Package com.marksayson.deletionchecker

Class DeletionChecker

java.lang.Object

com.marksayson.deletionchecker.DeletionChecker

public final class DeletionChecker
extends Object

Local, exact deletion-state lookup over a packed, immutable binary dataset.

Answers whether an entity ID has been deleted without any runtime network call. The dataset is a set of packed binary files, one per entity type, plus a manifest describing the release. A consumer selects which entity types to load with load(java.nio.file.Path, java.util.Set<java.lang.String>); unselected types are never opened and incur no memory or startup cost.

The dataset is immutable for the lifetime of the instance — new deletions or a changed entity-type selection are picked up only by loading a new instance against an updated dataset. Lookups are exact: no false positives or negatives.

The mapped files are held for the life of the process; there is no close. Their pages are off-heap, file-backed memory that counts toward process RSS and container memory limits, not toward the Java heap.

Instances are thread-safe after load(java.nio.file.Path, java.util.Set<java.lang.String>) returns.

Method Summary

Modifier and Type	Method	Description
String	datasetVersion()	Returns the loaded release's datasetVersion — the ISO-8601 timestamp identifying which generation run's deletion data this instance holds.
<T> List<T>	filter(String entityType, List<T> items, Function<T,String> idExtractor)	Returns the items whose extracted identifier is not recorded as deleted for entityType, in their original order.
boolean	isDeleted(String entityType, String id)	Returns whether id is recorded as deleted for entityType.
static DeletionChecker	load(Path datasetDirectory, Set<String> entityTypes)	Loads the requested entity types from the dataset in datasetDirectory.
Instant	loadedAt()	Returns the instant load(java.nio.file.Path, java.util.Set<java.lang.String>) completed, i.e.

Methods inherited from class java.lang.Object

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

Method Details

load

public static DeletionChecker load(Path datasetDirectory,
 Set<String> entityTypes)
                            throws IOException

Loads the requested entity types from the dataset in datasetDirectory.

Reads and verifies manifest.json, confirms every requested entity type is listed in it, then for each requested type opens its packed file, checks its format version and header entity type, and verifies its checksum against the manifest. Any failure fails the load; partially loaded state is discarded. Entity types not requested are never opened.

Parameters:

datasetDirectory - the directory holding manifest.json and the packed files

entityTypes - the entity types to load

Returns:

a checker over the loaded entity types

Throws:

NullPointerException - if datasetDirectory or entityTypes is null

IOException - if the manifest or a requested type's file cannot be read

IllegalArgumentException - if a requested entity type is not listed in the manifest

InvalidManifestException - if the manifest is malformed or fails its checksum

UnsupportedManifestVersionException - if the manifest's schema version is not the one this build reads

CorruptDatasetException - if a requested type's file is corrupt, its header entity type does not match the manifest, or its checksum does not match the manifest

UnsupportedFormatVersionException - if a requested type's file has a binary format version this build does not read

isDeleted

public boolean isDeleted(String entityType,
 String id)

Returns whether id is recorded as deleted for entityType.

The result is exact — true only if this identifier is present in the loaded dataset for this entity type. Comparison is on the identifier's raw UTF-8 bytes; no hashing and no Unicode normalization is performed, so the caller must supply the identifier exactly as issued by the authoritative source.

Parameters:

entityType - the entity type to query; must have been requested when this instance was loaded

id - the identifier to check

Returns:

true if the identifier is deleted for this entity type, false otherwise

Throws:

IllegalArgumentException - if entityType was not requested when this instance was loaded, or if id is null, empty, exceeds 64 bytes when UTF-8 encoded, or contains an unpaired surrogate

filter

public <T> List<T> filter(String entityType,
 List<T> items,
 Function<T,String> idExtractor)

Returns the items whose extracted identifier is not recorded as deleted for entityType, in their original order.

Equivalent to calling isDeleted(java.lang.String, java.lang.String) on each item's extracted identifier and keeping those that are not deleted, but resolves entityType once for the whole batch.

Type Parameters:

T - the item type

Parameters:

entityType - the entity type to query; must have been requested when this instance was loaded

items - the items to filter

idExtractor - maps an item to the identifier to check

Returns:

a new list containing only the items that are not deleted, in input order

Throws:

NullPointerException - if items or idExtractor is null

IllegalArgumentException - if entityType was not requested when this instance was loaded, or if an extracted identifier is null, empty, exceeds 64 bytes when UTF-8 encoded, or contains an unpaired surrogate

datasetVersion

public String datasetVersion()

Returns the loaded release's datasetVersion — the ISO-8601 timestamp identifying which generation run's deletion data this instance holds.

Returns:

the dataset version string

loadedAt

public Instant loadedAt()

Returns the instant load(java.nio.file.Path, java.util.Set<java.lang.String>) completed, i.e. when this instance's dataset became active. Operators use it with datasetVersion() to spot instances running a stale release.

Returns:

the load-completion instant