diff options
| author | Rutger Broekhoff | 2024-05-02 20:27:40 +0200 | 
|---|---|---|
| committer | Rutger Broekhoff | 2024-05-02 20:27:40 +0200 | 
| commit | 17a3ea880402338420699e03bcb24181e4ff3924 (patch) | |
| tree | da666ef91e0b60d20aa0b01529644c136fd1f4ab /script/synckv6/oeuf-synckv6.sh | |
| download | oeuf-17a3ea880402338420699e03bcb24181e4ff3924.tar.gz oeuf-17a3ea880402338420699e03bcb24181e4ff3924.zip  | |
Initial commit
Based on dc4ba6a
Diffstat (limited to 'script/synckv6/oeuf-synckv6.sh')
| -rwxr-xr-x | script/synckv6/oeuf-synckv6.sh | 43 | 
1 files changed, 43 insertions, 0 deletions
diff --git a/script/synckv6/oeuf-synckv6.sh b/script/synckv6/oeuf-synckv6.sh new file mode 100755 index 0000000..6b24347 --- /dev/null +++ b/script/synckv6/oeuf-synckv6.sh  | |||
| @@ -0,0 +1,43 @@ | |||
| 1 | #!/usr/bin/env bash | ||
| 2 | |||
| 3 | set -eu | ||
| 4 | set -o pipefail | ||
| 5 | |||
| 6 | export AWS_ACCESS_KEY_ID="$S3_ACCESS_KEY_ID" | ||
| 7 | export AWS_SECRET_ACCESS_KEY="$S3_SECRET_ACCESS_KEY" | ||
| 8 | |||
| 9 | set +x | ||
| 10 | all_files=() | ||
| 11 | declare -A metafiles | ||
| 12 | |||
| 13 | while IFS=' ' read -r size filename; do | ||
| 14 | if [[ "$filename" == *.parquet.meta.json ]]; then | ||
| 15 | metafiles["$filename"]=1 | ||
| 16 | else | ||
| 17 | all_files+=($filename) | ||
| 18 | fi | ||
| 19 | done < <(rclone ls \ | ||
| 20 | --s3-provider "$S3_PROVIDER" \ | ||
| 21 | --s3-region "$S3_REGION" \ | ||
| 22 | --s3-endpoint "$S3_ENDPOINT" \ | ||
| 23 | --s3-env-auth \ | ||
| 24 | :s3:$S3_BUCKET) | ||
| 25 | |||
| 26 | files=() | ||
| 27 | for filename in "${all_files[@]}"; do | ||
| 28 | if [[ -v metafiles["$filename.meta.json"] ]]; then | ||
| 29 | files+=($filename) | ||
| 30 | fi | ||
| 31 | done | ||
| 32 | |||
| 33 | echo "Found ${#files[@]} relevant KV6 Parquet files" | ||
| 34 | echo "Synching this directory with these files" | ||
| 35 | |||
| 36 | printf "%s\n" "${files[@]}" | rclone copy \ | ||
| 37 | --s3-provider "$S3_PROVIDER" \ | ||
| 38 | --s3-region "$S3_REGION" \ | ||
| 39 | --s3-endpoint "$S3_ENDPOINT" \ | ||
| 40 | --s3-env-auth \ | ||
| 41 | --progress \ | ||
| 42 | --files-from - \ | ||
| 43 | :s3:$S3_BUCKET ./ | ||