FoRecoML
The machine-learning arm of a forecast reconciliation toolkit, four months old and already sharing its sibling's plumbing.
A side-by-side editorial comparison of carrier and nuggets — release velocity, themes, recent moves, and the top alternatives to consider.
carrier's 0.3.0 quietly rewrites how crated functions see each other — and breaks code that relied on the old rules.
carrier packages an R function together with the data it needs so it can be shipped to another process or machine. 0.3.0 changed the semantics of functions passed through ...: they are now re-crated inside the main crate environment instead of inheriting from wherever they were defined. That fixes helper functions failing to resolve each other remotely, and stops local environments dragging large objects into the payload.
nuggets keeps compounding on the 2.0 rewrite — more pattern families, lighter install.
nuggets searches for association rules, contrasts and other conditional patterns in the GUHA tradition, with a C++ core behind dig() and an interactive explore() app for reading results. Since the 2.0 rewrite of that core, every release has widened the same three surfaces: more pattern families to mine, more of explore() to inspect them in, and steady performance work underneath. The most recent tag optimises dig() on sparse crisp data with a sparse bit chain and adds clustering characteristics to explore() for association rules.
carrier packages an R function together with the data it needs so it can be shipped to another process or machine. 0.3.0 changed the semantics of functions passed through ...: they are now re-crated inside the main crate environment instead of inheriting from wherever they were defined. That fixes helper functions failing to resolve each other remotely, and stops local environments dragging large objects into the payload.
Each release tightens the boundary between a crate and its surroundings. 0.2.0 introduced .parent_env defaulting to baseenv() to cut the crate off from the global search path and made unnamed ... arguments an error instead of a silent drop; 0.3.0 finishes the job by controlling the environment chain of the packed closures themselves. The package is trading backwards compatibility for predictable, self-contained payloads — a reasonable bet for something whose failures otherwise surface on a remote worker.
The 0.1.0 notes promised an automatic dependency-detection mode to replace explicit packing; nothing since has delivered it, and with the environment semantics now settled that is the obvious next move — though the entries give no date for it.
nuggets searches for association rules, contrasts and other conditional patterns in the GUHA tradition, with a C++ core behind dig() and an interactive explore() app for reading results. Since the 2.0 rewrite of that core, every release has widened the same three surfaces: more pattern families to mine, more of explore() to inspect them in, and steady performance work underneath. The most recent tag optimises dig() on sparse crisp data with a sparse bit chain and adds clustering characteristics to explore() for association rules.
Two forces are shaping the package. One is coverage: baseline, complement and paired-baseline contrasts, correlations, tautologies, ancestors and clustering have all been added as first-class dig_ or explore_ surfaces, so the same search engine now answers a widening set of questions. The other is weight — Shiny packages moved from Imports to Suggests, BH and RcppThread dropped, XSIMD updated, parse_condition() rewritten in C++ — which keeps a package with an interactive app from forcing that app's dependencies on every user. Deprecations are handled through lifecycle rather than removed abruptly.
Expect the sparse-data optimisation to extend from crisp to fuzzy data, and explore() to keep gaining tabs as each new pattern family lands, on the roughly six-week cadence the 2.2 line has held.
Other Infra & APIs products tracked by Sparkpulse, ranked by recent ship velocity. Each card links to a full editorial trajectory and lets you pivot into a head-to-head comparison with either carrier or nuggets.
The machine-learning arm of a forecast reconciliation toolkit, four months old and already sharing its sibling's plumbing.
Forecast reconciliation with a real object model, five years after it started returning bare matrices.
A textbook data package whose whole job is to stay installable, and whose releases prove how much work that is.
A graph-centrality package that spent 2026 making its existing measures usable at scale, then went quiet.
A test-theory package that grew into a graphical-model toolkit, now spending its releases paying down the API debt that growth created.
projoint spent a year on CRAN paperwork, then shipped a correctness fix it flagged itself.
See all carrier alternatives → · See all nuggets alternatives →
Latest ship moves from both products, interleaved chronologically. ⚡ = editorial spark.
Both compete on the same themes — r-package — within Infra & APIs. nuggets is currently shipping more aggressively (velocity 2.5 vs 0.0), with 0 editorial sparks in the last 30 days against 0. See the at-a-glance table above for a side-by-side breakdown of velocity, recent sparks, and editorial themes.
Sparkpulse doesn't pick a winner — we score release velocity, not feature parity. nuggets is currently shipping more aggressively (velocity 2.5 vs 0.0), with 0 editorial sparks in the last 30 days against 0. For your specific use case, the alternatives sections above list other Infra & APIs products to evaluate alongside.
Top carrier alternatives in Infra & APIs are ranked by recent ship velocity. Browse the "carrier alternatives" section above for the current picks, or visit /alternatives/carrier for the full list with editorial commentary on each.
Top nuggets alternatives in Infra & APIs are ranked by recent ship velocity. Browse the "nuggets alternatives" section above for the current picks, or visit /alternatives/nuggets for the full list with editorial commentary on each.