- Jupyter
Jupyter Notebook is an open-source web application that allows you to create and share documents that contain live code, equations, visualizations and narrative text. In our course, Jupyter notebook will be our daily tool to write, test, and sharing our codes and works. It's very useful for us to learn and make reproducible works. The good advantage of Jupyter notebook includes:
- The Notebook has support for over 40 programming languages, including Python, R, Julia, and Scala.
- Notebooks can be shared with others using email, Dropbox, GitHub and the Jupyter Notebook Viewer.
- Your code can produce rich, interactive output: HTML, images, videos, LaTeX, and custom MIME types.
It is suggested to enter virtual environment before using Jupyter notebook. Because in the following study, we may need to install some packages and modules in Jupyter notebook. It's necessary to keep those files in the same path, so Jupyter can source the modules when you want to use.
The following are the usual path to setup jupyter environment. For users in CVA 517 LAB, please see here.
pyvenv venvsource venv/bin/activateWhen you see (venv) appear in front of your command line prompt, that means the you are in the virtual environment. Always check this prefix to make sure you are working in the right environment.
You can use deactivate command to exit current virtual environment.
Now you can install and enter Jupyter notebook.
Note: you only need to do this once for every virtual environment
pip3 install jupyterAlways run your Jupyter notebook in the virtual environment.
jupyter notebookBy default the notebook will be available at http://localhost:8888/tree
So, next time when you use your jupyter notebook, you just need to type following commands in your terminal
source venv/bin/activate
jupyter notebook$ control+CPay attention to the text. Then you will get the following picture. Please input y in 5 seconds.
Due to the jupyter and the python conflict, there are problems of installing jupyter by the usual way. Instead, the following will work. For more details explanation, please see here.
pyvenv venv
source venv/bin/activate
pip install --upgrade pip
pip3 install jupyter
pip3 install 'ipython ==6.5.0'
pip3 install 'prompt-toolkit ==1.0.15'You can use NBViewer to check out a Jupyter notebook hosted on GitHub. You only need to input the GitHub URL link into the input box.
Sometimes, you change codes after first preview on NBViewer. You find the codes are updated on GitHub. However, NBViewer still shows the outdated codes used before. This is a common "cache" problem. You can try to trigger a cache invalidation by adding query string in the URL.
Suppose this is your original NBviewer link:
https://nbviewer.jupyter.org/xxxxx/yyyyy.ipynb
You can add anything after the ? mark:
https://nbviewer.jupyter.org/xxxxx/yyyyy.ipynb?fffffff
The fffffff here can be arbitrary. Note, in your submission of the work, if the an NBviewer link is required, please paste the one that shows exactly the result you want, i.e. the one with some taililng querystring with which you can see the latest result. Or else, we may see the old content.
- click new to create a new python 3 notebook
- write codes like you usually do in text editors, and press
shift+returnto run the code. It will return the results or errors under the cell. - use
! pip3 install module_nameto install modules in jupyter notebook. - in front of every cell, there is an in [ ] sign, the number in [] means the sequences of cells, and if there is * in [], means that this cell is still running, you can either wait it finish or click
stopunder the kernel to exit from the running, pressingItwice will also do the trick. - cell. run cell run step by step. run all above to run and check the previous steps of coding.
- kernel. kernel is a tool for interactive input and output all the things you did from the beginning. By clicking
restart, you can give a variable another value. type()more to check the object. It is very useful when we write complicated codings. Eg:a = 1, type(a)help()to know the details. Eg: help(str.strip)printstep by step to check where the error is. (In Jupyter, you can just input the variables without the function of print.) Like, type data in cell equals to print(data).
- Rerun all the cells
- Restart the kernel
- Clear the output cells
You do not want to install and re-install dependencies every time. It is more convenient to setup a virtual environment and install common dependencies/ modules for data analysis.
You can download this requirement.txt and then run the following command (inside virtual environment)
pip install -r requirements.txtplease see here.
Different libraries to save .html locally.
#plotly
plotly.offline.plot(data, filename='file name')
#pyecharts
bar/line.render('file name.html')
#bokeh
output_file("file name.html")Import the html file it generate on your local computer.
from IPython.display import IFrame
IFrame('file_name.html', width=700, height=400)This is caused by the temporary cache in the browser. The first solution is you can change another browser to visit the link. The second, add several ??? in the nbviwer link to refresh the link, then the chart will display correctly.
If you have any questions, or seek for help troubleshooting, please create an issue here
In Jupyter notebook, you can write shell commands after !. cat is essentially a shell command that reads the content of a file and output to the screen. It is a common way to check if the output (to a file) is intended. In notes-week-01.md, we have learned some useful commands like cd, pwd and ls. Those can all be used here. Also recall how we install new Python modules in a Jupyter notebook: !pip install <package-name>.