interpretation of special characters in Python

[TP]

Hi everybody,

I am new to Python, I try to understand how Python treats special
characters. For example, if I execute the following line in a shell
console, I obtain a colored string:

$ python -c "print '\033[30;44m foo \033[0m'"

So, it works.
Some time ago, I have made a lot of shell variables with all possible colors
(with shell functions executed each time I launch a new shell). For
example:

$ echo -e $esc$ColorBlack OnDarkblue foo $esc$ColorReset

gives the same result than the python command above.
To know the corresponding non-interpreted characters, I can use the -n
option of echo:

$ echo -n $esc$ColorBlack OnDarkblue foo $esc$ColorReset
\033[30;44m foo \033[0m

So, it is exactly the string above, as expected.

My problem arises when it comes to get these shell variables ( $esc,
$ColorBlackOnDa rkblue, $ColorReset) in a Python script, with os.environ, in
the following 5-line script:

import os
Color = os.environ['ColorBlackOnDa rkblue']
ColorReset = os.environ['ColorReset']
Esc = os.environ['esc']
print '%s%s%s%s%s' % (Esc, Color, " foo ", Esc, ColorReset)

Run this script color.py, but after having defined these shell variables in
a shell:

$ export esc="\033"
$ export ColorBlackOnDar kblue="[30;44m"
$ export ColorReset="[0m"

When I execute the Python script, I do not obtain any special character
interpretation in Python:

$ python color.py
\033[30;44m foo \033[0m

Why? What is the problem? Is there any solution?
I really want to get my shell color variables.

Thanks a lot

--
TP (Tribulations Parallèles)

"Allez, Monsieur, allez, et la foi vous viendra." (D'Alembert).

[Peter Pearson]

On Sun, 06 Jul 2008 23:42:26 +0200, TP <Tr**********@P aralleles.inval idwrote:

>
$ python -c "print '\033[30;44m foo \033[0m'"

[writes an escape sequence to stdout]

$ echo -e $esc$ColorBlack OnDarkblue foo $esc$ColorReset

[also writes an escape sequence to stdout]

$ echo -n $esc$ColorBlack OnDarkblue foo $esc$ColorReset
\033[30;44m foo \033[0m

[snip, shuffle]

$ export esc="\033"
$ export ColorBlackOnDar kblue="[30;44m"
$ export ColorReset="[0m"

import os
Color = os.environ['ColorBlackOnDa rkblue']
ColorReset = os.environ['ColorReset']
Esc = os.environ['esc']
print '%s%s%s%s%s' % (Esc, Color, " foo ", Esc, ColorReset)

[snip]

$ python color.py
\033[30;44m foo \033[0m

The string "\033" is 4 characters long. Your shell variable
"esc" is 4 characters long. Your Python program prints
those four characters. You want it to re-interpret those 4
characters into a single escape character.

One of this group's regular participants can (I hope) tell
us three breathtakingly elegant ways to do that. I'm sorry
I can't.

When you run echo, it recognizes the 4-character "esc" as a
convention for representing a single character, and performs
the re-interpretation for you. When you tell python
"print '\033[30;44m foo \033[0m'", python interprets
the "\033" as a single character.

--
To email me, substitute nowhere->spamcop, invalid->net.

[TP]

Dennis Lee Bieber wrote:

Off-hand, I'd probably try first with:

csi = "\033["

and then define your

colorblackondar kblue = $csi"30;44m"

Thanks for your answer.
I have tried this slight modification, but it does not change anything on my
terminal.
--
TP (Tribulations Parallèles)

"Allez, Monsieur, allez, et la foi vous viendra." (D'Alembert).

[TP]

Peter Pearson wrote:

Thanks for your answer.

When you run echo, it recognizes the 4-character "esc" as a
convention for representing a single character, and performs
the re-interpretation for you. When you tell python
"print '\033[30;44m foo \033[0m'", python interprets
the "\033" as a single character.

So, the python print command *can* interpret these 4-character as a single
character. It would be odd if there were no possibility to do the same
thing when the characters are (i) stored in a python variable, or (ii) come
from the environment variables. Does anybody know any way to re-interpret a
string in Python? I have tried to play with "eval" (as in bash), but it
does not yield anything.

--
TP (Tribulations Parallèles)

"Allez, Monsieur, allez, et la foi vous viendra." (D'Alembert).

[Ben Finney]

TP <Tr**********@P aralleles.inval idwrites:

Peter Pearson wrote:

When you tell python "print '\033[30;44m foo \033[0m'", python
interprets the "\033" as a single character.

So, the python print command *can* interpret these 4-character as a
single character.

Not "interpret" , no.

It's more accurate to say that the Python *compiler* will translate
"\033" within a string literal into a single character in the compiled
executable byte code.

That is, the input to the Python interpreter will not have "\033" in
the string literal at all, but instead a single character produced by
the Python compiler at that point in the byte code.

