Make UCDetect a little less insane
This commit is contained in:
parent
2dd1da8333
commit
47c678bd63
5 changed files with 31 additions and 43 deletions
|
@ -25,8 +25,8 @@ std::string Detect(const std::string &file) {
|
|||
return UCDetect(file).Single();
|
||||
}
|
||||
|
||||
void DetectAll(const std::string& file, CharsetListDetected &list) {
|
||||
UCDetect(file).List(list);
|
||||
CharsetListDetected DetectAll(const std::string& file) {
|
||||
return UCDetect(file).List();
|
||||
}
|
||||
|
||||
} // namespace util
|
||||
|
|
|
@ -23,11 +23,14 @@
|
|||
|
||||
#include "../../universalchardet/nsCharSetProber.h"
|
||||
|
||||
#include <boost/range/algorithm.hpp>
|
||||
|
||||
namespace agi {
|
||||
namespace charset {
|
||||
|
||||
UCDetect::UCDetect(const std::string &file): nsUniversalDetector(NS_FILTER_ALL) {
|
||||
UCDetect::UCDetect(const std::string &file)
|
||||
: nsUniversalDetector(NS_FILTER_ALL)
|
||||
{
|
||||
{
|
||||
agi::scoped_ptr<std::ifstream> fp(io::Open(file, true));
|
||||
|
||||
|
@ -35,7 +38,7 @@ UCDetect::UCDetect(const std::string &file): nsUniversalDetector(NS_FILTER_ALL)
|
|||
// be able to do anything useful with it anyway
|
||||
fp->seekg(0, std::ios::end);
|
||||
if (fp->tellg() > 100 * 1024 * 1024) {
|
||||
list.insert(CLDPair(1.f, "binary"));
|
||||
list.emplace_back(1.f, "binary");
|
||||
return;
|
||||
}
|
||||
fp->seekg(0, std::ios::beg);
|
||||
|
@ -58,7 +61,7 @@ UCDetect::UCDetect(const std::string &file): nsUniversalDetector(NS_FILTER_ALL)
|
|||
}
|
||||
|
||||
if (binaryish > bytes / 8) {
|
||||
list.insert(CLDPair(1.f, "binary"));
|
||||
list.emplace_back(1.f, "binary");
|
||||
return;
|
||||
}
|
||||
}
|
||||
|
@ -67,49 +70,40 @@ UCDetect::UCDetect(const std::string &file): nsUniversalDetector(NS_FILTER_ALL)
|
|||
|
||||
DataEnd();
|
||||
|
||||
if (mDetectedCharset) {
|
||||
list.insert(CLDPair(1.f, mDetectedCharset));
|
||||
} else {
|
||||
|
||||
if (mDetectedCharset)
|
||||
list.emplace_back(1.f, mDetectedCharset);
|
||||
else {
|
||||
switch (mInputState) {
|
||||
case eHighbyte: {
|
||||
for (PRInt32 i=0; i<NUM_OF_CHARSET_PROBERS; i++) {
|
||||
if (mCharSetProbers[i]) {
|
||||
float conf = mCharSetProbers[i]->GetConfidence();
|
||||
if (conf > 0.01f) {
|
||||
list.insert(CLDPair(conf, mCharSetProbers[i]->GetCharSetName()));
|
||||
}
|
||||
}
|
||||
if (!mCharSetProbers[i]) continue;
|
||||
|
||||
float conf = mCharSetProbers[i]->GetConfidence();
|
||||
if (conf > 0.01f)
|
||||
list.emplace_back(conf, mCharSetProbers[i]->GetCharSetName());
|
||||
}
|
||||
|
||||
break;
|
||||
}
|
||||
case ePureAscii:
|
||||
list.insert(CLDPair(1.f, "US-ASCII"));
|
||||
list.emplace_back(1.f, "US-ASCII");
|
||||
break;
|
||||
|
||||
default:
|
||||
throw UnknownCharset("Unknown chararacter set.");
|
||||
throw UnknownCharset("Unknown character set.");
|
||||
}
|
||||
|
||||
if (list.empty() && (mInputState == eHighbyte))
|
||||
throw UnknownCharset("Unknown chararacter set.");
|
||||
throw UnknownCharset("Unknown character set.");
|
||||
|
||||
|
||||
} // if mDetectedCharset else
|
||||
}
|
||||
|
||||
std::string UCDetect::Single() {
|
||||
/// @todo Add a debug log here since this shouldn't happen.
