Feature request: ability to show/query manylinux wheels download per glibc version

[mayeut]

Is your feature request related to a problem? Please describe. Originally a request of a package maintainer in https://github.com/mayeut/manylinux-timeline/issues/529 The ability to show/query manylinux wheels downloads per glibc version allows package maintainers to decide when to drop older glibc versions based on actual download statistics rather than just some distributions EOL. It seems it would be a good thing to have in clickpy

Describe the solution you'd like This would require the following columns to be imported from BigQuery:

• details.distro.libc.lib
• details.distro.libc.version
• file.filename

Backfilling this information is most likely not necessary. It would likely require some new mv like:

https://github.com/mayeut/manylinux-timeline/issues/529

CREATE OR REPLACE TABLE pypi.pypi_downloads_per_day_by_version_by_glibc
(
    `project` String,
    `version` String,
    `date` Date,
    `glibc_version` String,
    `count` Int64
)
ENGINE = SummingMergeTree
ORDER BY (project, version, date, glibc_version)

CREATE MATERIALIZED VIEW pypi.pypi_downloads_per_day_by_version_by_glibc_mv TO pypi.pypi_downloads_per_day_by_version_by_glibc
(
    `project` String,
    `version` String,
    `date` Date,
    `glibc_version` String,
    `count` Int64
) AS
SELECT
    project,
    version,
    date,
    libc_version AS glibc_version,
    count() AS count
FROM pypi.pypi
WHERE libc_lib = "glibc" and match(filename, "(?:-|\.)manylinux[^-\.]+\.(?:[^-\.]+\.)*whl$")
GROUP BY
    project,
    version,
    date,
    glibc_version

Describe alternatives you've considered Use custom BigQuery queries with pypinfo

Additional context The overall (not per package) statistics can be seen there https://mayeut.github.io/manylinux-timeline/

[gingerwizard]

We'll do this in October @mayeut