There might already be questions like this, but none of them answer my question. I have a script that loads a Python script in the directory and then displays a output with PHP:
<?php
$param1 = "first";
$param2 = "second";
$param3 = "third";
$command = "scripts/sg.py";
$command .= " $param1 $param2 $param3 2>&1";
header('Content-Type: text/html; charset=utf-8');
echo '<meta http-equiv="Content-Type" content="text/html; charset=UTF-8" />';
echo "<style type='text/css'>
</style>";
$pid = popen( $command,"r");
echo "<body><pre>";
while( !feof( $pid ) )
{
echo fread($pid, 256);
flush();
ob_flush();
echo "<script>window.scrollTo(0,99999);</script>";
usleep(100000);
}
pclose($pid);
echo "</pre><script>window.scrollTo(0,99999);</script>";
echo "<br /><br />Script finalizado<br /><br />";
?>
And is the Python code it should run, that is located in the directory:
#!/usr/bin/env python
# -*- coding: utf-8 -*-
# Script Python Example
import time
import sys
print "Initializing Python Script"
secret = 1337
guess = 0
count = 0
#Help the user give number range.
print 'Guess s number it`s between 0 <-> 2000 '
while guess != secret:
guess = input("Guess: ")
if guess < secret:
print "to small"
elif guess > secret:
print "to big"
count += 1
print 'You guessed the number in %s try' % count
The Python actually works! However it seems that Python's inputs don't work, they create a EOF error (end of file error).
Can anybody help me and suggest a way to create a Python interpreter that runs the Python file found in the directory. Just like, skuplt.org, but instead of running code a client users, it runs a Python file located in a directory, like stated above.
popen opens a pipe that is one-way only; you can read from it or write to it, but not both.
You want to use proc_open instead - see http://php.net/manual/en/function.proc-open.php
Related
I try to run a Python script as a background process from PHP on Windows using exec() on this way:
<?PHP
$python = 'C:\\Users\\User\\anaconda3\\python.exe';
$py_script = 'C:\\wamp\\www\\lab\\ex\\simple_test.py';
$py_stdout = '> temp\\'.session_id()."_std.txt";
$py_stderror = '2> temp\\'.session_id()."_stde.txt";
exec("$py_bg $python $py_script $py_stdout $py_stderror &");
The script called and worked correctly, but PHP still waiting for the script.
I removed the end & as I foundout it's only work on Linux and after searching other Q&A find this sulotion:
exec("start /B $py_bg $python $py_script $py_stdout $py_stderror");
But same result. How can I solve this problem?
=== UPDATE:
I used start /B in the wrong way, I changed my code to this:
<?PHP
$python = 'C:\\Users\\User\\anaconda3\\python.exe';
$py_script = 'C:\\wamp\\www\\lab\\ex\\simple_test.py';
$py_stdout = '> temp\\'.session_id()."_std.txt";
$py_stderror = '2> temp\\'.session_id()."_stde.txt";
$py_cmd = "$python $py_script $py_arg_1 $py_std $py_stde";
pclose(popen("start /B ". $py_cmd, "a"));
But now a Warning in PHP for popen():
Warning: popen(start /B ...,a): No error in C:\wamp\www\lab\start.php on line 50
and an other for pclose():
Warning: pclose() expects parameter 1 to be resource, bool given in ...
I checked PHP: popen - Manual and see there a is not a valid mode, but I see this on several answers around here!
however:
The mode. Either 'r' for reading, or 'w' for writing.
By changing mode to r, the script call and run in the background correctly and there is not an error or warning on PHP or Py.
<?PHP
$python = 'C:\\Users\\User\\anaconda3\\python.exe';
$py_script = 'C:\\wamp\\www\\lab\\ex\\simple_test.py';
$py_stdout = '> temp\\'.session_id()."_std.txt";
$py_stderror = '2> temp\\'.session_id()."_stde.txt";
$py_cmd = "$python $py_script $py_arg_1 $py_std $py_stde";
pclose(popen("start /B ". $py_cmd, "r"));
Background:
I'm in a position where I'm placing data into the command line and I need a php loop (what will become a server of sorts) to read STDIN and just echo what it reads to the shell its running in.
The following terrible code works when the process is running in the same shell as the content echoed:
<?php
echo getmypid();
$string = "/proc/" . getmypid() . "/fd/0";
while (true) {
fwrite(STDOUT, fgets(fopen($string, 'r'), 4096) . " worked\n");
}
?>
I've tried many variants:
<?php
echo getmypid();
$string = "/proc/" . getmypid() . "/fd/0";
while (true) {
$fo = fread(STDIN, 1024);
fwrite(STDOUT, $fo);
}
?>
The problem is that whenever I write to this loop from a separate terminal, the output appears in the other terminal but is not processed by the loop.
When I enter text in the same terminal, the text is echoed right back.
I need a way to get command line data into this loop from any source.
I used PHP to call python script successfully and got the result . But I have to wait for the end of script running without anything output. It looks not friendly to my customer.
How can I return the script results to the PHP web in realtime ?
