libjxl VS crux

https://github.com/libjxl/libjxl/blob/main/doc/format_overvi... is a pretty detailed but good overview. The highlights are variable size DCT (up to 128x128), ANS entropy prediction, and chroma from luminance prediction. https://github.com/libjxl/libjxl/blob/main/doc/encode_effort... also gives a good breakdown of features by effort level.

For JPEG XL, refer to its format overview [1]. In short its lossless mode uses a combination of multiple techniques: the rANS coding with an alias table, LZ77, reversible color transforms, a general vector quantization that subsumes palettes, a modified Haar transform and a learnable meta-adaptive decision tree for context modelling.

One good thing about JPEG XL is that its lossy mode also largely uses the same tool, with a major addition of specialized quantization and context modelling for low- and high-frequenty components.

[1] https://github.com/libjxl/libjxl/blob/main/doc/format_overvi...

The cjxl source is here. If you spot any reason why -x strip=exif may not work, tell me.

The problem is that the instructions for actually running the WASM file are not that clear... the docs the author mentions shows how to compile to WASM, which is easy enough, but then here's the instructions to make that actually work in the browser:

https://github.com/libjxl/libjxl/blob/main/tools/wasm_demo/R...

Yeah, you need some mysterious Python script, a JS service worker at runtime, choose whether you want the WASM or WASM_SIMD target, use a browser that supports Threads and SIMD if you chose that, make sure to serve everything with the appropriate custom HTTP headers... just reading that, I can see that to get this stuff working on non-browser WASM targets would likely require expertise in WASM, which is the point of the OP. WASM's UX is just not there yet.

https://github.com/libjxl/libjxl#usage

> Specifically for JPEG files, the default cjxl behavior is to apply lossless recompression and the default djxl behavior is to reconstruct the original JPEG file (when the extension of the output file is .jpg).

I mean compiling a bleeding edge kicad, inkscape or jpeg-xl is easy. But will probably trash your system if you already have an older version installed.

But you look at your image viewer that could have the lossless indicator? (and there is an issue open to add this indicator to the jxl files)

https://github.com/libjxl/libjxl/issues/432

You might want to use libjxl directly, e.g. for visually lossless images: cjxl --effort 9 --brotli_effort 11 --distance 1.0 --num_threads (nproc) --lossless_jpeg 0 input.png output.jxl on linux (if you're on windows/mac, just replace the (nproc) with the number of cpu threads you have, e.g. --num_threads 16).

Hi, a couple of my colleagues spent some time working on this integration with our open source database product (https://opencrux.com), and I'm curious to know - has anyone done similar things to connect Corda with a secondary off-the-shelf query engine?

For more details, see the release notes.

The context is that I work on on https://opencrux.com, which offers a bi-temporal Datalog query layer (as well as SQL) that more or less addresses the intersection of the two, since Datalog is great for expressing recursive queries.

I suppose another somewhat important distinction, once again performance related, is that graph databases will typically track index statistics to aid with query planning. For example, Crux uses stored knowledge of attribute-value cardinalities (recently via HyperLogLog) to optimise the join order of a query - this can make a big difference when attempting to traverse large graphs efficiently.

Agreed, recursive querying & bitemporal modelling in SQL are non-trivial problems, and the combination of the two is harder still. For an alternative perspective on tackling such problems I'd suggest looking at Datalog, which makes recursion a breeze, and a database with first-class bitemporality - both of which feature in https://opencrux.com (which I happen to work on :))

I work on Crux so can share a few details about our implementation of Datalog. The query is compiled into a kind of Worst-Case Optimal Join algorithm [0] which means that certain types of queries (e.g. cyclic graph-analytical queries, like counting triangles) are generally more efficient than what is possible with a non-WCOJ query execution strategy. However, the potency of this approach relies on the query planner calculating a good ordering of variables for the join order, and this is a hard problem in itself.

Crux is usually very competent at selecting a sensible variable ordering but when it makes a bad choice your query will take an unnecessary performance hit. The workaround for these situations is to break your query into smaller queries (since we don't wish to support any kind of hinting). Over the longer term we will be continuing to build more intelligent heuristics that make use of advanced population statistics. For instance we are about to merge a PR that uses HyperLogLog to inform attribute selectivity: https://github.com/juxt/crux/pull/1472

[0] https://cs.stanford.edu/people/chrismre/papers/paper49.Ngo.p...