Hello everyone, I'm happy to announce the https://github.com/danielsz/cohere-clojure released under the MIT License. Cohere is a OpenAI contender with a focus on the enterprise and developer-friendliness. Also with a more relaxed free trial policy (non-expiring API key). My unofficial port of the SDK allows to execute tasks not possible via the REST API alone. Please share and enjoy!

https://github.com/babashka/babashka: Native, fast starting Clojure interpreter for scripting 1.3.183 (2023-08-22) • https://github.com/babashka/babashka/issues/1592: expose sci.core in babashka • https://github.com/babashka/babashka/issues/1596: Fix clojure.java.browse/browse-url truncates URLs with multiple query parameters on Windows • https://github.com/babashka/babashka/issues/1599: propagate error from run when task does not exist • Bump clj-yaml to 1.0.27 • https://github.com/babashka/babashka/issues/1604: throw FileNotFoundException when requiring namespace whose file cannot be found (as JVM Clojure does) • Bump integrant CI tests • https://github.com/babashka/babashka/issues/1600: use pagesize of 64K on linux aarch64, so it works on Asahi linux • Expose selmer.parser/resolve-arg • https://github.com/babashka/babashka/issues/1610: expose babashka.http-client.websocket namespace • Bump babashka.http-client to 0.4.14 • https://github.com/babashka/babashka/issues/1568: warn when task overrides built-in command

[com.wsscode/pathom3 "2023.08.22-alpha"] is out! This release includes some general improvements and a few breaking changes, most notably is the one related to placeholder nodes, which now won't add data to the children by default. This change is important because in case the user wants to have some custom params support, before this change it would also add the data down, which might be undesirable. There is a new plugin to add the placeholders data back for the ones that are using the previous feature. Changes on this version: • BREAKING: ::p.error/missing-output is now converged to ::p.error/attribute-missing (issue #149) • BREAKING: Placeholders won't override entity data via params by default (issue #187) • New pbip/placeholder-data-params plugin to get back the placeholder entity data behavior • Mutation not found error goes through ::pcr/wrap-mutate, allowing the user to capture and transform the not found error case • Remove Insufficient data calling resolver ... error due to problematic behavior with optionals (issue #192) • Improve runner error messages, elide at path part of the message when at root • Fix process-one return value when its false (previously it would return nil instead of false) (issue #195) • Include failing key on smart map errors • Support applyTo on resolvers and mutations (issue #203) https://clojars.org/com.wsscode/pathom3

https://github.com/techascent/tech.ml.dataset is out. TMD is an in-memory columnar data processing system similar to pandas or R's dplyr. If none of that makes sense to you then think of a very very optimized pathway for dealing with sequences of maps. • all the research from ham-fisted that applies has been moved up through the libraries with special attention to reduction as the primary pathway so for instance columns now have a very optimized reduction pathway. • many, many issues fixed. • many small tweaks here and there for performance. I know several of you have been waiting for 7 to come out of beta -- well, its out and we won't be changing API's any time soon 🙂. Enjoy 🙂

I’m open-sourcing https://github.com/recbus/caribou, a migration tool for Datomic. It has been tested against Datomic Cloud and Datomic Local (formerly dev-local) in varying conditions for the past six months. It is intended to work against Datomic Pro/On-Prem as well, but this capability has not yet been tested. Caribou’s tenets are that migrations are arbitrary transactions best expressed as a dependency graph and that the only reliable source of history for migrations is the database itself. At Sun Tribe Trading, we use Caribou for ensuring databases are “shaped” correctly before our application code can see it. Shaping involves arbitrary Datomic transactions including schema addition, seed data additions and data corrections. In addition to managing the ongoing evolution of our production database, Caribou is particularly useful for migrating in-memory test databases.