For instance ,for code below , I want to the PHP web will show output message in realtime instead of show them together at the end . How can I change my code?
Thank you .
PHP Code:
<?php
$k = $_REQUEST['k'];
if (!empty($k))
{
$k = trim($k);
$a = array();
exec('python ./some.py '.$k, $a);
echo $a[0];
}
?>
Python Code:
#!/usr/bin/env python
#-*- coding:utf-8 -*-
import sys
def do_some(a):
print 'test1'
time.sleep(30)
print 'test2'
if __name__ == '__main__':
print 'Now the python scritp running'
time.sleep(20)
a = sys.argv[1]
if a:
print 'Now print something'
T = do_some(a)
By specification, exec stop the calling program until the end of the callee. After that, you get back the output in a variable.
If you want to send data as soon as they are produced, you should use popen. It will fork a new process, but will not block the caller. So you can perform other tasks, like looping to read the sub-process output line by line to send it to your client. Something like that:
$handle = popen("python ./some.py ", 'r');
while(!feof($handle)) {
$buffer = fgets($handle);
echo "$buffer<br/>\n";
ob_flush();
}
pclose($handle)
As a simple proof of concept, I tried to share a string between forked processes from node to node or from node to php.
Take this simple php code that should log the output of stdin according to the php docs:
echo 'test' | php -r 'echo trim(fgets(STDIN));'
Working fine, but when I'm spawning the process from nodejs:
spawner.js
var fs = require('fs'); var spawn = require('child_process').spawn;
//dummy stdin file
var stdin = fs.openSync('stdin_file', 'w+');
//write the string
fs.writeSync(stdin, 'test');
spawn('php', ['stdin_test.php'], {
cwd: __dirname,
detached: true,
//to fully detach the process nothing should be piped from or to the parent process
stdio: [stdin, fs.openSync('out.log', 'a'), fs.openSync('err.log', 'a')]
})
stdin_test.php
<?php
error_log('php://stdin');
//this should log 'test' but outputs a newline
error_log(trim(fgets(STDIN)));
$t = fopen('/dev/stdin', 'r');
error_log('/dev/stdin:');
//this is working as expected
error_log(trim(fgets($t)));
Why is php://stdin empty? Is it safe to use /dev/stdin? What is the difference between /dev/stdin and php://stdin anyway?
Note that I have this behavior between 2 node processes too: process.stdin is empty but /dev/stdin has the expected result.
Gist available here
stdin man reference
I tested with the following script ( stdin_test.php ) using:
> echo test | php stdin_test.php
stdin_test.php
<?
echo 'STDIN :' ;
echo trim(fgets(STDIN)) ;
echo PHP_EOL;
$stdin_stream = fopen('php://stdin', 'r');
echo 'php://stdin :';
echo trim(fgets($stdin_stream));
echo PHP_EOL;
fclose($stdin_stream);
$stdin_file = fopen('/dev/stdin', 'r');
echo '/dev/stdin :';
echo trim(fgets($stdin_file));
echo PHP_EOL;
fclose($stdin_file);
I get back :
STDIN :test
php://stdin :
/dev/stdin :
If I then comment out the line:
//echo trim(fgets(STDIN));
I get back:
STDIN :
php://stdin :test
/dev/stdin :
If I comment out both of the first stdin echoes (and the file handler pointers), I get:
STDIN :
php://stdin :
/dev/stdin : test
Looking at documentation on php://input and how it is one-time usable unless (after 5.6) "the request body is saved" which is typical for POST requests but not PUT requests (apparently). This has me thinking that they are called "streams" because you get to walk in them once.
Rewind your stdin stream in JS before spawning PHP, else the file pointer will sit at the end of what you just wrote.
I have a problem running another file from php. I want my php params to be the output of running a python file that calls another file itself.
Here is my php file:
<?php
if (isset($_POST['submit'])) {
$params = solve();
}
function solve() {
exec("python array.py", $output);
return $output;
}
?>
If array.py is simply:
if __name__ == "__main__":
print 1
print 2
print 3
print 4
I will get 1,2,3,4 for my output, but I as soon as I change array.py to the following file that calls os.system, I don't get anything. So the new array.py is:
import os
def main():
os.system("python test.py") #test.py creates tmp.txt with 4 lines w/ values 1,2,3,4
def output():
f = open("tmp.txt", "r")
myReturn = []
currentline = f.readline()
while currentline:
val = currentline[:-1] #Getting rid of '\n'
val = int(val)
myReturn = myReturn + [val]
currentline = f.readline()
f.close()
return myReturn
if __name__ == "__main__":
main()
o = output()
print o[0]
print o[1]
print o[2]
print o[3]
Also if I just run test.py, the output is the file tmp.txt:
1
2
3
4
So now, when I run my php file, the output tmp.txt is not even created in the directory and as a result I don't get any output from my php either.
I am not sure why this is happening because when I just run array.py myself, I get the desired output, and the tmp file is created.
EDIT:
I forgot to include: import os above.
Change exec to:
exec("python array.py 2>&1", $output)
Or check the web server or php error log. This will return the error output from the python script to your php script (not normally what you want in production).