Ecosyste.ms: Awesome
An open API service indexing awesome lists of open source software.
https://github.com/jupyter/colaboratory
[deprecated] Jupyter CoLaboratory, goto google colab now
https://github.com/jupyter/colaboratory
Last synced: about 1 month ago
JSON representation
[deprecated] Jupyter CoLaboratory, goto google colab now
- Host: GitHub
- URL: https://github.com/jupyter/colaboratory
- Owner: jupyter
- License: bsd-3-clause
- Archived: true
- Created: 2014-03-06T19:40:00.000Z (over 10 years ago)
- Default Branch: master
- Last Pushed: 2017-12-28T17:27:23.000Z (almost 7 years ago)
- Last Synced: 2024-09-23T19:03:07.767Z (about 2 months ago)
- Language: JavaScript
- Homepage: https://colab.research.google.com
- Size: 1.01 MB
- Stars: 740
- Watchers: 76
- Forks: 111
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- Contributing: CONTRIBUTING.md
- License: LICENSE
Awesome Lists containing this project
- awesome-starred - jupyter/colaboratory - [deprecated] Jupyter CoLaboratory, goto google colab now (others)
README
# CoLaboratory
**This repository, jupyter/colaboratory is no longer being actively developed or maintained. For the latest in Colaboratory, check out [Colab from google research](https://colab.research.google.com/). That site is an up to date version maintained by the Google Colaboratory team.** For new issues see: https://github.com/googlecolab/colabtools
----------------------
This repo contains two related tools:
1. The CoLaboratory Chrome App.
2. CoLaboratory with Classic Jupyter Kernels.
Both of these create and store notebooks in Google Drive and allow for
collaborative editing of notebooks. The difference is that the Chrome
App executes all code inside the Chrome browser using the
[PNaCl Sandbox](https://developer.chrome.com/native-client/nacl-and-pnacl)),
while coLaboratory Classic executes code via local Jupyter kernels
(such as the IPython kernel) that have complete access to the host
system and files.
## Setup
First clone this repo:
```
git clone --recursive https://github.com/jupyter/colaboratory
```
## Installing CoLaboratory Classic Frontend for Local IPython Kernels
Run
```
cd colaboratory
pip install -r requirements.txt
```
to install the dependencies.
NOTE: This must be run from the colaboratory directory.
If you did not use the `--recursive` flag when cloning, you will get errors like:
```
[tornado.access] WARNING | 404 GET /static/closure/css/common.css
```
To fix this, run `git submodule init && git submodule update`.
Start IPython notebook:
```
python -m colaboratory
```
This launches the web application.
Navigate to ```http://127.0.0.1:8888/static/colab/welcome.html``` in your browser.
## Installing the CoLaboratory Chrome App
Run
```
python colaboratory/install_chrome.py
```
This creates an unpacked Chrome App, in the ```build_chrome/``` directory.
To install this app in Chrome, follow the instructions for installing an unpacked extension
(extensions are apps for these purposes), at https://developer.chrome.com/extensions/getstarted#unpacked.
The extension is located in ```colaboratory/build_chrome/```.
## Caveats/Requirements
### IPython Version
If you already had IPython installed, and the version you had installed was not 2.x, then Colaboratory will not work. You can either upgrade (or downgrade) to 2.x, or use virtual env. To check the version of IPython you have installed, run ```ipython --version```.
### Loading Python Libraries
Currently there is no way to install new libraries in the PNaCl kernel.
### Running from colaboratory.jupyter.org
The website [colaboratory.jupyter.org](http://colaboratory.jupyter.org) can be used for interactive editing of notebooks, but cannot yet connect to a local kernel. This is due to https issues that we are currently working on. However, running the Classic frontend as described above already works, as the website runs on localhost.
### The Collaboration Model
CoLaboratory's collaboration model is evolving. The current model is a single collaborative notebook with separate kernels. This can lead to a mismatch between a user's kernel state and the state of the notebook.
To understand how a state mismatch can manifest, consider the scenario below. Bob and Sue are working on the same notebook at the same time. Both Bob and Sue will have their own kernel state. Bobs changes will change the notebook Sue sees, but Sue's state is unchanged. If sue tries to access the variable Bob created, she will get an error.
## Development
You can simply edit JavaScript code in place and these changes will be reflected automatically. You may need to refresh the browser several times to clear the cache after making changes.
If you add new files or add/remove goog.require statements at the start of JavaScript files, you need to run the script build_colab_deps.py in order to compile the list of dependencies.