It would be odd if there were no possibility to do the same thing
when the characters are (i) stored in a python variable, or (ii)
come from the environment variables.

Since those are not the input to the Python compiler, they can't be
translated this way.

--
\ “You know I could rent you out as a decoy for duck hunters?� |
`\ —Groucho Marx |
_o__) |
Ben Finney

[TP]

TP wrote:

So, the python print command *can* interpret these 4-character as a single
character. It would be odd if there were no possibility to do the same
thing when the characters are (i) stored in a python variable

Sorry, it works when using variables. Try for example:

col="[0;31m"
esc="\033"
colreset="[0m"
print esc + col + "foobar" + esc + colreset

--
TP (Tribulations Parallèles)

"Allez, Monsieur, allez, et la foi vous viendra." (D'Alembert).

[Peter Otten]

Peter Pearson wrote:

On Sun, 06 Jul 2008 23:42:26 +0200, TP <Tr**********@P aralleles.inval id>
wrote:

>>
$ python -c "print '\033[30;44m foo \033[0m'"

[writes an escape sequence to stdout]

>$ echo -e $esc$ColorBlack OnDarkblue foo $esc$ColorReset

[also writes an escape sequence to stdout]

>$ echo -n $esc$ColorBlack OnDarkblue foo $esc$ColorReset
\033[30;44m foo \033[0m

[snip, shuffle]

>$ export esc="\033"
$ export ColorBlackOnDar kblue="[30;44m"
$ export ColorReset="[0m"

import os
Color = os.environ['ColorBlackOnDa rkblue']
ColorReset = os.environ['ColorReset']
Esc = os.environ['esc']
print '%s%s%s%s%s' % (Esc, Color, " foo ", Esc, ColorReset)

[snip]

>$ python color.py
\033[30;44m foo \033[0m

The string "\033" is 4 characters long. Your shell variable
"esc" is 4 characters long. Your Python program prints
those four characters. You want it to re-interpret those 4
characters into a single escape character.

One of this group's regular participants can (I hope) tell
us three breathtakingly elegant ways to do that. I'm sorry
I can't.

When you run echo, it recognizes the 4-character "esc" as a
convention for representing a single character, and performs
the re-interpretation for you. When you tell python
"print '\033[30;44m foo \033[0m'", python interprets
the "\033" as a single character.

Peter Pearson's explanation is spot-on. You get the 4-character sequence
instead of the escape code chr(27).

$ export esc="\033"
$ python
Python 2.5.1 (r251:54863, Mar 7 2008, 03:39:23)
[GCC 4.1.3 20070929 (prerelease) (Ubuntu 4.1.2-16ubuntu2)] on linux2
Type "help", "copyright" , "credits" or "license" for more information.

>>import os
os.environ["esc"]

'\\033'

>>print _

\033

If you want to interpret "\\033" as "\033" you have to perform the
conversion explicitly. Fortunately there already is an encoding that
understands these escape sequences for characters:

>>esc = os.environ["esc"].decode("string-escape")
esc

'\x1b'

>>print "%s[30;44malles so schoen bunt hier%s[0m" % (esc, esc)

alles so schoen bunt hier

Peter

[TP]

Peter Otten wrote:

>>>esc = os.environ["esc"].decode("string-escape")
esc

'\x1b'

>>>print "%s[30;44malles so schoen bunt hier%s[0m" % (esc, esc)

alles so schoen bunt hier

Thanks a lot for your help. It works perfectly.
Indeed, one can read in the documentation concerning encodings:

"Produce a string that is suitable as string literal in Python source code"

--
TP (Tribulations Parallèles)

"Allez, Monsieur, allez, et la foi vous viendra." (D'Alembert).

[Peter Pearson]

On Mon, 07 Jul 2008 10:05:56 +0200, TP <Tr**********@P aralleles.inval idwrote:

TP wrote:

>So, the python print command *can* interpret these 4-character as a single
character. It would be odd if there were no possibility to do the same
thing when the characters are (i) stored in a python variable

Sorry, it works when using variables. Try for example:

col="[0;31m"
esc="\033"
colreset="[0m"
print esc + col + "foobar" + esc + colreset

I don't understand exactly what you mean by "Sorry", but
let me direct your attention to the fact that the "interpretation "
step about which we're talking happens during the parsing
of the string literal (i.e., while the statement esc="\033" is
processed), not during the execution of the print statement.
To prove this assertion, simply print len(esc). You will see
that esc is a single character long.

--
To email me, substitute nowhere->spamcop, invalid->net.

[TP]

Peter Pearson wrote:

I don't understand exactly what you mean by "Sorry"

I means: please forgive me for having said that it does not work with
variables, because it is completely false.

Thanks one more time

Julien

--
TP (Tribulations Parallèles)

"Allez, Monsieur, allez, et la foi vous viendra." (D'Alembert).