WindowsError: [Error 193] %1 is not a valid Win32 application

[alsudais]

Hello,

I keep getting this error message. Any help is much appreciated. The "runTagger.sh" is in the same folder..

Attached is a screen capture of the issue.

Kareem

[ianozsvald]

I'm not on Windows so it'll be tricky to diagnose. I'm guessing that the first thing to fix is line 71 (the subprocess.Popen line in check_script_is_present) - without fixing this we won't fix the second problem (line 47). Can you first try the following at the command line (not in Python, just purely at the Cmd prompt):

./runTagger.sh --help

If this runs then we can guess that Python is at fault. If this fails then the problem is how the script is run (separate to Python).

[ianozsvald]

Depending on what we learn it might also be reading up on the Windows notes for subprocess:
http://docs.python.org/2/library/subprocess.html#converting-an-argument-sequence-to-a-string-on-windows

[alsudais]

You're right, windows doesn't know how to run this runTagger.sh file. When I tried running ./runTagger.sh --help, I got an alert message asking me to specify a program that can open it.

Do you think I shouldn't work on changing lines 71 and 47?

[ianozsvald]

.sh files are for Linux/Mac (which uses BSD under the hood). Windows won't run a .sh file by default (unless perhaps you have Cygwin installed). However the runTagger.sh file is very simple. Let's try to debug it. Do not edit the Python files, the problem isn't there (at least not yet). Let's understand the shell script first.

runTagger.sh has two commands. The first is a set which does some sanity checks, you don't need that. The second line runs Java and the "$@" takes any command line arguments (e.g. --help or the --output-format that I use) and substitutes them in place of "$@".

First I need to know what works on Windows. Can you try running exactly the following command at the command line? If it works then we can fix the Python wrapper to avoid using the runTagger.sh script at all (which I think is the easier fix). Try this and send me the result:

java -XX:ParallelGCThreads=2 -Xmx500m -jar ark-tweet-nlp-0.3.2.jar --help

You should see some output starting:

RunTagger [options] [ExamplesFilename]
  runs the CMU ARK Twitter tagger on tweets from ExamplesFilename, 
  writing taggings to standard output. Listens on stdin if no input filename.

Options:
  --model <Filename>        Specify model filename. (Else use built-in.)
...<more stuff>

DO YOU SEE THE ABOVE OUTPUT?

If you try this and it complains that it can't find java, just try Java by self:

java -version
#and you should get something like:
java version "1.7.0_17"
Java(TM) SE Runtime Environment (build 1.7.0_17-b02)
Java HotSpot(TM) 64-Bit Server VM (build 23.7-b01, mixed mode)

If you don't get the above (and you still get a Java error) then you'll have to get Java running on your machine before we can go any further. I'm assuming you have a working Java that you can call from the command line.

[alsudais]

Yep, I do see that output after running the first command.

RunTagger [options] [ExamplesFilename]
runs the CMU ARK Twitter tagger on tweets from ExamplesFilename,
writing taggings to standard output. Listens on stdin if no input filename.
Options:
--model Specify model filename. (Else use built-in.)
--just-tokenize Only run the tokenizer; no POS tags.

...

I'm not sure if it matters, but here is the java version I'm using. It's seems it's a bit older than the one you're using.

java -version
java version "1.6.0_24"
Java(TM) SE Runtime Environment (build 1.6.0_24-b07)
Java HotSpot(TM) 64-Bit Server VM (build 19.1-b02, mixed mode)

[ianozsvald]

Good, we can fix the problem. I probably can't work on this until tonight
(I'm at work) but I'll try to change the way the script is called this
evening. Feel free to hack with the Popen if you feel you know what you're
doing? Regarding your java version - that shouldn't matter.

On 4 April 2013 10:43, kareem180 notifications@github.com wrote:

Yep, I do see that output after running the first command.

RunTagger [options] [ExamplesFilename]
runs the CMU ARK Twitter tagger on tweets from ExamplesFilename,
writing taggings to standard output. Listens on stdin if no input filename.

Options:
--model Specify model filename. (Else use built-in.)
--just-tokenize Only run the tokenizer; no POS tags.
...

I'm not sure if it matters, but here is the java version I'm using. It's
seems it's a bit older than the one you're using.

java -version
java version "1.6.0_24"
Java(TM) SE Runtime Environment (build 1.6.0_24-b07)
Java HotSpot(TM) 64-Bit Server VM (build 19.1-b02, mixed mode)

—
Reply to this email directly or view it on GitHubhttps://github.com//issues/2#issuecomment-15888276
.

