如何使用python源文件的'coding'标头正确读取其内容？

Question

Python source files often come with a coding header similar to the following

# -*- coding: iso-8859-1 -*-

How can I this line to properly parse the contents of such a file? Is there a better way than manually opening the file in binary mode, reading one line, and checking if it contains the header? Is there a library that does this?

---

Background: this comes in the context of fixing this bug , which crashes elpy when used in conjunction with python3 and importmagic. The code that I'm trying to fix uses

with open(filename) as fd:
    success = subtree.index_source(filename, fd.read())

and crashes on non-utf-8 files. Ideally I would like to keep changes to a minimum.

Answer 1

There is tokenize.open() that does exactly that: it opens a Python source file using the character encoding specified in the coding header ( encoding declaration ).

You could decode on-the-fly remote Python files too .