emory/tools/compare.cpp

/*
 * This Source Code Form is subject to the terms of the Mozilla Public
 * License, v. 2.0. If a copy of the MPL was not distributed with this
 * file, You can obtain one at http://mozilla.org/MPL/2.0/.
 *
 * Copyright 2013 Danny Robson <danny@nerdcruft.net>
 */

#include "emory/chunk/params.hpp"
#include "emory/chunk/find.hpp"
#include "emory/chunk/region.hpp"
#include "emory/chunk/match.hpp"
#include "emory/chunk/ostream.hpp"

#include <cruft/util/io.hpp>
#include <cruft/util/parse/value.hpp>

#include <cstddef>
#include <iostream>


enum {
    ARG_SELF,
    ARG_BITS,
    ARG_WINDOW,
    ARGS_MINIMUM,
    ARGS_MAXIMUM,
    ARGS_TARGET,
    ARGS_SOURCE,

    NUM_ARGS,
};


int main (int argc, char const **argv)
{
    if (argc < NUM_ARGS) {
        std::cerr << "usage: " << argv[ARG_SELF] << " <bits> <window> <minimum> <maximum> <target> <source> [...]\n";
        return EXIT_FAILURE;
    }

    emory::chunk::params const p {
        .bits    = cruft::parse::from_string<std::size_t> (argv[ARG_BITS  ]),
        .window  = cruft::parse::from_string<std::size_t> (argv[ARG_WINDOW]),
        .minimum = cruft::parse::from_string<std::ptrdiff_t> (argv[ARGS_MINIMUM]),
        .maximum = cruft::parse::from_string<std::ptrdiff_t> (argv[ARGS_MAXIMUM]),
    };

    std::clog << "Hashing target\n";
    std::vector<emory::chunk::region> target;
    emory::chunk::find<emory::chunk::static_hash> (
        std::back_inserter (target),
        cruft::mapped_file (argv[ARGS_TARGET]), p
    );

    std::sort (target.begin (), target.end (), emory::chunk::region::digest_ordering {});
    std::cout << "Found " << target.size () << " chunks\n";

    std::vector<emory::chunk::match<int>> found;

    for (int i = ARGS_SOURCE; i != argc; ++i) {
        auto const path = argv[i];
        std::clog << "Hashing source: " << path << '\n';
        std::vector<emory::chunk::region> source;
        emory::chunk::find<emory::chunk::static_hash> (
            std::back_inserter (source),
            cruft::mapped_file (path), p
        );
        std::sort (source.begin (), source.end (), emory::chunk::region::digest_ordering {});

        std::clog << "Finding common\n";
        auto const &source_matches = common (target, source);
        std::clog << "Discovered " << source_matches.size () << " blocks\n";

        std::transform (
            std::begin (source_matches),
            std::end   (source_matches),
            std::back_inserter (found),
            [&] (auto const &j) -> emory::chunk::match<int>
        {
            return {
                .src = {           i, j.src.data },
                .dst = { ARGS_TARGET, j.dst.data },
            };
        });

        std::sort (
            std::begin (found),
            std::end   (found),
            emory::chunk::match<int>::src_ordering {}
        );

        found.erase (
            std::unique (
                std::begin (found),
                std::end   (found),
                emory::chunk::match<int>::dst_equality {}
            ),
            found.end ()
        );
    }

    std::clog << "Finalising\n";
    std::size_t matching = 0;
    std::vector<std::size_t> source_bytes (argc - ARGS_SOURCE, 0);
    for (auto const &i: found) {
        std::cout << i << '\n';
        auto const size = i.dst.data.offset.second - i.dst.data.offset.first;
        matching += size;
        source_bytes[i.src.id - ARGS_SOURCE] += size;
    }

    std::size_t const total = std::accumulate (
        target.begin (),
        target.end (),
        0u,
        [] (auto const &a, auto const &b) { return a + b.offset.second - b.offset.first; }
    );

    std::cout << "Found " << found.size () << " chunks. " << matching << "/" << total << " bytes for a factor of " << float (matching) / total << "\n";

    for (int i = ARGS_SOURCE; i != argc; ++i)
        std::cerr << argv[i] << " contributed: " << source_bytes[i - ARGS_SOURCE] << '\n';
}