I want to do some code cleanup and make all of the strings within a module consistently double-quoted strings. My problem is there are tons of them and going through by hand (alt+enter, arrow, enter) is tedious and my regex-Fu is weak. (keep running into oddball cases where single quotes are in doc strings, etc)
Is there some faster way to convert all single-quoted strings to double-quoted strings in Pycharm?
In Preferences, type double-quoted into the search bar. This will suggest Editor->Intentions. There's an option there for converting double-quoted strings to single-quoted strings and vice-versa. Check this to see if it is enabled.
When you select one instance of such a string, and click on the 'light bulb', this may, or may not, offer you the option to apply it to the entire file by tapping the right arrow.
If not, it may at least ease your pain in that you can simply click once on each instance and use the light bulb to fix that on. Finding documentation on Intentions isn't all that fruitful, regarding how to apply them across a file or include them in code inspection. This may be a difference between the Community edition and the Full Version.
Though, as @acushner's comment suggests, and in my personal experience and preference, single-quoting seems the more consistently-used styling, and arguably sub rosa canon.
(To be honest, it's so ingrained in me at this point that I find myself single-quoting prose. But I reckon that's a personal problem, and TMI. 'Be yourself.')
I think there's an easier solution not being discussed, which is PyCharm's find and replace feature to ignore comments and string literals, which seems like it would do exactly what you want without much work and with little room for edge cases:
To ignore triple single quotes, you can use this regex, which I copied from this SO post:
(?<!')'(?!')
'\'foo\''; 2) Your solution still cannot be applied to a whole file, since it would blindly replace the enclosing single quotes of strings containing double quotes (like '"foo"'), without escaping them; 3) There is a bug in PyCharm that causes the opening quote of a prefixed string like an f-string to be considered "inside" the string, and thus ignored. Only the closing quote of f"foo{bar}" will be found! –
Accouchement f'foo \'{bar}\'' would be found by the recipe of this answer –
Accouchement f'foo{bar}' –
Accouchement f"foo{bar['baz']}" are found (they can't be simply replaced!) –
Accouchement I know you asked how to do it in PyCharm, but I'll propose a tool instead, since in my tests it works flawlessly. You can still run it from within PyCharm's terminal window, which makes this a valid answer to your question ;-).
The tool (source code below) is based on Python's tokenizer. This guarantees that only single quotes enclosing string literals can be replaced.
Furthermore, conversions that would not be straightforward and safe, and possibly not even desirable, are avoided. The file positions that may need manual intervention are reported.
With the help of the find command you can apply this tool to a whole source tree at once.
import argparse
import re
from tokenize import generate_tokens, STRING, untokenize
# Tool for the automatic (and manual) conversion of single-quoted strings to double-quoted ones
# Reference for string literals: https://docs.python.org/3/reference/lexical_analysis.html#string-and-bytes-literals
LONG_STRING_REGEX = re.compile(r"([fruFRU]*)'''(.*)'''", re.DOTALL)
SHORT_STRING_REGEX = re.compile(r"([fruFRU]*)'(.*)'") # only if LONG_STRING_REGEX doesn't match
def single_to_double_quotes(infilename, outfilename, long_too):
convs_done = 0
manual_conv_starts = []
output_tokens = []
with open(infilename) as infile:
for t in generate_tokens(infile.readline):
if t.type == STRING:
if m := LONG_STRING_REGEX.fullmatch(t.string):
if long_too:
prefix, value = m.groups()
if '"""' in value:
manual_conv_starts.append(t.start)
else:
t = (STRING, f'{prefix}"""{value}"""', t.start, t.end, t.line)
convs_done += 1
elif m := SHORT_STRING_REGEX.fullmatch(t.string):
prefix, value = m.groups()
if '"' in value or "'" in value:
manual_conv_starts.append(t.start)
else:
t = (STRING, f'{prefix}"{value}"', t.start, t.end, t.line)
convs_done += 1
output_tokens.append(t)
with open(outfilename, 'w') as outfile:
outfile.write(untokenize(output_tokens))
return convs_done, manual_conv_starts
