Linux command line using csv file [sentiment analysis]

snake0

Hello all,

I am trying to run a process built with the Sentiment Analysis wizard via command line and cannot find the way to do it.

The wizard built a bunch of subdirectories but I managed to organize them into the following:
repositories/Analyze
    Analyze.properties
    process.properties
    process.rmp
    input.ioo, input,md, input.properties

I also ran this with the GUI and it works, so I know these are the only files required. What I actually want to do is use the process with a new csv file each time in a script, so I presume I get rid of the three input.* files.

I found the command line tool and for a start tried to just run it with the 3 input files that were already compiled:

---

scripts/rapidminer-batch.sh -f /var/www/rapidminer-studio/repositories/Analyze/process.rmp

---

However this process fails.

process.rmp contains the following at the top of the file:

---

    <input>
      <location>input</location>
    </input>
    <output/>
    <macros>
      <macro>
        <key>label</key>
        <value>Sentiment</value>
      </macro>
      <macro>
        <key>label_positive_class</key>
        <value>positive</value>
      </macro>
      <macro>
        <key>label_negative_class</key>
        <value>negative</value>
      </macro>
    </macros>

---

So I should probably be getting rid of the 3 input files and the <input><location>input</location></input> part in process.rmp. The macros denote the format of my csv file originally used, so presumably as long as I use another csv file with the same format it should work.

Only problem is, I have no idea how to do this via the command line. Could anyone help?