Recreate a git repo from a bunch of PyPI releases.

pypi2git.sh

#! /usr/bin/env bash
#
# Recreate a git repo from a bunch of PyPI releases.
#
set -o errexit
set -o nounset
set -o pipefail
#set -o xtrace

PACKAGE=arcgis
REPO=arcgis-python
GHUSER=dotysan

main() {
    rm -fr "$REPO"
    reset_repo
    setup_repo

    get_all_rels

    pkg2git

    pushd "$REPO"
    git push --set-upstream origin master
    popd

    rm -fr "$PACKAGE"-1* "$PACKAGE"-2* "$PACKAGE"-3*
    gh gist create "$0" --public --desc 'Recreate a git repo from a bunch of PyPI releases.'
}

reset_repo() {
    local repo="$GHUSER/$REPO"

    if ! gh auth status
    then
        echo 'Paste a token from https://github.com/settings/tokens here, the press Ctrl-D to close stdin.'
        gh auth login --hostname github.com --with-token
    fi

    # switch from https default to ssh
    gh config set git_protocol ssh
    # note, everything is stored in ~/.config/gh/

    if gh repo view "$repo"
    then
        gh repo delete "$repo" --confirm
    fi

    gh repo create "$repo" --description="ESRI's Python API reconstructed." --private --clone
}

setup_repo() {
    pushd "$REPO"
    cat >.gitattributes <<-EOF
	# macros
	[attr]lfs filter=lfs diff=lfs merge=lfs -text

	# images
	*.png lfs

	# libraries
	*.pyd lfs
	*.so lfs
EOF
    git lfs track
    git add .gitattributes
    git commit -m 'Enable git LFS for PNG images'
    popd
}

get_all_rels() {
    local rel tb tbmt bn mt
    get_releases "$PACKAGE" |while read -r rel
    do
        tb="${rel##*/}"
        if [[ ! -s "$tb" ]]
        then wget -N "$rel"
        fi
        tbmt=$(stat -c%Y "$tb")

        bn="${tb::-7}"
        if [[ ! -d "$bn" ]]
        then
            mkdir "$bn"
            pushd "$bn"
            tar --extract --strip-components=1 --file="../$tb"
            ~/src/sh/lt.sh
            popd
        fi
    done
}

get_releases() {
    curl "https://pypi.org/pypi/$1/json" |jq --raw-output '.releases[] |.[] |.url'
}

pkg2git() {
    local rel mt rev

#    # extracted tarball subdirs sorted by mtime
#    find -mindepth 1 -maxdepth 1 -type d -name '*[0-9]*' -printf '%T@ %p\n' \
#      |sort |while read -r mt rel
    # extracted tarball subdirs sorted by name (will fail with 10 less than 9)
    find -mindepth 1 -maxdepth 1 -type d -name '*[0-9]*' -printf '%T@ %p\n' \
      |sort -k2 |while read -r mt rel
    do  echo "$rel"
        mt=$(date -d "@$mt" --rfc-2822)

        rsync --archive --delete \
            --exclude=.git/ \
            --exclude=.gitattributes \
            --exclude='*/widgets/js/dist/*' \
            "$rel/" "$REPO/"

        pushd "$REPO"

        # no point in tracking bin libs
        find -type f \( \
            -name '*.pyd' -o \
            -name '*.so' \
          \) -delete

        # standardize \n without \r
        find -type f \( \
            -name '*.cfg' -o \
            -name '*.css' -o \
            -name '*.in' -o \
            -name '*.js' -o \
            -name '*.json' -o \
            -name '*.md' -o \
            -name '*.py' -o \
            -name '*.ts' -o \
            -name '*.txt' -o \
            -name .gitignore -o \
            -name PKG-INFO \
          \) -print0 |xargs -r0 dos2unix --quiet

        git add --all
        rev="${rel##*-}"
        GIT_COMMITTER_DATE="$mt" git commit --date="$mt" -m "$rev" ||:

        popd
        #if grep ^1 <<<"$rev"
        #then echo good
        #else read -n1 -p 'Press any key to continue...' </dev/tty
        #fi
    done
}

main
exit 0

        #git branch "$rev"
        #git tag -a "$rev" -m "Release $rev"

        #git push origin "$rev"
        #git push origin "$rev" --tags