Convert python 3.x code to python 2.x

background

I've been using python3 for a long time, but it's a new development, and I wrote it in 3 because it was developed in python2, but in the end, I needed to unify it, so I investigated it. ..

Install 3to2

1. Download the zip from https://pypi.python.org/pypi/3to2
2. Open the zip and cd 3to2-x.x.x
3. python setup.py install

Try 3to2

You can check the Option by typing the following command in the terminal

3to2 --help

Usage: 3to2 [options] file|dir ...

Options:
  -h, --help            show this help message and exit
  -d, --doctests_only   Fix up doctests only
  -f FIX, --fix=FIX     Each FIX specifies a transformation; default: all
  -j PROCESSES, --processes=PROCESSES
                        Run 3to2 concurrently
  -x NOFIX, --nofix=NOFIX
                        Prevent a fixer from being run.
  -l, --list-fixes      List available transformations (fixes/fix_*.py)
  -v, --verbose         More verbose logging
  -w, --write           Write back modified files
  -n, --nobackups       Don't write backups for modified files.
  --no-diffs            Don't show diffs of the refactoring

For the time being, I decided to write directly using -w (cutting Branch). Otherwise, only the part that changes from 3 to 2 will be displayed.

3to2 -w aaaa.py

 Run and succeed!

# Other

 However, this alone may still cause an error.

## 1. Character code related

SyntaxError: Non-ASCII character '\xe6' in file aaaa.py on line 3, but no encoding declared; see http://python.org/dev/peps/pep-0263/ for details

 In this case, write `# coding = <encoding name>` on the first line according to the support of URL.
 In the Encoding name part, write ʻutf-8`

## 2. All Strings are preceded by ʻu`, which causes an error

TypeError: "delimiter" must be string, not unicode

 For example, if you convert a pandas dataframe:


 Before conversion (python3)
`df.to_csv('aaa.tsv', sep='\t')`

 After conversion (python2)
`df.to_csv(u'aaa.tsv', sep=u'\t')`

 And it says `TypeError:" delimiter "must be string, not unicode`, so
 Remove ʻu` before sep ='\ t'

## 3. Similarly, a mysterious error appears in the `to_csv` part.

UnicodeEncodeError: 'ascii' codec can't encode characters in position 0-2: ordinal not in range(128)

 In Python2, it seems that csv does not support unicode, so

 Add ʻencoding ='utf-8'`,

df.to_csv('aaa.tsv', sep='\t', encoding='utf-8')

 If so, it will be solved

## 4. cld2 can no longer be installed

 After making python 2.7.6

 `pip intall cld2` fails with the error` No package'libffi' found`

 For mac
 Install with `brew install libffi`. → Success

 I'm not sure because I haven't done it for Ubuntu, but is it okay to use ʻapt-get install libffi` or `libffi-dev`? .. ??

## 5. Unicode-related mystery error 2

TypeError: initializer for ctype 'char *' must be a str or list or tuple, not unicode

 The one passed to cld2 above was unicode, so change it as follows

 Change before

reliable, _, lang = cld2.detect(name)

 After change

reliable, _, lang = cld2.detect(name.encode('utf-8'))

# Conclusion from the error

 Python2 has a lot of errors in Unicode, so be careful!


# bonus

## version control
 For those who use both Python 2 and 3, it may be convenient to use `pyenv` and separate it by Local.

### Check which version you have

pyenv versions

### Install the required version

pyenv install 2.7.10

### Set locally

pyenv local 2.7.10

 This will create a file called .python-verion with the version inside. If you delete this file, you can use the global python version.


## Note the 3to2 execution Python version
 I installed python3, so it is said that there is no `3to2` command in the place where I changed to local version python2 with pyenv. So, when changing multiple files, you need to think about whether to execute `3to2` in python3 and then switch to python2 or install both python3 and 2.