Final Cleanup for v1.0.0 #641
Merged
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Migrate from deprecated RedisZSetCommands.Range to org.springframework.data.domain.Range in all lexicographic query implementations to fix deprecation warnings marked for removal in Spring Data Redis 3.4.5. enabled. Also added lombok.config to configure Lombok behavior for the tests module. - Fixed text-blocks trailing whitespace warning in MetamodelGeneratorTest - Added AutoService dependency to test annotation processor classpath to resolve "Cannot find annotation method" warning - Updated lombok.config documentation to clarify that @nonnull warnings on primitives are intentional for constructor generation in test fixtures
Add missing Javadoc comments for public constructors and methods in lexicographic query classes. Update Spring Boot Javadoc URL to prevent redirect warning. - Add Javadoc for constructors and getters in marker classes - Document constructor parameters in LexicographicIndexer and LexicographicQueryExecutor - Fix Spring Boot docs URL from /docs/current/api/ to /api/java/
Configure build to skip Javadoc tasks for all demo modules since they are
example applications not part of the published API. This reduces build time
and avoids generating unnecessary documentation for non-library code.
- Add condition in root build.gradle to skip Javadoc for :demos: modules
- Skip both javadoc and javadocJar tasks using onlyIf { false }
- Keep Javadoc enabled for main library modules (redis-om-spring, redis-om-spring-ai)
Upgrade Antora documentation generator and resolve all build warnings for cleaner
documentation builds.
- Update Antora CLI and site generator from 3.1.4 to 3.1.12
- Update Node.js to 20.17.0 LTS and npm to 10.9.2
- Fix deprecated npm package warnings (inflight, glob, gulp-vinyl-zip)
- Replace deprecated {attachmentsdir} with modern Antora 3.x attachment syntax
- Update api-reference.adoc to use xref:attachment$ instead of link:{attachmentsdir}
- Change javadoc-base-url to use relative path instead of deprecated attribute
Replaced deprecated project.buildDir and project.projectDir with layout.buildDirectory and layout.projectDirectory to avoid execution-time access. This fixes Gradle 8.x deprecation warnings and ensures compatibility with future Gradle versions that will enforce configuration cache.
…#639) Fixes ClassCastException when count() method receives "num_docs" as a String from Redis FT.INFO command instead of a numeric type. The method now properly handles both String and Number types by parsing or converting them to long. - Parse String values using Long.parseLong() - Convert Number values using longValue() - Return 0 for null or unexpected types - Added test case to verify fix in EntityStreamsIssuesTest Closes redis#639
…icts (redis#637) Added @ConditionalOnMissingBean to the default redisEnhancedMappingContext bean configuration, allowing users to provide their own custom mapping context with a custom keyspace resolver without encountering bean definition override issues. The fix maintains @primary on the default bean for backward compatibility while allowing users to override it by providing their own bean with the same name and @primary annotation. Example usage: ```java @bean(name = "redisEnhancedMappingContext") @primary public RedisEnhancedMappingContext customMappingContext() { RedisEnhancedMappingContext context = new RedisEnhancedMappingContext(); context.setKeySpaceResolver(type -> "tenant:prefix:" + type.getSimpleName()); return context; } ``` - Added @ConditionalOnMissingBean(name = "redisEnhancedMappingContext") - Maintained @primary for backward compatibility - Added comprehensive test demonstrating custom keyspace resolver - Verified backward compatibility with existing tests Closes redis#637
Enhanced the indexExistsFor method in RediSearchIndexer to handle different error messages across Redis versions when checking for non-existent indexes. The method now recognizes multiple error message patterns: - "Unknown index name" - Redis Stack / Redis 7.x - "no such index" - Redis 8.0 (reported in issue) - "index does not exist" - Alternative format - "not found" - Generic pattern Also fixed NPE when checking existence for unmapped entity classes by adding null check for index name. - Made error detection case-insensitive for robustness - Added comprehensive test coverage for different error patterns Closes redis#636 style: apply code formatting
…xception (redis#622) Both RediSearchQuery and RedisEnhancedQuery were returning entity objects for existsBy* query methods, causing ClassCastException when the repository method expected a boolean return type. Modified executeQuery methods in both query classes to detect when the return type is boolean/Boolean and return the result of checking if any documents were found (searchResult.getTotalResults() > 0) instead of attempting to return the entity object. Added tests to verify existsBy* methods work correctly for both @document and @RedisHash entities.
) Added throwOnSaveAllFailure configuration property to control whether saveAll operations throw exceptions on failures or just log warnings. Changes: - Added redis.om.spring.repository.throw-on-save-all-failure property (defaults to false for backward compatibility) - Modified SimpleRedisEnhancedRepository to use syncAndReturnAll() instead of sync() to capture operation responses - Both repositories now check pipeline responses for JedisDataException and either throw RuntimeException or log warnings based on config - Added tests to verify saveAll error handling works correctly This allows applications to detect and handle bulk save failures programmatically when Redis encounters issues like memory limits.
