doc/user/tutorials/running-external-program.textile

   1 ---
   2 layout: default
   3 navsection: userguide
   4 navmenu: Tutorials
   5 title: "Running external programs"
   6 navorder: 18
   7 ---
   8
   9 h1. Running external programs
  10
  11 This tutorial demonstrates how to use Crunch to run an external program by writting a wrapper using the Python SDK.
  12
  13 *This tutorial assumes that you are "logged into an Arvados VM instance":{{site.basedoc}}/user/getting_started/ssh-access.html#login, and have a "working environment.":{{site.basedoc}}/user/getting_started/check-environment.html*
  14
  15 Start by entering the @crunch_scripts@ directory of your git repository:
  16
  17 <notextile>
  18 <pre><code>$ <span class="userinput">cd you/crunch_scripts</span>
  19 </code></pre>
  20 </notextile>
  21
  22 Next, using your favorite text editor, create a new file called @run-md5sum.py@ in the @crunch_scripts@ directory.  Add the following code to compute the md5 hash of each file in a collection:
  23
  24 <pre><code class="userinput">{% include run-md5sum.py %}</code></pre>
  25
  26 Make the file executable:
  27
  28 notextile. <pre><code>$ <span class="userinput">chmod +x run-md5sum.py</span></code></pre>
  29
  30 Next, add the file to @git@ staging, commit and push:
  31
  32 <notextile>
  33 <pre><code>$ <span class="userinput">git add run-md5sum.py</span>
  34 $ <span class="userinput">git commit -m"run external md5sum program"</span>
  35 $ <span class="userinput">git push origin master</span>
  36 </code></pre>
  37 </notextile>
  38
  39 You should now be able to run your new script using Crunch, with "script" referring to our new "run-md5sum.py" script.
  40
  41 <notextile>
  42 <pre><code>$ <span class="userinput">cat &gt;the_job &lt;&lt;EOF
  43 {
  44  "script": "run-md5sum.py",
  45  "script_version": "you:master",
  46  "script_parameters":
  47  {
  48   "input": "c1bad4b39ca5a924e481008009d94e32+210"
  49  }
  50 }
  51 EOF</span>
  52 $ <span class="userinput">arv -h job create --job "$(cat the_job)"</span>
  53 {
  54  ...
  55  "uuid":"qr1hi-xxxxx-xxxxxxxxxxxxxxx"
  56  ...
  57 }
  58 $ <span class="userinput">arv -h job get --uuid qr1hi-xxxxx-xxxxxxxxxxxxxxx</span>
  59 {
  60  ...
  61  "output":"4d164b1658c261b9afc6b479130016a3+54",
  62  ...
  63 }
  64 </code></pre>
  65 </notextile>