Public-Mirror/Textractor
1
0
Fork 1
mirror of https://github.com/Artikash/Textractor.git synced 2024-12-23 17:04:12 +08:00
Code Issues Projects Releases Wiki Activity

upgrade repetition remover and add tests

Browse Source
This commit is contained in:
Akash Mozumdar 2018-12-28 11:57:06 -05:00
parent bdc083a62a
commit 54c1b508d5
3 changed files with 58 additions and 14 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
extensions/CMakeLists.txt
View File

@ -12,6 +12,8 @@ add_library(Google\ Translate SHARED googletranslate.cpp extensionimpl.cpp)
add_library(Regex\ Filter SHARED regexfilter.cpp extensionimpl.cpp)
add_library(Remove\ Repetition SHARED removerepeat.cpp extensionimpl.cpp)
add_executable(Extension_Tests extensiontester.cpp)
target_link_libraries(Bing\ Translate winhttp Qt5::Widgets)
target_link_libraries(Extra\ Window Qt5::Widgets)
target_link_libraries(Google\ Translate winhttp Qt5::Widgets)

11
extensions/extensiontester.cpp Normal file
View File

@ -0,0 +1,11 @@
#include "common.h"
#include <filesystem>
int main()
{
wchar_t path[MAX_PATH] = {};
GetModuleFileNameW(NULL, path, MAX_PATH);
*(wcsrchr(path, L'\\') + 1) = 0;
for (auto file : std::filesystem::directory_iterator(path))
if (file.path().extension() == L".dll") LoadLibraryW(file.path().c_str());
}

59
extensions/removerepeat.cpp
View File

@ -1,22 +1,34 @@
#include "extension.h"
#include "defs.h"
void RemoveRepeatedChars(std::wstring& sentence)
{
int repeatNumber = 0;
wchar_t prevChar = sentence[0];
for (auto c : sentence)
if (c == prevChar) repeatNumber++;
else break;
if (repeatNumber == 1) return;
std::vector<int> repeatNumbers(sentence.size() + 1, 0);
int repeatNumber = 1;
wchar_t prevChar = L'\0';
for (auto nextChar : sentence)
if (nextChar == prevChar) repeatNumber++;
else
{
prevChar = nextChar;
++repeatNumbers.at(repeatNumber);
repeatNumber = 1;
}
if ((repeatNumber = std::distance(repeatNumbers.begin(), std::max_element(repeatNumbers.begin(), repeatNumbers.end()))) == 1) return;
for (int i = 0; i < sentence.size(); i += repeatNumber)
for (int j = i; j < sentence.size(); ++j)
if (sentence[j] != sentence[i])
if ((j - i) % repeatNumber != 0) return;
else break;
std::wstring newSentence = L"";
for (int i = 0; i < sentence.size(); i += repeatNumber) newSentence.push_back(sentence[i]);
std::wstring newSentence;
for (int i = 0; i < sentence.size();)
{
newSentence.push_back(sentence.at(i));
for (int j = i; j <= sentence.size(); ++j)
{
if (j == sentence.size() || sentence.at(i) != sentence.at(j))
{
i += (j - i) % repeatNumber == 0 ? repeatNumber : 1;
break;
}
}
}
sentence = newSentence;
}
@ -38,3 +50,22 @@ bool ProcessSentence(std::wstring& sentence, SentenceInfo sentenceInfo)
RemoveCyclicRepeats(sentence);
return true;
}
TEST(
{
std::wstring repeatedChars = L"aaaaaaaaaaaabbbbbbcccdddaabbbcccddd";
RemoveRepeatedChars(repeatedChars);
assert(repeatedChars.find(L"aaaabbcd") == 0);
std::wstring cyclicRepeats = L"abcdeabcdefabcdefgabcdefgabcdefgabcdefgabcdefg";
RemoveCyclicRepeats(cyclicRepeats);
assert(cyclicRepeats == L"abcdefg");
InfoForExtension tester{ "hook address", 0, nullptr };
std::wstring empty = L"", one = L" ", normal = L"This is a normal sentence. ‚Í‚¢<E2809A>B";
ProcessSentence(empty, { &tester });
ProcessSentence(one, { &tester });
ProcessSentence(normal, { &tester });
assert(empty == L"" && one == L" " && normal == L"This is a normal sentence. ‚Í‚¢<E2809A>B");
}
);