fix(search): fatals on invalid short wildcard terms

Searches such as "wiki a*" could crash in QueryEvaluator::opAnd() with
a TypeError. The query parser still emitted the wildcard term, but the
full

fix(search): fatals on invalid short wildcard terms

Searches such as "wiki a*" could crash in QueryEvaluator::opAnd() with
a TypeError. The query parser still emitted the wildcard term, but the
fulltext search later dropped it because its non-wildcard base was below
the minimum search term length.

That left the evaluator with an AND operator whose right-hand operand was
missing, causing a stack underflow during RPN evaluation.

Fix this in two places:
- filter invalid short wildcard terms already in Tokenizer::getWords()
when wildcard parsing is enabled, so they never enter the parsed query
- harden QueryEvaluator against missing operands for AND/OR/NOT to avoid
fatals if tokens disappear during later processing

Add regression tests covering the parser output for "wiki a*" and the
evaluator behavior when an operand is missing.

show more ...

SearchIndex: rector, phpcs, type hint fixes

SearchIndex: rewrite MetadataSearch to use Collection classes

Replace MetadataIndex usage in MetadataSearch with the new Collection/Index
architecture. This completes the read-path migration so data

SearchIndex: rewrite MetadataSearch to use Collection classes

Replace MetadataIndex usage in MetadataSearch with the new Collection/Index
architecture. This completes the read-path migration so data written by the
Collection-based Indexer is read back correctly using TupleOps tuple format.

Generalize FrequencyCollectionSearch into CollectionSearch that works with any
AbstractCollection type (Frequency, Lookup, Direct) and handles both
split-by-length and non-split index layouts transparently. DirectCollection
participates via resolveTokenFrequencies() which maps token RID = entity RID.

Key changes:
- AbstractCollection gains isSplitByLength(), resolveTokenFrequencies(),
getEntitiesWithData(), and groupToSuffix() with validation
- Index groups are now int (0 = non-split, positive = token length)
- CollectionSearch provides both addTerm()/execute() for fulltext and
lookup() for metadata-style search (exact/wildcard/callback)
- MetadataSearch delegates entirely to collection APIs
- Shared filterPages() replaces duplicated page filtering logic
- All callers updated from MetadataIndex to MetadataSearch
- Tests moved to Search namespace with full coverage for new APIs

show more ...

SearchIndex: rewrite FulltextSearch to use FrequencyCollectionSearch

Replace FulltextIndex->lookupWords() with FrequencyCollectionSearch which
correctly handles the compact tuple format written by t

SearchIndex: rewrite FulltextSearch to use FrequencyCollectionSearch

Replace FulltextIndex->lookupWords() with FrequencyCollectionSearch which
correctly handles the compact tuple format written by the new Indexer.

Introduce QueryEvaluator with typed stack entries (PageSet,
NamespacePredicate, NegatedEntry) for RPN query evaluation. NOT wraps
its operand instead of computing a universe complement, so AND with a
negated operand becomes efficient set subtraction. The full page index
is only loaded for standalone negative or namespace-only queries.

Move QueryParser and QueryEvaluator into the new Search\Query namespace
along with the stack entry types.

Simplify FulltextSearch to orchestration: parse query, look up words,
evaluate, filter, sort. Replace FT_SNIPPET_NUMBER constant with
maxSnippets property. Combine ACL/existence/time filtering into a
single pass.

show more ...