phundrak/lzw-assignment
1
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity

Optimized compression function

Browse Source
This commit is contained in:
Phuntsok Drak-pa 2018-04-03 16:55:25 +02:00
parent 3ce2754211
commit 3eeb2bed7c
1 changed files with 25 additions and 13 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

38
src/main.cc
View File

@ -38,9 +38,9 @@ using uvec = std::vector<uint32_t>; // chaine encodée
size_t size_str(dic_t &dict, uint32_t nr_chaine) { size_t size_str(dic_t &dict, uint32_t nr_chaine) {
for(const auto &[key, value] : dict) { for(const auto &[key, value] : dict) {
if(value == nr_chaine) { if(value == nr_chaine) {
if(key.first == 0xFFFFFFFF) if(std::get<0>(key) == 0xFFFFFFFF)
return 1; return 1;
return 1 + size_str(dict, key.first); return 1 + size_str(dict, std::get<0>(key));
} }
} }
return 0; return 0;
@ -77,6 +77,14 @@ std::pair<bool, uint32_t> dico(dic_t &dictionary, uint32_t nr_chaine, uint8_t c)
typename std::remove_reference<decltype(e)>::type>( typename std::remove_reference<decltype(e)>::type>(
dictionary.size()) + dictionary.size()) +
255)); 255));
// auto &e = dictionary[std::make_tuple(nr_chaine, c, len)];
// return (e) ? std::make_pair(true, e)
// : std::make_pair(
// false,
// (e = static_cast<
// typename std::remove_reference<decltype(e)>::type>(
// dictionary.size()) +
// 255));
} }
/** /**
@ -93,19 +101,32 @@ std::pair<bool, uint32_t> dico(dic_t &dictionary, uint32_t nr_chaine, uint8_t c)
* \return std::vector<uint16_t> * \return std::vector<uint16_t>
*/ */
const uvec compress(const ustring &text, dic_t &dictionary) { const uvec compress(const ustring &text, dic_t &dictionary) {
std::puts("Compressing...");
uvec res{}; uvec res{};
uint32_t w = 0xFFFFFFFF; uint32_t w = 0xFFFFFFFF;
uint32_t len = 0;
#ifdef Debug
size_t progress = 0;
#endif
for (const auto &c : text) { for (const auto &c : text) {
if(size_str(dictionary , w) > 16) { ++len;
#ifdef Debug
printf("\rprogress: %zu / %zu", progress++, text.size());
#endif
if (len > 9) {
res.push_back(static_cast<uint32_t>(w)); res.push_back(static_cast<uint32_t>(w));
w = c; w = c;
} else if (const auto &[exists, pos] = dico(dictionary, w, c); exists) { len = 0;
} else if (const auto &[exists, pos] = dico(dictionary, w, c);
exists) {
w = pos; w = pos;
} else { } else {
res.push_back(static_cast<uint32_t>(w)); res.push_back(static_cast<uint32_t>(w));
w = c; w = c;
len = 0;
} }
} }
printf("\n");
return res; return res;
} }
@ -201,15 +222,6 @@ int main(int argc, char *argv[]) {
printf("Compression ratio: %.10f\n", printf("Compression ratio: %.10f\n",
static_cast<double>(str.size()) / static_cast<double>(comp_str.size())); static_cast<double>(str.size()) / static_cast<double>(comp_str.size()));
#ifdef Debug
printf("// compressed file: hex "
"/////////////////////////////////////////////\n");
for (size_t i = 0; i < comp_str.size(); ++i)
printf("%04x%s", str[i],
(((i + 1) % 16 == 0) ? "\n" : ((i + 1) % 2 == 0 ? " " : "")));
printf("\n\n");
#endif
printf("Number of custom words in the dictionary: %zu\n", dictionary.size()); printf("Number of custom words in the dictionary: %zu\n", dictionary.size());
for(const auto c : comp_str) for(const auto c : comp_str)