Hi DataSketches community, I'm Yadu, a Scala developer, and I've recently started working with Apache DataSketches. While working with sketches, I found myself wanting a simple way to inspect their serialized contents without having to write a program against the library.
That led me to build *SketchProbe* <https://yadavan88.github.io/sketchprobe/>, a small open-source tool I'd like to share with the community and get your feedback on. SketchProbe aims to make that easier without requiring a separate program. *What it supports:* - *Input formats:* Base64, hex, MongoDB Extended JSON, and raw bytes. - *Decoding:* HLL, Theta, KLL, Classic Quantiles, and Frequent Items. - *Other families:* Recognizes their headers and metadata, although decoding isn't implemented yet. - *CLI:* Native binaries for macOS/Linux, Homebrew, and a JAR. - *Browser:* The same decoder compiled to WebAssembly. Everything runs locally in the browser; no data is uploaded. *Try it:* https://yadavan88.github.io/sketchprobe/ *Source:* https://github.com/yadavan88/sketchprobe (Apache-2.0) The tool is read-only, and I've also put some effort into handling malformed or potentially hostile input safely. I'd particularly appreciate feedback on two things: 1. *What should I decode next?* I'm considering CPC, Tuple and REQ. Are there other families or formats that would be more useful? 2. *Decoding correctness:* If you notice anything that doesn't match the reference implementations, please let me know. I'd be happy to investigate and fix it. I'm still relatively new to DataSketches, and so far I've mainly been working with HLL, Frequent Items, and KLL. There's plenty more for me to learn. Thanks for building and maintaining such a useful library. Looking forward to hearing your thoughts! Best regards, Yadu
