Improved ESP-index: A Practical Self-index for Highly Repetitive Texts

Abstract

Abstract. While several self-indexes for highly repetitive texts exist, developing a practical self-index applicable to real world repetitive texts remains a challenge. ESP-index is a grammar-based self-index on the notion of edit-sensitive parsing (ESP), an efficient parsing algorithm that guarantees upper bounds of parsing discrepancies between different appearances of the same subtexts in a text. Although ESP-index performs efficient top-down searches of query texts, it has a serious issue on binary searches for finding appearances of variables for a query text, which resulted in slowing down the query searches. We present an improved ESP-index (ESP-index-I) by leveraging the idea behind succinct data structures for large alphabets. While ESP-index-I keeps the same types of efficiencies as ESP-index about the top-down searches, it avoid the binary searches using fast rank/select operations. We experimentally test ESP-index-I on the ability to search query texts and extract subtexts from real world repetitive texts on a large-scale, and we show that ESPindex-I performs better that other possible approaches.

A Dynamic Compressed Self-Index for Highly Repetitive Text Collections

“…We performed all the experiments on one core of a quad-core Intel(R) Xeon(R) E5-2680 v2 (2.80 GHz) CPU with 256 GB of memory. We compared our TST-index with the ESP-index [18] and the RLFM-index [7]. The ESP-index provides a baseline for evaluating the effectiveness of the TST-index, while the RLFM-index is a stateof-the-art self-index for highly repetitive text collections.…”

“…Table 1: Summary of self-indexes for highly repetitive text collections. Here, occ c ≥ occ is the number of candidate occurrences of a given pattern as obtained by the ESP-index [18].…”

“…We can extend signature encoding as a self-index for pattern search. A key idea for pattern search is that there exists a common variable sequence of length O(log |P | log * M ) representing every substring P in the parse tree of G. This sequence is called the core of P and is computable from P , as described elsewhere [18,15] in detail.…”

“…, 20}, D = {1 → A, 2 → B, 3 → 1 1 , 4 → 2 1 , 5 → 2 2 , 6 → 1 2 , 7 → (3, 4), 8 → (3, 5, 3), 9 → (4, 3), 10 → (4, 3, 4, 6), 11 → 7 3 , 12 → 8 1 , 13 → 9 1 , 14 → 10 1 , 15 → 9 3 , 16 → (11, 12, 13), 17 → (14, 15), 18 → 16 1 , 19 → 17 1 , 20 → (18, 19)}, and S = 20. Assume that LC 4N (SE T 1 ) = (3, 4) 3 , (3, 5, 3), (4, 3), (4, 3, 4, 6), (4, 3) 3 , LC 4N (SE T 3 ) =(11,12,13),(14,15), and LC 4N (SE T 5 ) =(18,19) hold. Then G is the signature encoding of T = ABABABABBAABABABAABABABA.…”

“…Takabatake et al[18] construct the grammar representing an input text by using a technique called alphabet reduction, which is essentially equal to LCPR. Although the definition of their grammar is slightly different from ours, we can use their search algorithm with our signature encoding.…”