Add Support for Multifile Distribution Downloads

[kylstevenson]

This adds support for downloading multiple files from a single distribution. Previously, the SDK assumed each distribution contained exactly one file. This update enables distributions to contain multiple files and provides methods to list, filter, and download them individually or in bulk.

What's Changed

New Features

1. List Distribution Files

A new method to discover all files available in a distribution:

Map<String, DistributionFile> files = fusion.listDistributionFiles(
    "my-catalog",       // catalog identifier
    "my-dataset",       // dataset identifier
    "my-series-member", // series member identifier
    "csv",              // distribution/file format
    0                   // maxResults (0 = all files)
);

2. Selective File Downloads

Download specific files from a distribution instead of all files:

List<String> fileNames = Arrays.asList("file1", "file2");
fusion.download("my-catalog", "my-dataset", "my-series-member", "csv", "/downloads", fileNames);

3. Multifile Streaming

Stream multiple files with individual InputStream access:

Map<String, InputStream> streams = fusion.downloadStream("my-catalog", "my-dataset", "my-series-member", "csv");
InputStream fileStream = streams.get("file1");

5. Checksum Validation changes for Glue Delivery

For datasets delivered via AWS Glue (currently the only delivery channel supporting multifile distributions), checksum validation is automatically skipped when checksums are not provided by the API. This prevents download failures for Glue-delivered multifile distributions where checksums may not be available.

The SDK now:

• Detects if a dataset uses the glue delivery channel
• Automatically skips checksum validation when no checksum is provided
• Maintains strict validation for all other delivery channels and single-file distributions

Breaking Changes

1. downloadStream() Return Type Changed

Before:

InputStream stream = fusion.downloadStream("catalog", "dataset", "series", "csv");

After:

Map<String, InputStream> streams = fusion.downloadStream("catalog", "dataset", "series", "csv");
InputStream stream = streams.get("filename");
// Or for single-file distributions:
InputStream stream = streams.values().iterator().next();

Impact: All existing code using downloadStream() must be updated to handle the Map<String, InputStream> return type.

2. download() File Naming Changed

Before:

• Single file downloaded with generated name: {catalog}_{dataset}_{series}.{distribution}
• Example: common_API_TEST_20230308.csv

After:

• All files downloaded with their original API names (unless no file name is provided)
• Files saved to: {downloadPath}/{filename}
• Example: downloads/file1.csv, downloads/file2.csv

Impact: Any code relying on the previous file naming convention needs to be updated to handle the new naming scheme. In the future we should return a File object to that specific file when downloading.

[Ramadevie]

LGTM