…different types (redis#342) Added new methods to SearchFieldPredicate interface to allow combining predicates with different field types, solving the generic type incompatibility issue when building complex queries. The existing and() and or() methods require compatible types due to Java's type system constraints. The new andAny() and orAny() methods use wildcards to allow combining predicates regardless of their field types, enabling queries like: - NAME_SPACE.eq("PERSONAL").andAny(RELATE_ID.eq(100L))
- Add detailed Jedis pool configuration documentation - Include examples for property-based and Java-based configuration - Document JedisClientConfigurationBuilderCustomizer approach - Add pool monitoring examples and common configuration scenarios - Provide imports for all code examples to improve clarity
…entities (redis#595) - Document the existing partial update functionality for JSON documents - Add examples for single entity updates using Query by Example - Include bulk update operations and ExampleMatcher usage - Explain how JSON.SET with path specifications enables atomic field updates - Clarify that this feature is already implemented and tested
- Update early-access workflow to handle Release Candidate (RC) versions - Change version check from SNAPSHOT-only to include RC versions - Configure JReleaser to mark RC and SNAPSHOT versions as prereleases - Fix workflow conditional to use EARLY_ACCESS flag instead of SNAPSHOT
…ployment (redis#625) - Modified early-access workflow to automatically append -SNAPSHOT to versions - Skip snapshot releases for already-released versions (with git tags) - Skip snapshot releases for RC and existing SNAPSHOT versions - Configure JReleaser to mark SNAPSHOT versions as prereleases - Enable continuous deployment for every push to main branch Fixes redis#625
- Add ClassLoaderAwareKeyspaceResolver to handle cross-class-loader entity lookups - Store keyspace settings by class name for compatibility with RestartClassLoader - Update SimpleRedisDocumentRepository and RedisJSONKeyValueAdapter to use resolver - Ensure TTL settings work correctly in @PostConstruct and when devtools is active The issue occurred because spring-boot-devtools uses a different class loader (RestartClassLoader) which caused entity classes to not be recognized by the KeyspaceConfiguration when comparing Class instances directly. This fix stores and retrieves settings by fully qualified class name as a fallback. Fixes redis#458
Implement full Map<String, T> field support for Redis JSON documents with: - Support for all value types: String, Boolean, numeric types, temporal types, UUID, Ulid, enums, and Point - Automatic indexing using JSONPath expressions ($.fieldName.*) - Repository query methods using MapContains pattern - EntityStream support for Map field queries - Metamodel generation for Map VALUES fields - Custom Gson serializers for Boolean values (1/0 format) - Comprehensive test coverage for all Map value types - Complete documentation with examples and best practices Note: Boolean values in Map fields are indexed as NUMERIC fields for consistency with Redis JSON serialization
Fix documentation that incorrectly described Redis OM Spring hash mapping as using Spring Data Redis' basic Set-based indexing. Redis OM Spring actually creates RediSearch indexes using FT.CREATE commands for Hash entities. Key corrections: - Clarify that Redis OM Spring uses RediSearch indexes, not Redis Sets - Add comparison table showing differences from Spring Data Redis - Show actual Redis commands (FT.CREATE ON HASH) generated - Explain O(log N) index-based query performance vs O(N) scans - Add migration guide from Spring Data Redis - Update examples to show RediSearch field types (TEXT, TAG, NUMERIC, GEO, VECTOR)
Fix JReleaser configuration to show actual commit authors instead of incorrectly parsed Java annotations as contributors. Issues fixed: - Enable contributor extraction from git history (was disabled) - Escape Java annotations (@reference, @RedisKey, etc.) in changelog to prevent GitHub from treating them as user mentions - Filter out bot accounts (GitHub Actions) from contributor list - Add regex replacer to wrap all @annotations in backticks This ensures releases show real contributors instead of fake users like "reference" and "RedisKey" that were being parsed from annotation mentions in the changelog.
Fix ClassCastException when retrieving document count from Redis index info. Redis can return num_docs as either String or Long depending on version/config. Applied robust type checking to all occurrences: - RedisJSONKeyValueAdapter.count() - RedisEnhancedKeyValueAdapter.count() - RedisFluentQueryByExample (already fixed in SearchStreamImpl) The fix handles: - String values: parse to Long - Number values: convert to long - Null/unexpected: return 0L safely This is more robust than PR redis#640's approach as it: - Handles multiple Redis return types - Avoids NullPointerException - Prevents unnecessary String conversions - Provides safe fallback for unexpected types
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
No description provided.