Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
This page is currently under development and will be updated soon.
Jupyter notebooks are a common way to run Python code for data science or data exploration. You can find more information about how TileDB Cloud manages JupyterLab instances here.
You can launch a JupyterLab instance by clicking on Launch Notebook
from the left menu, select and an image and the server size, and click on the Start
button. After about 30 seconds, you are all set to start writing some code. To shut down the notebook server, simply click on the Shut down
button (it also takes a few seconds).
The notebook will also automatically shutdown after 30 minutes if you close the tab with the notebook server still running.
This page is currently under development and will be updated soon.
You can also make a notebook public, effectively sharing it with every user on TileDB Cloud. This can make your notebook discoverable by other users that wish to explore public notebooks, or you can discover useful code that others wish to share with the world.
To make a notebook public, you just need to navigate to the Settings
of that notebook and click on Make public
as shown below. Similarly, you can always switch the notebook back to private mode at any time.
When making a notebook public, you do not get charged when other users run the notebook. You only get charged when you run public notebooks.
This page group contains simple recipes for managing your notebooks. You can find the contents below:
This page is currently under development and will be updated soon.
To create a Jupyter notebook, you first need to navigate to the notebooks
tab from the sidebar under your assets
and then click the "plus" icon. That will prompt a dialog asking you to specify the notebook name, the path of the cloud storage space where the physical notebook will live and your cloud credentials that will allow TileDB Cloud to access that storage space. Once you create the notebook, you can launch it, edit and save it as you would otherwise do for any Jupyter notebook.
Alternatively you can create a notebook from the launch a JupyterLab notebook instance. Once you are in the Jupyter notebook environment, navigate to File -> New -> TileDB notebook
.
You can deregister a notebook by navigating to its Settings
and clicking on Deregister Notebook
.
Deleting (deregistering) a notebook does not physically delete your notebook from the physical cloud storage. It simply deregisters the notebook from TileDB Cloud. Your data will still be accessible by you outside of TileDB Cloud if you own the appropriate AWS access keys.
Renaming a notebook (in Settings
) is under the danger zone, because from that point onwards you (and all the users you shared the notebook with) will have to change your code to add the new notebook name. The notebook will still be shared and accessible by the other users, but they will need to add the new name to their code. In other words, TileDB Cloud does not currently support automatic redirection of notebook URIs upon renaming.
If you are using a notebook through TileDB Cloud's embedded Jupyter Lab environment, you can create any array or file inside your dedicated EBS volume. Just make sure you use ~/path/to/your/array
for array names, as the current working directory of the launched Jupyter environment might be different from your EBS home directory.
This page is currently under development and will be updated soon.
You can share a registered notebook with any other user on TileDB Cloud. To share a notebook, find it on Assets -> Notebooks
and either click on the sharing button located on the right end of the notebook card in the list, or click on the notebook card and navigate to the Sharing
tab. The added member will appear in the notebook members list, where you will be able to change the access policy or revoke access from the user. Users get notified by email when someone shares a notebook with them.
When sharing a notebook with other users, you do not get charged when those users run the notebook. You only get charged when you run notebooks.
When sharing with a member, TileDB Cloud uses auto-complete to facilitate finding a username you are looking for. Similar to GitHub/GitLab, the usernames are considered public information (in contrast to full names and emails that are protected). Please email us at privacy@tiledb.com
if you wish your username to be excluded from auto-complete.
Note that the notebook URL when you are viewing its Overview
is shareable, and another user can view it on their browser if they have access to it. URLs of public notebooks can be viewed by users, even if they are not logged in.
TileDB Cloud notebook servers include a persistent home directory for you to install custom packages.
TileDB Cloud supports pip
, conda
, cran
, and other methods of installing packages. When installing a package, you should install it in your home directory if you want it persisted after reboots.
For Python, you can use pip
to install packages into your home directory with the --user
option.
If a package from pip
is installed without the --user
flag, it will not be persisted and will not be available upon reboot
Both conda
and mamba
are available in the notebook environment to install any available packages. Creating a custom environment in your home directory will allow you to install and persist any packages.
If conda install
or mamba install
is used outside of a custom environment, your packages will not be persisted.
R packages can be installed from CRAN by setting the `lib location to the home directory for persistence.
If package installations are done without setting the library path to your home directory, it's likely the installed packages will not be persisted.
This page is currently under development and will be updated soon.
You can explore public notebooks, adding a variety of filters, from the Explore
page.
TileDB Cloud support uploading notebooks from git repositories. Its common that you might have notebooks stored in source control such as github or gitlab and wish to automatically upload them to TileDB Cloud for usage.
TileDB provides public github actions that let you upload a notebook.
Example usage includes:
Gitlab CI doesn't offer a marketplace or public template support currently. Instead find the bellow example for setting up a gitlab ci setup.
An API Token is required to access TileDB Cloud. Its recommended to set this as a secret in gitlab. This should be set as an environmental variable for TILEDB_REST_TOKEN
.