medialab/xan 0.57.0

v0.57.0

on GitHub

The temporal update.

Breaking

• xan select -n will not error anymore on empty inputs and, generally, empty files should not trigger selection errors when using commands with -n/--no-headers.
• xan heatmap -C/--cram becomes a flag accepting either auto, always or never.
• Dropping -C short flag for xan sort --cells (it could be confused with --columns or --check).
• Completely overhauled how datetimes work in moonblade.
• xan separate will not trim splitted values with some modes by default anymore.
• Dropping xan network --stats in favor of -f stats.
• -D becomes short flag for xan network --degrees instead of --disjoint-keys.
• xan separate --capture-groups is dropped in favor of -c/--captures & -C/--all-captures.
• Renaming xan search --breakdown shortflag to -b to allow for future -B/--before-context.

Features

• Adding xan matrix count & xan matrix adj.
• Adding front_coding window function.
• Timestamp support with xan plot -LT.
• Adding xan rename -n/--no-headers support for -p/--prefix & -x/--suffix.
• Adding xan from -f parquet (requires the parquet feature).
• Adding xan to latex.
• Adding xan top -L/--lexicographic.
• Adding xan heatmap flags: -w/--width, -F/--fill, -a/--align, -U/--unit, -Z/--show-normalized, -A/--ascii, -l/--label & -v/--values.
• Adding new gradients to xan heatmap.
• Adding range & repeat moonblade functions.
• Adding xan sort --columns.
• Adding xan view -T/--tee.
• Adding now, fractional_days, to_timezone, to_local_timezone, with_timezone, with_local_timezone, without_timezone, to_timestamp, to_timestamp_ms, from_timestamp, from_timestamp_ms, span, date & time moonblade functions.
• Better type inference with xan stats, and the type & types aggregation functions, now including more types for temporal values (zoned_datetime, datetime, date & time).
• Adding xan input -T/--tolerant.
• Adding xan separate --trim.
• Adding xan grep -B/--before-context & -A/--after-context.
• Adding xan network -f=components, -S/--simple, --union-find, --minify & --sample-size <n>.
• Adding xan plot --timezone.
• Adding xan hist --log shorthand flag for --scale=log.
• Adding log_dist sparkline column to xan stats -q output.
• Adding dist & log_dist aggregation functions.
• Adding xan search -L/--levenshtein <k> & -D/--damerau-levenshtein <k>.

Fixes

• Fixing xan separate automatic column prefix extraction.
• Fixing xan heatmap -n.
• Fixing xan heatmap --repeat-headers --cram always not repeating x-axis legend.
• Fixing correctness of xan plot -T and increase resolution to microseconds.
• Fixing moonblade column-related functions returning incorrect results wrt -n/--no-headers.
• xan search should now properly error when handling invalid utf-8 in relevant modes.
• Fixing xan search -iR & xan search -i --replacement-column.

Performance

• Improving performance of xan complete, xan top, xan plot -T & xan hist.
• Improving overall performance of xan network.
• Slightly optimizing xan vocab by allowing needless heap allocation & indirection.
• Improving performance and memory usage of xan separate.

Quality of Life

• Adding proper help to xan heatmap.