Get all locations-of-interest.csv files from git repo

all-locations-of-interest.sh

#!/bin/bash
# Get all versions of a file in a git repo

# Where to output new files
output_dir="/tmp/all_loc_csvs"

# Which file are we interested in?
file_of_interest="locations-of-interest/august-2021/locations-of-interest.csv"

mkdir -p "${output_dir}"

all_git_blobs=$(
    git rev-list --objects --all \
        | grep ${file_of_interest} \
        | cut -d ' ' -f 1
)

for blob in ${all_git_blobs}; do
    out_dirname_blob=$(dirname "${output_dir}/by-blob-id/${blob}/${file_of_interest}")

    mkdir -p "${out_dirname_blob}"

    out_filename=$(basename "${file_of_interest}")

    git cat-file -p "${blob}" > "${out_dirname_blob}/${out_filename}"
done


all_git_commits=$(
    git log --format=oneline \
        | cut -d ' ' -f 1
)

for commit in ${all_git_commits}; do
    if ! git ls-tree -r "${commit}" | grep "${file_of_interest}" > /dev/null; then
        continue
    fi
    commit_date=$(git show -s --format=%ct "${commit}")

    out_dirname_commit=$(dirname "${output_dir}/by-commit/${commit}/${file_of_interest}")
    out_dirname_date=$(dirname "${output_dir}/by-date/${commit_date}/${file_of_interest}")

    mkdir -p "${out_dirname_commit}"
    mkdir -p "${out_dirname_date}"

    out_filename=$(basename "${file_of_interest}")

    git show "${commit}:${file_of_interest}" > "${out_dirname_commit}/${out_filename}"
    git show "${commit}:${file_of_interest}" > "${out_dirname_date}/${out_filename}"
done