|
||||
if (list.empty()) {
|
||||
throw UnknownCharset("Unknown chararacter set.");
|
||||
typedef std::pair<float, std::string> const& result;
|
||||
boost::sort(list, [](result lft, result rgt) { return lft.first > rgt.first; });
|
||||
}
|
||||
|
||||
CharsetListDetected::const_iterator i_lst = list.begin();
|
||||
return i_lst->second;
|
||||
}
|
||||
|
||||
|
||||
std::string UCDetect::Single() const {
|
||||
return list.front().second;
|
||||
}
|
||||
|
||||
} // namespace util
|
||||
} // namespace agi
|
||||
|
|
|
@ -34,10 +34,6 @@ namespace agi {
|
|||
namespace charset {
|
||||
|
||||
class UCDetect : public nsUniversalDetector {
|
||||
|
||||
/// For insertion into CharsetListDetected
|
||||
typedef std::pair<float, std::string> CLDPair;
|
||||
|
||||
/// List of detected character sets.
|
||||
CharsetListDetected list;
|
||||
|
||||
|
@ -51,12 +47,11 @@ public:
|
|||
UCDetect(const std::string &file);
|
||||
|
||||
/// @brief Detect character set of a file using UniversalCharDet
|
||||
/// @param out[out] Map to load list into ordered by confidence
|
||||
void List(CharsetListDetected &out) { out = list; }
|
||||
CharsetListDetected List() const { return list; }
|
||||
|
||||
/// @brief Return a single character set (highest confidence)
|
||||
/// @return Character set
|
||||
std::string Single();
|
||||
std::string Single() const;
|
||||
};
|
||||
|
||||
} // namespace util
|
||||
|
|
|
@ -19,8 +19,8 @@
|
|||
#ifndef LAGI_PRE
|
||||
#include <fstream>
|
||||
#include <functional>
|
||||
#include <map>
|
||||
#include <string>
|
||||
#include <vector>
|
||||
#endif
|
||||
#include <libaegisub/exception.h>
|
||||
|
||||
|
@ -32,18 +32,17 @@ DEFINE_BASE_EXCEPTION_NOINNER(CharsetError, agi::Exception)
|
|||
DEFINE_SIMPLE_EXCEPTION_NOINNER(UnknownCharset, CharsetError, "charset/unknown")
|
||||
|
||||
/// List of detected encodings.
|
||||
typedef std::map<float, std::string, std::greater_equal<float> > CharsetListDetected;
|
||||
typedef std::vector<std::pair<float, std::string>> CharsetListDetected;
|
||||
|
||||
/// @brief Return a complete list of detected character sets ordered by precedence.
|
||||
/// @param file File to check
|
||||
/// @param[out] list Map to load detected list into.
|
||||
void DetectAll(const std::string &file, CharsetListDetected &list);
|
||||
/// @return List of possible charsets sorted by probability
|
||||
CharsetListDetected DetectAll(std::string const& file);
|
||||
|
||||
/// @brief Returns the character set with the highest confidence
|
||||
/// @param file File to check
|
||||
/// @return Detected character set.
|
||||
std::string Detect(const std::string &file);
|
||||
|
||||
|
||||
} // namespace util
|
||||
} // namespace agi
|
||||
|
|
|
@ -55,7 +55,7 @@ wxString GetEncoding(wxString const& filename) {
|
|||
agi::charset::CharsetListDetected list;
|
||||
|
||||
try {
|
||||
agi::charset::DetectAll(from_wx(filename), list);
|
||||
list = agi::charset::DetectAll(from_wx(filename));
|
||||
} catch (const agi::charset::UnknownCharset&) {
|
||||
/// @todo If the charset is unknown we need to display a complete list of character sets.
|
||||
}
|
||||
|
|
Loading…
Reference in a new issue