From Zero to Binder in Python!
Sarah Gibson, The Alan Turing Institute
The Turing Way - making reproducible Data Science "too easy not to do"
To follow these instructions on your own machine, follow this link: http://bit.ly/zero-to-binder-python
Binder can take a long time to load, but this doesn't necessarily mean that your Binder will fail to launch. You can always refresh the window if you see the "... is taking longer to load, hang tight!" message.
- Running Code is more complicated than Displaying Code
- What Binder Provides
- 1. Creating a repo to Binderize
- 2. Launch your first repo!
- 3. Run
- 4. Pinning Dependencies
- 5. Check the Environment
- 6. Sharing your Work
- 7. Accessing data in your Binder
- 8. Get data with
- Beyond Notebooks...
- Now over to you!
Running Code is more complicated than Displaying Code
GitHub is a great service for sharing code, but the contents are static.
How could you run a GitHub repository without installing complicated requirements? Or even in your browser?
To run code, you need:
- Hardware on which to run the code
- Software, including:
- The code itself
- The programming language (e.g. Python, R, Julia, and so on)
- Relevant packages (e.g. pandas, matplotlib, tidyverse, ggplot)
What Binder Provides
Binder is a service that provides your code and the hardware and software to execute it.
You can create a link to a live, interactive version of your code!
- An example binder link:
https://mybinder.org/v2/gh/trekhleb/homemade-machine-learning/master?filepath=notebooks%2Fanomaly_detection%2Fanomaly_detection_gaussian_demo.ipynb From this repo: https://github.com/trekhleb/homemade-machine-learning
- Notice that the Binder link has a similar structure to the GitHub repo link (
- The "filepath" argument opens a specific notebook (
.ipynbfile) in the repo
- Notice that the Binder link has a similar structure to the GitHub repo link (
1. Creating a repo to Binderize
- Create a new repo on GitHub called "my-first-binder".
- Make sure the repository is public, not private!
- Don't forget to initialise with a README!
- Create a file called
hello.pyvia the web interface with
print("Hello from Binder!")on the first line and commit to the
Why did the repo have to be public?
<mybinder.org> cannot access private repositories as this would require a secret token. The Binder team choose not to take on the responsibility of handling secret tokens as <mybinder.org> is a public service and proof of technological concept. If accessing private repositories is a feature you/your team need, we advise that you look into building your own BinderHub.
2. Launch your first repo!
- Go to https://mybinder.org
- Type the URL of your repo into the "GitHub repo or URL" box.
It should look like this:
- As you type, the webpage generates a link in the "Copy the URL below..." box
It should look like this:
- Copy it, open a new browser tab and visit that URL.
- You will see a "spinner" as Binder launches the repo.
If everything ran smoothly, you'll see a Jupyter Notebook interface.
What's happening in the background? - Part 1
While you wait, BinderHub (the backend of Binder) is:
- Fetching your repo from GitHub
- Analysing the contents
- Creating a Docker file based on your repo
- Launching that Docker image in the Cloud
- Connecting you to it via your browser
- In the top right corner, click "New"
- In the new tab with the terminal, type
python hello.pyand press return
Hello from Binder! should be printed to the terminal.
4. Pinning Dependencies
It was easy to get started, but our environment is barebones - let's add a dependency!
- In your repo, create a file called
- Add a line that says:
- Check for typos! Then commit to the
- Visit https://mybinder.org/v2/gh/your-username/my-first-binder/HEAD again in a new tab
This time, click on "Build Logs" in the big, horizontal, grey bar. This will let you watch the progress of your build. It's useful when your build fails or something you think should be installed is missing.
N.B.: Sometimes Binder's build logs prints things in red font, such as warnings that
pip is not up-to-date (
pip is often out of date because it's regularly updated!) or installation messages, especially if you're using R.
These red messages don't necessarily mean there's a problem with your build and it will fail - it's just an unfortunate font colour choice!
What's happening in the background? - Part 2
This time, BinderHub will read
requirements.txt and install version
1.14.5 of the
More on pinning dependencies
In the above example, we used two equals signs (
==) to pin the version of
This tells Binder to install that specific version.
Another way to pin a version number is to use the greater than or equal to sign (
>=) to allow any version above a particular one to be installed.
This is useful when you have a lot of dependencies that may have dependencies on each other and allows Binder to find a configuration of your dependencies that do not conflict with one another whilst avoiding any earlier versions which may break or change your code.
Finally, you could not provide a version number at all (just the name of the library/package) and Binder will install the latest version of that package.
N.B.: These operations to pin dependencies are most likely specific to Python.
