parquet_metadata.json

{
  "hep-adl-ethz": {
    "s3_prefix": "hep-parquet/native/Run2012B_SingleMu-65536000.parquet/",
    "tables": [
      {
        "url_regex": "s3://hep-adl-ethz/hep-parquet/native/(Run2012B_SingleMu)-65536000.parquet/[^/]*\\.snappy\\.parquet",
        "table_key_format": "{0}"
      }
    ]
  },
  "overturemaps-us-west-2": {
    "s3_prefix": "release/2024-04-16-beta.0/",
    "tables": [
      {
        "url_regex": "s3://overturemaps-us-west-2/release/2024-04-16-beta\\.0/theme=([^/]*)/type=[^/]*/[^/]*\\.zstd\\.parquet",
        "table_key_format": "{0}"
      }
    ]
  },
  "daylight-openstreetmap": {
    "s3_prefix": "parquet/",
    "tables": [
      {
        "url_regex": "s3://daylight-openstreetmap/parquet/([^/]*)/release=v1\\.45/type=[^/]*/[^/]*",
        "format": "parquet",
        "table_key_format": "{0}"
      }
    ]
  },
  "aws-public-blockchain": {
    "s3_prefix": "v1.0/",
    "tables": [
      {
        "url_regex": "s3://aws-public-blockchain/v1\\.0/(btc)/(transactions)/date=[^/]*/part-[^/]*\\.snappy\\.parquet",
        "table_key_format": "{0}-{1}"
      },
      {
        "url_regex": "s3://aws-public-blockchain/v1\\.0/(eth)/(logs)/date=[^/]*/[^/]*\\.snappy\\.parquet",
        "table_key_format": "{0}-{1}"
      }
    ]
  },
  "aws-roda-hcls-datalake": {
    "tables": [
      {
        "url_regex": "s3://aws-roda-hcls-datalake/(clinvar_summary_variants)/([^/]*)/[^/]*\\.snappy\\.parquet",
        "table_key_format": "{0}-{1}"
      },
      {
        "url_regex": "s3://aws-roda-hcls-datalake/(opentargets_latest)/([^/]*)/[^/]*snappy\\.parquet",
        "table_key_format": "{0}-{1}"
      },
      {
        "url_regex": "s3://aws-roda-hcls-datalake/(thousandgenomes_dragen)/(var_partby_samples)/[^/]*/[^/]*\\.snappy\\.parquet",
        "table_key_format": "{0}-{1}"
      },
      {
        "url_regex": "s3://aws-roda-hcls-datalake/(gnomad)/(chrm)/[^/]*snappy\\.parquet",
        "table_key_format": "{0}-{1}"
      },
      {
        "url_regex": "s3://aws-roda-hcls-datalake/(gnomad)/(sites)/partition_0=[^/]*/[^/]*snappy\\.parquet",
        "table_key_format": "{0}-{1}"
      },
      {
        "url_regex": "s3://aws-roda-hcls-datalake/(gtex_8)/([^/]*)/[^/]*\\.snappy\\.parquet",
        "table_key_format": "{0}-{1}"
      }
    ]
  }
}