#Tutorial: Run Scripts on Penn's server, but display Spyder on your own computer!

![Demo screenshot](https://ww1.sinaimg.cn/large/006tKfTcgy1flevabqykkj31kw0zkqqj.jpg)

This Spyder instance actually runs a biglab.seas.upenn.edu. It is "tunneled" to my Macbook, so that I can utilize the computational power of BigLab, the friendly IDE of Spyder, while working away from home in my pajamas on this 37°F cold day!

1. If you are on a Mac, **download and install [XQuartz](https://www.xquartz.org/)**, which enables your Mac to accept remote windows. For other Operating Systems, please see [this](https://www.seas.upenn.edu/cets/answers/x11-forwarding.html).

2. **Add the following lines** to the `~/.zshrc`:

   ```shell
   # This line enables you to type "biglab" and connect to biglab.seas.upenn.edu directly:
   alias biglab="ssh -tY <pennkey>@biglab.seas.upenn.edu 'screen -S foo -rd || screen -S foo zsh'"

   # This script allows iTerm to accept X11 windows from remote servers. Copied from <http://hints.macworld.com/article.php?story=2004082505230779>.
   DISPLAY=""
   for x in 0 1 2 3 4 5 6 7 8 9
   do
     if [ -O /tmp/.X$x-lock ]
         then
         DISPLAY=:$x.0
         break
     fi
   done
   if [ -z "$DISPLAY" ]
       then
       echo "$USER has no X11 DISPLAY open" 1>&2
   fi
   export DISPLAY
   ```

3. **Start XQuartz**, and set it aside. This will be our "client program" accepting windows running on UPenn's servers.

4. Open a terminal, and **connect to biglab.seas.upenn.edu** by typing `biglab`. 

5. When connected, **install Anaconda**:

   ```Shell
   wget https://repo.continuum.io/archive/Anaconda2-5.0.1-Linux-x86_64.sh
   bash Anaconda2-5.0.1-Linux-x86_64.sh
   ```

6. When finished, you can **set up virtual environments** by following instructions from Homework 4, Part II, 2.3.

7. With a virtual environment activated or not, you can **start Spyder** now: `spyder`. After a while, you can see a Spyder instance showing up on your desktop. Keep in mind that this is actually running on biglab!

## Known Bugs

1. Tensorflow will always throw an error upon finishing, due to its "unqiue" way of quitting a program.
2. Button symbols are replaced by weird Korean characters, generating lots of `Failed to compute left/right minimum bearings for "FontAwesome"` error from the terminal.

Neither of these bugs affect the proper execution of our programs, so I'd let them go for now.