Scanning a series of build reports

Vincent J. Carey, stvjc at channing.harvard.edu

July 16, 2023

Source: vignettes/scan.Rmd

Introduction

Build reports on 3.17 were collected from 2023-06-15 to 2023-07-15.

The report.tgz are cached using a cron job that in turn uses R and bbsBuildArtifacts. This takes care of locating and renaming the report.tgz files.

# m h  dom mon dow   command
37 12 * * * Rscript -e 'library(bbsBuildArtifacts); setup_artifacts(version="3.17", hostnames=c(linux="nebbiolo1", macos="merida1"))' >> mylog.log
37 13 * * * Rscript -e 'library(bbsBuildArtifacts); setup_artifacts(version="3.18", hostnames=c(linux="nebbiolo2", macos="lconway"))' >> mylog.log

The entries of the log file look like:

bsBuildArtifacts ArtifSet instance.
  2230 pkg paths for type bioc, Bioconductor version 3.17.
  23 extra file paths.
  tarball production date: 2023-07-14
R version: 4.3.1 (2023-06-16) -- "Beagle Scouts"
Platforms:
  Linux (Ubuntu 22.04.2 LTS)
  macOS 12.6.4 Monterey
Use paths(aset)[...] to retrieve selected paths.
bbsBuildArtifacts ArtifSet instance.
  2214 pkg paths for type bioc, Bioconductor version 3.18.
  26 extra file paths.
  tarball production date: 2023-07-15
R version: 4.3.1 (2023-06-16) -- "Beagle Scouts"
Platforms:
  Linux (Ubuntu 22.04.2 LTS)
  macOS 12.6.5 Monterey
Use paths(aset)[...] to retrieve selected paths.

There are version-, host-, package-, and phase-specific data collected on all software packages every time setup_artifacts is run. Our task is to harvest the information in ways that are useful to developers and core.

Time series extraction

We serialized the results of scan_artifs for dates ranging from 2023-06-15 to 2023-07-15. This was done for 3.17 and 3.18. Two types of information are retrieved in these scans – the number of OK outcomes for installation, and the number of OK outcomes for checksrc.

library(bbsBuildArtifacts)
data(scan_demo_3.17)
scan_demo_3.17
## bbsBuildArtifacts artifset_scan for version 3.17
##  start 2023-06-15, end 2023-07-15

We visualize the series:

plot_scan_inst(scan_demo_3.17) + ggplot2::ylim(2120,2250)
## Warning: Removed 2 rows containing missing values (`geom_line()`).

Similarly for checksrc outcomes:

plot_scan_chk(scan_demo_3.17) + ggplot2::ylim(1590,1720)
## Warning: Removed 2 rows containing missing values (`geom_line()`).