David Brochart
@davidbrochart.bsky.social
Better not to start, once begun better to finish.
September 17, 2025 at 12:09 PM
Sneak preview of what's coming to Microverse: real-time collaboration 🚀
This is not a re-implementation of JupyterLab's RTC so that it runs in the browser, it is the same Jupyter Collaboration package.
That's exactly what I expect Microverse to bring to the table: more compatibility with JupyterLab.
This is not a re-implementation of JupyterLab's RTC so that it runs in the browser, it is the same Jupyter Collaboration package.
That's exactly what I expect Microverse to bring to the table: more compatibility with JupyterLab.
August 7, 2025 at 12:14 PM
Sneak preview of what's coming to Microverse: real-time collaboration 🚀
This is not a re-implementation of JupyterLab's RTC so that it runs in the browser, it is the same Jupyter Collaboration package.
That's exactly what I expect Microverse to bring to the table: more compatibility with JupyterLab.
This is not a re-implementation of JupyterLab's RTC so that it runs in the browser, it is the same Jupyter Collaboration package.
That's exactly what I expect Microverse to bring to the table: more compatibility with JupyterLab.
JupyterGIS-tiler can do a lot of processing in the backend. For instance you can clip a raster layer with a vector polygon, and have sub-pixel rendering.
Here is an example of DEM clipping with the Amazon river watershed:
github.com/geojupyter/j...
Here is an example of DEM clipping with the Amazon river watershed:
github.com/geojupyter/j...
August 6, 2025 at 10:14 AM
JupyterGIS-tiler can do a lot of processing in the backend. For instance you can clip a raster layer with a vector polygon, and have sub-pixel rendering.
Here is an example of DEM clipping with the Amazon river watershed:
github.com/geojupyter/j...
Here is an example of DEM clipping with the Amazon river watershed:
github.com/geojupyter/j...
Microverse is deployed on GitHub pages here:
davidbrochart.github.io/microverse/
Note that it doesn't work in Firefox and that it's quite slow to start up (see dev network).
It's still very early stage, but it's improving. One interesting feature is how kernels can be shared between browser windows.
davidbrochart.github.io/microverse/
Note that it doesn't work in Firefox and that it's quite slow to start up (see dev network).
It's still very early stage, but it's improving. One interesting feature is how kernels can be shared between browser windows.
July 28, 2025 at 4:50 PM
Microverse is deployed on GitHub pages here:
davidbrochart.github.io/microverse/
Note that it doesn't work in Firefox and that it's quite slow to start up (see dev network).
It's still very early stage, but it's improving. One interesting feature is how kernels can be shared between browser windows.
davidbrochart.github.io/microverse/
Note that it doesn't work in Firefox and that it's quite slow to start up (see dev network).
It's still very early stage, but it's improving. One interesting feature is how kernels can be shared between browser windows.
Microverse is a new in-browser version of JupyterLab. Don't we already have JupyterLite, you might ask?
While there is no web server in JupyterLite, there is one in microverse, and it is Jupyverse.
I think this will open up new possibilities.
While there is no web server in JupyterLite, there is one in microverse, and it is Jupyverse.
I think this will open up new possibilities.
July 16, 2025 at 8:56 AM
Microverse is a new in-browser version of JupyterLab. Don't we already have JupyterLite, you might ask?
While there is no web server in JupyterLite, there is one in microverse, and it is Jupyverse.
I think this will open up new possibilities.
While there is no web server in JupyterLite, there is one in microverse, and it is Jupyverse.
I think this will open up new possibilities.
My talk at PyData Paris on "Parallel processing using CRDTs" has been accepted and I want to illustrate it with the image bellow.
The small ships are all the peers that are concurrently mutating the global state (the big ship). They can all diverge temporarily but they tend to converge eventually.
The small ships are all the peers that are concurrently mutating the global state (the big ship). They can all diverge temporarily but they tend to converge eventually.
June 17, 2025 at 1:54 PM
My talk at PyData Paris on "Parallel processing using CRDTs" has been accepted and I want to illustrate it with the image bellow.
The small ships are all the peers that are concurrently mutating the global state (the big ship). They can all diverge temporarily but they tend to converge eventually.
The small ships are all the peers that are concurrently mutating the global state (the big ship). They can all diverge temporarily but they tend to converge eventually.
I'm working on a new in-browser JupyterLab. The difference with JupyterLite is that there is a server running in the browser 🚀
It's actually Jupyverse, a FastAPI-based Jupyter server.
It's actually Jupyverse, a FastAPI-based Jupyter server.
June 4, 2025 at 1:33 PM
I'm working on a new in-browser JupyterLab. The difference with JupyterLite is that there is a server running in the browser 🚀
It's actually Jupyverse, a FastAPI-based Jupyter server.
It's actually Jupyverse, a FastAPI-based Jupyter server.
jupytergis-tiler allows to visualize local or remote Xarray datasets in JupyterGIS.
It doesn't matter if the data is huge and doesn't fit in memory, the tiler will render to the screen just what you need to see.
All powered by @developmentseed.org's awesome TiTiler!
github.com/geojupyter/j...
It doesn't matter if the data is huge and doesn't fit in memory, the tiler will render to the screen just what you need to see.
All powered by @developmentseed.org's awesome TiTiler!
github.com/geojupyter/j...
May 23, 2025 at 7:48 AM
jupytergis-tiler allows to visualize local or remote Xarray datasets in JupyterGIS.
It doesn't matter if the data is huge and doesn't fit in memory, the tiler will render to the screen just what you need to see.
All powered by @developmentseed.org's awesome TiTiler!
github.com/geojupyter/j...
It doesn't matter if the data is huge and doesn't fit in memory, the tiler will render to the screen just what you need to see.
All powered by @developmentseed.org's awesome TiTiler!
github.com/geojupyter/j...
And it looks like Litestar is not going to catch up.
January 30, 2025 at 2:17 PM
And it looks like Litestar is not going to catch up.
Marimo 0.10.10 is out, with real-time collaboration 🎉
This is using the same technology as Jupyter (CRDTs). It's great to see how the two projects can benefit from each other.
As a contributor to both projects I think there can be a future where they share more functionalities.
This is using the same technology as Jupyter (CRDTs). It's great to see how the two projects can benefit from each other.
As a contributor to both projects I think there can be a future where they share more functionalities.
January 10, 2025 at 2:54 PM
Marimo 0.10.10 is out, with real-time collaboration 🎉
This is using the same technology as Jupyter (CRDTs). It's great to see how the two projects can benefit from each other.
As a contributor to both projects I think there can be a future where they share more functionalities.
This is using the same technology as Jupyter (CRDTs). It's great to see how the two projects can benefit from each other.
As a contributor to both projects I think there can be a future where they share more functionalities.
I just released jpterm 0.2.20, with an update to Textual v1.0 🚀
Textual's ability to maximize the focused widget is awesome from an accessibility point of view. For instance here you can view only the notebook cell that you are currently editing.
Textual's ability to maximize the focused widget is awesome from an accessibility point of view. For instance here you can view only the notebook cell that you are currently editing.
December 15, 2024 at 8:25 AM
I just released jpterm 0.2.20, with an update to Textual v1.0 🚀
Textual's ability to maximize the focused widget is awesome from an accessibility point of view. For instance here you can view only the notebook cell that you are currently editing.
Textual's ability to maximize the focused widget is awesome from an accessibility point of view. For instance here you can view only the notebook cell that you are currently editing.
Hacking around with real-time collaboration in Marimo.
November 21, 2024 at 4:39 PM
Hacking around with real-time collaboration in Marimo.