CodeDweller/roller.hpp

// roller.hpp
// Copyright (C) 2022 MicroNeil Research Corporation.
//
// This software is released under the MIT license. See LICENSE.TXT.
//
// Roller is a naive rolling hash system for rapid string searches.
// Roller32 accommodates matches up to 4 bytes.
// Roller64 accommodates matches up to 8 bytes.

#pragma once

#include <cstdint>
#include <string>
#include <vector>

namespace codedweller {

    class Roller32 {
        private:
            uint_fast32_t roller;

        public:
            Roller32() : roller(0) {}
            uint_fast32_t value() const { return roller; }
            uint_fast32_t add(uint8_t byte) { 
                roller = roller << 8 | byte;
                return roller;
            }
    };

    class Roller64 {
        private:
            uint_fast64_t roller;

        public:
            Roller64() : roller(0) {}
            uint_fast64_t value() const { return roller; }
            uint_fast64_t add(uint8_t byte) {
                roller = roller << 8 | byte;
                return roller;
            }
    };

    class RollerMatch32 {
        private:
            uint_fast32_t match;
            uint_fast32_t mask;

        public:
            RollerMatch32(const std::vector<unsigned char> pattern) {
                match = mask = 0;
                size_t ingest = std::min(sizeof(uint_fast32_t), pattern.size());
                Roller32 matcher;
                Roller32 masker;
                for(size_t count = 0; count < ingest; count++) { 
                    matcher.add(pattern.at(count));
                    masker.add(0xFF);
                }
                match = matcher.value();
                mask = masker.value();
            }

            RollerMatch32(const std::string pattern) {
                match = mask = 0;
                size_t ingest = std::min(sizeof(uint_fast32_t), pattern.size());
                Roller32 matcher;
                Roller32 masker;
                for(size_t count = 0; count < ingest; count++) { 
                    matcher.add(pattern.at(count));
                    masker.add(0xFF);
                }
                match = matcher.value();
                mask = masker.value();
            }

            bool matches(const Roller32 roller) {
                return (match == (roller.value() & mask));
            }

    };

    class RollerMatch64 {
        private:
            uint_fast64_t match;
            uint_fast64_t mask;

        public:
            RollerMatch64(const std::vector<unsigned char> pattern) {
                match = mask = 0;
                size_t ingest = std::min(sizeof(uint_fast64_t), pattern.size());
                Roller64 matcher;
                Roller64 masker;
                for(size_t count = 0; count < ingest; count++) { 
                    matcher.add(pattern.at(count));
                    masker.add(0xFF);
                }
                match = matcher.value();
                mask = masker.value();
            }

            RollerMatch64(const std::string pattern) {
                match = mask = 0;
                size_t ingest = std::min(sizeof(uint_fast64_t), pattern.size());
                Roller64 matcher;
                Roller64 masker;
                for(size_t count = 0; count < ingest; count++) { 
                    matcher.add(pattern.at(count));
                    masker.add(0xFF);
                }
                match = matcher.value();
                mask = masker.value();
            }

            bool matches(const Roller64 roller) {
                return (match == (roller.value() & mask));
            }
    };


}