If this is your first time here, welcome to Occam!
In this guide, you will learn the basics of creating new things in Occam.
This guide assumes you are learning how to use Occam. If you would like to install Occam yourself, then go to the installation guide instead.
Occam is an archival tool focusing on the repeatability of software. That is, being able to build and run software once and then, in the future, run it again as close to the same way as possible. It does this by keeping track of software and its dependencies and storing them in a highly available way.
This site you are currently on serves as a portal to creating and using software that has been made using the Occam platform. You, too, will be able to create or wrap existing software using Occam and share that for others to use.
This guide serves as a basic tutorial for creating a script in Occam, which is only one small way that Occam can be used. It will step through the process in a very matter-of-fact way, but you are free to explore on your own. Throughout the site, you will see icons that will pop up a helpful message about that area of the site when you invoke them.
The first thing you will want to do is sign up for your own account. You can do many things as a guest (without logging on), but you will be unable to create and save new objects.
To create a new account, click and fill in the fields with your desired username and password. These will be used each time you wish to log on to this particular Occam server and revisit your work.
You will be logged in and presented with your dashboard page. Navigate to the Collection page to see a list of your ongoing work. Naturally, this will be empty for a new account. So we will create a simple script to start with.
First, we will create a small script. Don't worry, it won't require much programming. We will use a template so that we get some of the more repetitive basic steps out of the way more quickly.
At the top of the collection page, use the button to bring up a dialog to create a new item in your collection.
Here, we could create the default item, which is an empty object or experiment (workflow), however we will start with a Python script using a template. Click on the template field and start typing "Python" and select Python Script from the dropdown.
Once you have that, it will automatically fill out the object type as "script" and you will then be able to enter in a name for your script. You can be imaginative, or just type "My Python Script" here.
Once you are done, press the button and it will navigate to your new script.
The initial template will have a simple Python script already provided that will do something simple. To test everything and become more familiar with the interface, let's just run that simple script as it is.
Navigate to the Run page using the tab bar to see the options we have to execute our script. We will simply press .
This will create a manifest for running your script, generate a job that will be queued, and then eventually invoke your script and show a live view of the running task. When it is done, the output should perform the simple script and show you:
Now, this is not very interesting, of course. So, let's modify the file.
Go to the Files page and you will see the list of files
that comprise your object. For our template, this is just a single
main.py script. You can open a file by navigating to that
file in the listing and clicking on it. This will open the file in a viewer and show you the following text:
So, let's add some code and see what happens. Let's update the file to the following (you can copy and paste from the section below:)
import numpy as np a = np.arange(15).reshape(3, 5) print(a)
This will use numpy to create a matrix and then print that matrix to the screen. The dimensions of the matrix are 3 by 5 and the values are supplied as an integer range from 1 to 15.
Make sure you save the file. Afterward, go back to the Run tab and press the button again.
This time, you will see the following output:
Traceback (most recent call last): File "/occam/QmeLUK8o7nXTa5A7JkR4odjBjLf8pu3hjQvjS7wephBPcE-5dtJT2WQVh6Jyx3UscpqQyeeBkAxms/main.py", line 2, in <module> import numpy as np ModuleNotFoundError: No module named 'numpy'
This means that numpy was not specified as a dependency to our script. By default, not every python library in the world is automatically available to your script. You may have a need for a specific version of a specific library, after all.
To faciliate this, we have to specify that we need numpy.
Navigate to the Details page to see the metadata for your script.
There will be a section called
Dependencies which lists each object that must be included
when our script is invoked.
There is a button at the end of the list that will allow you to append a new dependency.
Invoking this button will bring up a dialog that will allow you to search for and then add such an object.
In the query field of the dialog, type
numpy. At least two objects should appear of the type
numpy‑2 is for Python 2.x, which is an older version of Python than we are using. We want the
Generally, you will want to include the python libraries without the '-2' suffix.
numpy and then click the button. It should show you a new object in the list with the name "numpy."
Now simply navigate once more to the Run page and press that button a final time. You will, if all goes well, see the following:
[[ 0 1 2 3 4] [ 5 6 7 8 9] [10 11 12 13 14]]
With this output, you now have a starting point for writing scientific Python. Enjoy!