Package com.marksayson.deletionchecker.format

Class PackedFileWriter

java.lang.Object

com.marksayson.deletionchecker.format.PackedFileWriter

public final class PackedFileWriter
extends Object

Serializes one entity type's identifiers into a packed binary file (format v2): header, prefix index, Bloom filter, identifier offset table, identifier data, with the file's CRC32C patched into the header.

Build-time only — never invoked on the lookup path. Callers pass identifiers already sorted by unsigned bytes and deduplicated; write(java.lang.String, java.util.List<byte[]>) verifies both and rejects violations rather than producing a file that would fail mysteriously at query time.

Field Summary

Fields

Modifier and Type	Field	Description
static final double	DEFAULT_BLOOM_FPR	Default Bloom-filter false-positive rate — see write(String, List, int, double).

Method Summary

Modifier and Type	Method	Description
static byte[]	write(String entityType, List<byte[]> sortedIdentifiers)	Serializes a packed file using PrefixIndex.DEFAULT_BUCKET_SIZE and DEFAULT_BLOOM_FPR.
static byte[]	write(String entityType, List<byte[]> sortedIdentifiers, int bucketSize)	Serializes a packed file with DEFAULT_BLOOM_FPR.
static byte[]	write(String entityType, List<byte[]> sortedIdentifiers, int bucketSize, double bloomFpr)	Serializes a packed file.

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Field Details

DEFAULT_BLOOM_FPR

public static final double DEFAULT_BLOOM_FPR

Default Bloom-filter false-positive rate — see write(String, List, int, double).

See Also:

• Constant Field Values

Method Details

write

public static byte[] write(String entityType,
 List<byte[]> sortedIdentifiers)

Serializes a packed file using PrefixIndex.DEFAULT_BUCKET_SIZE and DEFAULT_BLOOM_FPR.

Parameters:

entityType - the entity type this file holds; ASCII, 1 to 64 bytes

sortedIdentifiers - the identifiers, already sorted by unsigned bytes and deduplicated

Returns:

the complete file bytes

write

public static byte[] write(String entityType,
 List<byte[]> sortedIdentifiers,
 int bucketSize)

Serializes a packed file with DEFAULT_BLOOM_FPR.

Parameters:

entityType - the entity type this file holds; ASCII, 1 to 64 bytes

sortedIdentifiers - the identifiers, already sorted by unsigned bytes and deduplicated

bucketSize - the prefix index bucket target size K; must be at least 1

Returns:

the complete file bytes

write

public static byte[] write(String entityType,
 List<byte[]> sortedIdentifiers,
 int bucketSize,
 double bloomFpr)

Serializes a packed file.

Parameters:

entityType - the entity type this file holds; ASCII, 1 to 64 bytes

sortedIdentifiers - the identifiers, already sorted by unsigned bytes and deduplicated

bucketSize - the prefix index bucket target size K; must be at least 1

bloomFpr - the Bloom filter's target false-positive rate; >= 1 disables the filter, otherwise must be positive

Returns:

the complete file bytes, with the CRC32C written into the header

Throws:

IllegalArgumentException - if sortedIdentifiers is not strictly ascending by unsigned bytes, if entityType is not 1 to 64 ASCII bytes, if bucketSize is less than 1, or if bloomFpr is not positive