parser = argparse.ArgumentParser(description='safely convert single quotes to double quotes')
parser.add_argument('infilename', metavar='INPUT_FILE', help='the input file')
parser.add_argument('outfilename', metavar='OUTPUT_FILE', help='the output file')
parser.add_argument('--long', '-l', action='store_true', help='also convert "long" (triple-quoted) strings')
args = parser.parse_args() # type: argparse.Namespace
infilename = args.infilename # type: str
outfilename = args.outfilename # type: str
long_too = bool(args.long)
print(f"input file: {infilename}")
print(f"long string conversion {('disabled', 'enabled')[long_too]}")
convs_done, manual_conv_starts = single_to_double_quotes(infilename, outfilename, long_too)
print(f"strings converted: {convs_done}")
print(f"possible manual conversions: {len(manual_conv_starts)}")
for start in manual_conv_starts:
print(f" {outfilename} line {start[0]} char {start[1] + 1}")
print(f"output file: {outfilename}")
Example output (without -l option):
input file: single_to_double_quotes_TEST.py
long string conversion disabled
strings converted: 5
possible manual conversions: 3
single_to_double_quotes_CONV.py line 3 char 5
single_to_double_quotes_CONV.py line 4 char 5
single_to_double_quotes_CONV.py line 12 char 7
output file: single_to_double_quotes_CONV.py
single_to_double_quotes_TEST.py:
n = 123
s = 'abc'
t = 'd"e"f'
u = 'g\'h\'i'
l = '''
"let's see"
'''
w, x, y, z = '''xyz''', '''"""yz''', '''x"""z''', '''xy"""'''
print(f'{s=}')
d = {'a': 1, 'b': 2, 'c': 3}
print(f"{d['a']=}")
print(f'{d["a"]=}')
single_to_double_quotes_CONV.py:
n = 123
s = "abc"
t = 'd"e"f'
u = 'g\'h\'i'
l = '''
"let's see"
'''
w, x, y, z = '''xyz''', '''"""yz''', '''x"""z''', '''xy"""'''
print(f"{s=}")
d = {"a": 1, "b": 2, "c": 3}
print(f"{d['a']=}")
print(f'{d["a"]=}')
On a copy of your source tree, like on a git checkout, you can do the following:
find . -name \*.py -exec python3 PATH/TO/TOOL.py {} {} \;
This will convert each file onto itself.
In case you are not familiar with argparse: start from
python3 TOOL.py -h
I recommend you modify the tool to best fit your needs.
'"don\'t do it" they said', "\"don't do it\" they said", '''"don't do it" they said''' and """"don't do it" they said"""? How to deal with, e.g., f"""{d['''"don't do it" they said''']}!""". I had a couple of working solutions, but they were not good enough to be published, and now I ran out of time. I'm sorry, I have to postpone this to some longer vacation. –
Accouchement This can be done using black formatting
Install black by doing pip install black
Then do black . to perform the formatting.
This will not just take care of the single quotes, but will format your code to a much readable way. There are a lot of options that you can provide to the black cli to customize how black does the formatting.
You can check those options here: https://black.readthedocs.io/en/stable/usage_and_configuration/the_basics.html#command-line-options
black . --quotes-only) –
Gardner I don't know if it helps anyone but I prefer to select one word with a double click instead of selecting the whole string between quotes and then press Alt + Enter. This way you open the actions menu and select "Convert single quoted string to double quotes" or vice versa. You don't even have to double click since Pycharm is smart enough to select the first pair of quotes. I find this very comfortable. I don't know if it is smart enough for very complicated nested strings but you can try. It would be nice though, if it would let us select multiple lines and fix them too but that's what black is here for, right?
install black in global
pip install black
1、use file watcher
add black to file watcher
enable it
2、or use blackconnect
3、I just found pycharm already support some usage
Ctrl+R Will bring up the find/replace dialog. Just put ' in the find field and " in the replace field.
© 2022 - 2024 — McMap. All rights reserved.
blackpython formatter, it opts for double-quoted strings by default and can easily be integrated with PyCharm. – Strangeblack– Restful:, then%s/'/"/gand hit enter. – Lacerationlib2to3, which ships with CPython. Another option is libcst. Putting aside the handling of f-strings and triple quoted strings, you should be able to parse a module into CST, change every string's quote type, and generate the source code again, with nothing else changing. – Flyweight