Each language has it's own quirks and a link to the different types of configuration files (which is what
requirements.txt is) is given at the bottom of this document.
5. Check the Environment
In the top right corner, click "New"
➡️"Python 3" to open a new notebook
Type the following into a new cell:
import numpy print(numpy.__version__) numpy.random.randn()
Note the two underscores either side of
Run the cell to see the version number and a random number printed out
- Press either SHIFT+RETURN or the "Run" button in the Menu bar
N.B.: If you save this notebook, it will not be saved to the GitHub repo. Pushing changes back to the GitHub repo through the container is not possible with Binder. Any changes you have made to files inside the Binder will be lost once you close the browser window.
6. Sharing your Work
Binder is all about sharing your work easily and there are two ways to do it:
- Share the https://mybinder.org/v2/gh/your-username/my-first-binder/HEAD URL directly
- Visit https://mybinder.org, type in the URL of your repo and copy the Markdown or ReStructured Text snippet into your
README.mdfile. This snippet will render a badge that people can click, which looks like this:
- Add the Markdown snippet from https://mybinder.org to the
README.mdfile in your repo
- The grey bar displaying a binder badge will unfold to reveal the snippets. Click the clipboard icon next to the box marked with "m" to automatically copy the Markdown snippet.
- Click the badge to make sure it works!
7. Accessing data in your Binder
Another kind of dependency for projects is data. There are different ways to make data available in your Binder depending on the size of your data and your preferences for sharing it.
Small public files
The simplest approach for small, public data files is to add them directly into your GitHub repository. They are then directly encapsulated into the environment and versioned along with your code.
This is ideal for files up to 10MB.
Medium public files
To access medium files from a few 10s MB up to a few hundred MB, you can add a file called
postBuild to your repo.
postBuild file is a shell script that is executed as part of the image construction and is only executed once when a new image is built, not every time the Binder is launched.
postBuild example for more uses of the
N.B.: New images are only built when Binder sees a new commit, not every time you click the Binder link. Therefore, the data is only downloaded once when the Docker image is built, not every time the Binder is launched.
Large public files
It is not practical to place large files in your GitHub repo or include them directly in the image that Binder builds. The best option for large files is to use a library specific to the data format to stream the data as you're using it or to download it on demand as part of your code.
For security reasons, the outgoing traffic of your Binder is restricted to HTTP or GitHub connections only. You will not be able to use FTP sites to fetch data on mybinder.org.
There is no way to access files which are not public from mybinder.org. You should consider all information in your Binder as public, meaning that:
- there should be no passwords, tokens, keys etc in your GitHub repo;
- you should not type passwords into a Binder running on mybinder.org;
- you should not upload your private SSH key or API token to a running Binder.
In order to support access to private files, you would need to create a local deployment of BinderHub where you can decide the security trade-offs yourselves.
N.B.: Building a BinderHub is not a simple task and is usually taken on by IT/RSE groups for reasons around managing maintenance, security and governance. However, that is not to say that they are the only groups of people who should/could build a BinderHub.
8. Get data with
- Go to your GitHub repo and create a file called
postBuild, add a single line reading:
wget -q -O gapminder.csv http://bit.ly/2uh4s3g
wgetis a program which retrieves content from web servers. This line extracts the content from the bitly URL and saves it to the filename denoted by the
-Oflag (capital "O", not zero), i.e.
wgetto do this quietly, i.e. don't print anything to the console.
- Update your
requirements.txtfile by adding a new line with
pandason it and another new line with
- These packages aren't necessary to download the data but we will use them to read the CSV file and make a plot
- Click the binder badge in your README to launch your Binder
Once the Binder has launched, you should see a new file has appeared that was not part of your repo when you clicked the badge.
Now visualise the data by creating a new notebook ("New"
%matplotlib inline import pandas data = pandas.read_csv("gapminder.csv", index_col="country") years = data.columns.str.strip("gdpPercap_") # Extract year from last 4 characters of each column name data.columns = years.astype(int) # Convert year values to integers, saving results back to dataframe data.loc["Australia"].plot()
See this Software Carpentry lesson for more info.
JupyterLab is installed into your containerized repo by default. You can access the environment by changing the URL you visit to:
N.B.: We've already seen how the
?filepath= argument can link to a specific file in the "What Binder Provides" section at the beginning of this workshop.
Here you can access:
- IPython consoles
- A text editor
If you use R, you can also open RStudio using
Now over to you!
Now you've binderized (bound?) this demo repo, it's time to binderize the example script and data you brought along!
Some useful links:
- Choosing languages:
- Configuration files:
- Example Binder repos:
- Getting data: