I did something similar to visualize the snapshots and files. But instead of using the static website, I was using the Java API to get the metadata from HDFS and send it back to the frontend. Something like this: https://observablehq.com/@capkurmagati/iceberg-metadata-visualization My actual implementation does not use vega for visualization but you get the idea.
My team plans to integrate Iceberg table into our existing Hive-based catalog system. I think my team can work with the community. Bests, On Wed, Mar 3, 2021 at 10:29 AM Ryan Blue <rb...@netflix.com.invalid> wrote: > This is great! If there's an easy way to make this available to other > people I think that's a great idea. > > On Tue, Mar 2, 2021 at 9:24 AM John Zhuge <jzh...@apache.org> wrote: > >> Nice! >> >> On Tue, Mar 2, 2021 at 8:39 AM Filip <filip....@gmail.com> wrote: >> >>> Ooops, sorry, forgot to post a link to the tools screenshot sample >>> https://i.imgur.com/FERzd8X.png >>> >>> On 2021/03/02 16:18:22, Filip <filip....@gmail.com> wrote: >>> > Hi devs, >>> > >>> > With a lot of help from plotly.js [1] and some very basic vanilla >>> > javascript (sorry, I peaked in javascript back in the days of >>> prototype and >>> > scriptaculous) we managed to craft a tool to visually analyze rows and >>> data >>> > files metrics of an Iceberg metadata file. >>> > >>> > Just drag and drop an Iceberg metadata file and you get something >>> looking >>> > like this. >>> > [image: sample.iceberg.viz.png] >>> > >>> > Best part is we can host this on github pages cause it's just a >>> javascript >>> > file and one lonely html file (no CSS stylesheet, sorry :| I peaked >>> around >>> > the emergence of reset.css files). >>> > Let me know if you think this would make a good addition to the repo. >>> > >>> > [1] https://github.com/plotly/plotly.js/blob/master/LICENSE >>> > >>> > -- >>> > Filip Bocse >>> > >>> >> >> >> -- >> John Zhuge >> > > > -- > Ryan Blue > Software Engineer > Netflix >