Ian Ozsvald (A.I. researcher)
ian@IanOzsvald.com

http://IanOzsvald.com
http://MorConsulting.com/
http://Annotate.IO http://annotate.io
http://SocialTiesApp.com/
http://TheScreencastingHandbook.com
http://FivePoundApp.com/
http://twitter.com/IanOzsvald
http://ShowMeDo.com

[ianozsvald]

Try this version. I've taken runTagger.sh out and used a directly call to java instead. Try running it directly from the parent directory (where ark-tweeet-nlp-0.3.2.jar is stored):

$ ls
ark-tweet-nlp-0.3.2.jar  ark-tweet-nlp-python  data  docs  dummy.txt  env  examples  LICENSE.txt  README.txt  runTagger.sh  scripts  twokenize.sh
ian@ian-Latitude-E6420 ~/workspace/cmu_pos_tweet_tagger/0.3.2/ark-tweet-nlp-0.3.2 $ python ark-tweet-nlp-python/CMUTweetTagger.py
Checking that we can see "java -XX:ParallelGCThreads=2 -Xmx500m -jar ark-tweet-nlp-0.3.2.jar", this will crash if we can't
Success.
Now pass in two messages, get a list of tuples back:
[[('this', 'O', 0.9526), ('is', 'V', 0.9988), ('a', 'D', 0.9981), ('message', 'N', 0.9985)], [('and', '&', 0.9996), ('a', 'D', 0.9958), ('second', 'A', 0.9428), ('message', 'N', 0.9986)]]

[alsudais]

I guess we can't see "java -XX:ParallelGCThreads=2 -Xmx500m -jar ark-tweet-nlp-0.3.2.jar". Both "CMUTweetTagger.py" and "ark-tweeet-nlp-0.3.2.jar" are in the same directory.

Checking that we can see "java -XX:ParallelGCThreads=2 -Xmx500m -jar ark-tweet-nlp-0.3.2.jar", this will crash if we can't
Traceback (most recent call last):
File "C:/Python27/ark-tweet-nlp-python-master/CMUTweetTagger.py", line 94, in
success = check_script_is_present()
File "C:/Python27/ark-tweet-nlp-python-master/CMUTweetTagger.py", line 85, in check_script_is_present
assert "RunTagger [options]" in lines[0]
IndexError: list index out of range

[ianozsvald]

Can you do "dir" and paste the result here? Let me double check what we
have.
On 5 Apr 2013 18:13, "kareem180" notifications@github.com wrote:

I guess we can't see "java -XX:ParallelGCThreads=2 -Xmx500m -jar
ark-tweet-nlp-0.3.2.jar". Both "CMUTweetTagger.py" and
"ark-tweeet-nlp-0.3.2.jar" are in the same directory.

Checking that we can see "java -XX:ParallelGCThreads=2 -Xmx500m -jar
ark-tweet-nlp-0.3.2.jar", this will crash if we can't

Traceback (most recent call last):
File "C:/Python27/ark-tweet-nlp-python-master/CMUTweetTagger.py", line 94,
in
success = check_script_is_present()
File "C:/Python27/ark-tweet-nlp-python-master/CMUTweetTagger.py", line 85,
in check_script_is_present
assert "RunTagger [options]" in lines[0]
IndexError: list index out of range

—
Reply to this email directly or view it on GitHubhttps://github.com//issues/2#issuecomment-15968569
.

[alsudais]

Here you go:

Directory of C:\Python27\ark-tweet-nlp-python-master
04/05/2013 10:06 AM

.
04/05/2013 10:06 AM ..
03/28/2013 06:56 AM 303 .gitignore
10/28/2012 09:08 AM 15,257,377 ark-tweet-nlp-0.3.2.jar
04/05/2013 10:09 AM 4,594 CMUTweetTagger.py
04/01/2013 02:54 PM 3,362 CMUTweetTagger.pyc
04/01/2013 03:01 PM data
04/01/2013 03:01 PM docs
04/01/2013 03:01 PM examples
10/28/2012 09:08 AM 2,086 LICENSE.txt
04/05/2013 10:06 AM 29,016 README.md
10/28/2012 09:08 AM 1,735 README.txt
10/28/2012 09:08 AM 141 runTagger.sh
04/01/2013 03:01 PM scripts
04/05/2013 10:09 AM 4,594 test.py
10/28/2012 09:08 AM 150 twokenize.sh
03/28/2013 06:56 AM 0 init.py
11 File(s) 15,303,358 bytes

[ianozsvald]

Since the jar file and the python file are in the same directory, I'd expect it to run. It runs as a command from the command line so there's just something wrong in how Python is calling out. I can't help you for a while I'm afraid - I get married in 2 weeks and then leave the country for 4 (without laptop), everything is going on hold until June.

If you figure out a patch for this issue (read up on Popen and Subprocess - particularly the Windows notes which are different to the Linux/Mac notes) I'd be happy to integrate it. I'd suggest starting with the command line that works (e.g. start with the --help version, then go to the version that passes in a file), then do a one-liner in Python at the console to make the same thing. Once you've figured out how to make that work then work on CMUTweetTagger.py to integrate the changes. Don't worry about Linux compatibility, we can figure that out later when my life goes back to normal.

You might also want to try asking clear, simple questions in the python-tutor mailing list, they can probably help with Windows Popen/Subprocess examples (just ask about the one-liner examples, not CMUTweetTagger.py).

Sorry I can't be of more help. Good luck!

[alsudais]

Thanks for your help, and good luck with the marriage and stuff. I'll try working on this and I'll let you know if I end up figuring out a way to resolve the issue.

Thanks again!

[tjflexic]

@kareem180 I find the same question. How to solve it? Thank you.

[alsudais]

We were not able to solve it.

[tjflexic]

@kareem180 Thank you. I wrote a new python wrapper with subprocess.call(subprocess.call(run_tagger_cmd + ' tmp.txt > tmp_tagger.txt', shell=True), then read the results from tmp_tagger.txt.