as Windows-1252.
• A Turkish table for ş, ğ, and ı does not fix É™. ə is not in that table.
• If the code throws, the text already has a mark that is not a cp1252 byte. Type that place by hand.
• If the letter is ?, the byte is gone. Take the text from the source again.

Save the fixed text as utf-8 again. Do not read the same file as cp1252 a second time.

def duzelt(metn: str) -> str:
    return metn.encode("cp1252").decode("utf-8")

print(duzelt("ödəyici"))