NEWS.md
In autocodebook: Automatic Codebook and Tracking for 'Spark' and 'dplyr' Pipelines

autocodebook 0.1.0

Initial CRAN release.
Core API: auto_mutate(), auto_summarise(), auto_filter() wrap dplyr verbs and register variable metadata (type, source columns, categories, source code) automatically.
Manual registration via cb_register(); codebook accessors via cb_get(), cb_reset(), cb_render(), cb_export() (HTML, CSV, DOCX, XLSX).
Eligibility tracking: track_step(), track_get(), track_render(), track_export().
CONSORT-style flow tree: track_split(), track_outcomes(), flow_table(), flow_get(), flow_reset().
flow_diagram() draws a publication-ready CONSORT-style flowchart (a ggplot object) from the captured eligibility steps and the flow tree: a vertical trunk (cohort baseline, aggregated exclusions box, eligible cohort), one column per subgroup, and outcome boxes stacked beneath each subgroup. The standardized report embeds it automatically.
flow_diagram_export() saves that flowchart by file extension: raster (.png, .jpg, .tiff), vector (.pdf, .svg, .eps), an .emf metafile editable in Word, a .docx document with the flowchart embedded as a vector image, or a .pptx slide whose boxes and labels are fully editable PowerPoint shapes. generate_report() now also writes flowchart.docx alongside the report.
Big-data optimizations for sparklyr (tbl_spark): cb_checkpoint(), per-call cache and assume_unique, session defaults via cb_set_verbose() and cb_set_default_cache().
Standardized HTML report via generate_report(), with eligibility flowchart, editable codebook export, and cross-sectional or longitudinal variable inspection.