MarkNahabedian/AnalyzeRegistryPackages.jl

Code for and results of running PackageAnalyzer.analyze on the Julia General Registry.

AnalyzeRegistryPackages*

Utility for running PackageAnalyzer.analyze on every package in a specified registry and collect the results in a tab separated values file.

One can run it on a local clone of the General Registry to avoid GitHub throttling:

using AnalyzeRegistryPackages
collect_stats_from_registry(joinpath(homedir(), ".julia/registries/General"))

For the Julia General registry I collected these results:

• raw data

• repository URIs

and made scatter plots

• lines of test code versus lines of source code

• lines of README file versus lines of source code

• lines of documentation versus lines of source code

• lines of README and documentation versus lines of source code

In the scatter plots, jll packages are red, Julia packages are green.

The following plots show how many of the currently registered (non-JLL) projects would pass at a given threshold value:

• satisfies-doc_min_fraction

• satisfies-readme_min_fraction

• satisfies-test_min_fraction

A number of registered projects were reported as unreachable by PackageAnalyzer. The function confirm_unreachable can be used to see if those repositories are still unreachable and record the HTTP status code to a tsv file.