google/sentencepiece
1
1
Fork 0
mirror of https://github.com/google/sentencepiece.git synced 2024-08-16 06:10:47 +03:00
Code Issues Packages Projects Releases Wiki Activity

Introduced new Filesystem API phase 2.

Browse Source
This commit is contained in:
Taku Kudo 2018-09-01 17:13:55 +09:00
parent f617fcb697
commit e9b3073703
5 changed files with 27 additions and 20 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

1
.gitignore vendored
View File

@ -5,6 +5,7 @@ Makefile.in
/py-compile
/test-driver
/ylwrap
/build
/autom4te.cache
/autoscan.log

6
src/common.h
View File

@ -64,12 +64,6 @@ static constexpr int64 kint64max = ((int64)(0x7FFFFFFFFFFFFFFF));
static constexpr uint32 kUnicodeError = 0xFFFD;
#ifdef OS_WIN
#define OUTPUT_MODE std::ios::binary | std::ios::out
#else
#define OUTPUT_MODE std::ios::out
#endif
#if defined(OS_WIN) && defined(UNICODE) && defined(_UNICODE)
#define WPATH(path) (::sentencepiece::win32::Utf8ToWide(path).c_str())
#else

20
src/filesystem.cc
View File

@ -13,8 +13,16 @@
// limitations under the License.!
#include "filesystem.h"
#include <iostream>
#include "util.h"
#ifdef OS_WIN
#define OUTPUT_MODE std::ios::binary | std::ios::out
#else
#define OUTPUT_MODE std::ios::out
#endif
namespace sentencepiece {
namespace filesystem {
@ -38,7 +46,15 @@ class PosixReadableFile : public ReadableFile {
return static_cast<bool>(std::getline(*is_, *line));
}
bool ReadAll(std::string *line) { return true; }
bool ReadAll(std::string *line) {
if (is_ == &std::cin) {
LOG(ERROR) << "ReadAll is not supported for stdin.";
return false;
}
line->assign(std::istreambuf_iterator<char>(*is_),
std::istreambuf_iterator<char>());
return true;
}
private:
util::Status status_;
@ -72,7 +88,7 @@ class PosixWritableFile : public WritableFile {
private:
util::Status status_;
std::ostream *os_;
}; // namespace filesystem
};
std::unique_ptr<ReadableFile> NewReadableFile(absl::string_view filename) {
return port::MakeUnique<PosixReadableFile>(filename);

19
src/sentencepiece_processor.cc
View File

@ -43,13 +43,11 @@ SentencePieceProcessor::SentencePieceProcessor() {}
SentencePieceProcessor::~SentencePieceProcessor() {}
util::Status SentencePieceProcessor::Load(util::min_string_view filename) {
std::ifstream ifs(WPATH(filename.data()), std::ios::binary | std::ios::in);
if (!ifs) {
return util::StatusBuilder(util::error::NOT_FOUND)
<< "\"" << filename.data() << "\": " << util::StrError(errno);
}
return Load(&ifs);
auto input = filesystem::NewReadableFile(string_util::ToSV(filename));
RETURN_IF_ERROR(input->status());
std::string proto;
CHECK_OR_RETURN(input->ReadAll(&proto));
return LoadFromSerializedProto(proto);
}
void SentencePieceProcessor::LoadOrDie(util::min_string_view filename) {
@ -57,10 +55,9 @@ void SentencePieceProcessor::LoadOrDie(util::min_string_view filename) {
}
util::Status SentencePieceProcessor::Load(std::istream *is) {
CHECK_OR_RETURN(is) << "input ifstream is null";
auto model_proto = port::MakeUnique<ModelProto>();
CHECK_OR_RETURN(model_proto->ParseFromIstream(is)) << "Model file is broken";
return Load(std::move(model_proto));
return util::StatusBuilder(util::error::UNIMPLEMENTED)
<< "std::stream API is deprecated. Use LoadFromSerializedProto() "
<< "to load model from any serialized blob object.";
}
util::Status SentencePieceProcessor::Load(const ModelProto &model_proto) {

1
src/util.h
View File

@ -19,7 +19,6 @@
#include <string.h>
#include <algorithm>
#include <fstream>
#include <memory>
#include <random>
#include <sstream>