Infidex

Infidex Search Engine A high-performance, hybrid search engine that combines TF-IDF vector space retrieval with multi-algorithm lexical matching. Features Two-Stage Architecture: Combines relevancy ranking (Stage 1) with coverage matching (Stage 2) TF-IDF Vector Model: Classic information retrieval with L2-normalized vectors Multi-Size N-grams: Character n-grams (2-3 or configurable) for morphological robustness Byte Quantization: 4x memory savings by storing weights as bytes (0-255) Bucket-Based Sorting: O(1) insertion and O(n) top-K retrieval using score buckets 5 Lexical Matching Algorithms: Exact word matching Fuzzy matching (Levenshtein distance ≤ 1) Joined/split word detection Prefix/suffix matching LCS (Longest Common Subsequence) Bit-Parallel Levenshtein: Ultra-fast edit distance for strings ≤64 characters