Mirrors
/
the-algorithm
mirror of https://github.com/twitter/the-algorithm.git
2
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

Merge 1020c610ef into 72eda9a24f

This commit is contained in:
Elijah 2023-07-17 21:41:37 -05:00 committed by GitHub
parent 72eda9a24f 1020c610ef
commit 0c1212dd5a
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
2 changed files with 46 additions and 46 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

12
twml/libtwml/include/twml/BlockFormatWriter.h
View File

@ -1,4 +1,5 @@
#pragma once
#include <fstream>
#include <twml/defines.h>
#include <cstdlib>
#include <cstdio>
@ -27,19 +28,18 @@ namespace twml {
class BlockFormatWriter {
private:
const char *file_name_;
FILE *outputfile_;
std::fstream outputfile_;
char temp_file_name_[PATH_MAX];
int record_index_;
int records_per_block_;
int pack_tag_and_wiretype(FILE *file, uint32_t tag, uint32_t wiretype);
int pack_varint_i32(FILE *file, int value);
int pack_string(FILE *file, const char *in, size_t in_len);
int write_int(FILE *file, int value);
int pack_tag_and_wiretype(std::fstream& fs, std::uint32_t tag, std::uint32_t wiretype);
int pack_varint_i32(std::fstream& fs, int value);
int pack_string(std::fstream& fs, const char *in, size_t in_len);
int write_int(std::fstream& fs, int value);
public:
BlockFormatWriter(const char *file_name, int record_per_block);
~BlockFormatWriter();
int write(const char *class_name, const char *record, int record_len) ;
int flush();
block_format_writer getHandle();

80
twml/libtwml/src/lib/BlockFormatWriter.cpp
View File

@ -11,7 +11,7 @@
#endif
#define MARKER_SIZE (16)
static uint8_t _marker[MARKER_SIZE] = {
static std::uint8_t _marker[MARKER_SIZE] = {
0x29, 0xd8, 0xd5, 0x06, 0x58, 0xcd, 0x4c, 0x29,
0xb2, 0xbc, 0x57, 0x99, 0x21, 0x71, 0xbd, 0xff
};
@ -20,29 +20,28 @@ namespace twml {
BlockFormatWriter::BlockFormatWriter(const char *file_name, int record_per_block) :
file_name_(file_name), record_index_(0), records_per_block_(record_per_block) {
snprintf(temp_file_name_, PATH_MAX, "%s.block", file_name);
outputfile_ = fopen(file_name_, "a");
std::fstream outputfile_(file_name_, std::ios::app);
if (!outputfile_.is_open()) {
throw std::runtime_error("Failed to open the output file");
}
}
BlockFormatWriter::~BlockFormatWriter() {
fclose(outputfile_);
}
// TODO: use fstream
int BlockFormatWriter::pack_tag_and_wiretype(FILE *buffer, uint32_t tag, uint32_t wiretype) {
uint8_t x = ((tag & 0x0f) << 3) | (wiretype & 0x7);
size_t n = fwrite(&x, 1, 1, buffer);
if (n != 1) {
int BlockFormatWriter::pack_tag_and_wiretype(std::fstream& fs, std::uint32_t tag, std::uint32_t wiretype) {
std::uint8_t x = ((tag & 0x0f) << 3) | (wiretype & 0x7);
fs.write(reinterpret_cast<char*>(&x), sizeof(x));
if (fs.fail()) {
return -1;
}
return 0;
}
int BlockFormatWriter::pack_varint_i32(FILE *buffer, int value) {
int BlockFormatWriter::pack_varint_i32(std::fstream& fs, int value) {
for (int i = 0; i < 10; i++) {
uint8_t x = value & 0x7F;
std::uint8_t x = value & 0x7F;
value = value >> 7;
if (value != 0) x |= 0x80;
size_t n = fwrite(&x, 1, 1, buffer);
if (n != 1) {
fs.write(reinterpret_cast<const char*>(&x), sizeof(x));
if (fs.fail()) {
return -1;
}
if (value == 0) break;
@ -50,21 +49,23 @@ namespace twml {
return 0;
}
int BlockFormatWriter::pack_string(FILE *buffer, const char *in, size_t in_len) {
if (pack_varint_i32(buffer, in_len)) return -1;
size_t n = fwrite(in, 1, in_len, buffer);
if (n != in_len) return -1;
int BlockFormatWriter::pack_string(std::fstream& fs, const char *in, size_t in_len) {
if (pack_varint_i32(fs, in_len)) return -1;
fs.write(in, in_len);
if (fs.fail()) {
return -1;
}
return 0;
}
int BlockFormatWriter::write_int(FILE *buffer, int value) {
uint8_t buff[4];
buff[0] = value & 0xff;
buff[1] = (value >> 8) & 0xff;
buff[2] = (value >> 16) & 0xff;
buff[3] = (value >> 24) & 0xff;
size_t n = fwrite(buff, 1, 4, buffer);
if (n != 4) {
int BlockFormatWriter::write_int(std::fstream& fs, int value) {
std::uint8_t buff[4];
buff[0] = static_cast<std::uint8_t>(value & 0xff);
buff[1] = static_cast<std::uint8_t>((value >> 8) & 0xff);
buff[2] = static_cast<std::uint8_t>((value >> 16) & 0xff);
buff[3] = static_cast<std::uint8_t>((value >> 24) & 0xff);
fs.write(reinterpret_cast<char*>(buff), sizeof(buff));
if (fs.fail()) {
return -1;
}
return 0;
@ -74,9 +75,9 @@ namespace twml {
if (record) {
record_index_++;
// The buffer holds max records_per_block_ of records (block).
FILE *buffer = fopen(temp_file_name_, "a");
if (!buffer) return -1;
if (ftell(buffer) == 0) {
std::fstream buffer(temp_file_name_, std::ios::app);
if (!buffer.is_open()) return -1;
if (buffer.tellg() == 0) {
if (pack_tag_and_wiretype(buffer, 1, WIRE_TYPE_VARINT))
throw std::invalid_argument("Error writting tag and wiretype");
if (pack_varint_i32(buffer, 1))
@ -90,7 +91,6 @@ namespace twml {
throw std::invalid_argument("Error writtig tag and wiretype");
if (pack_string(buffer, record, record_len))
throw std::invalid_argument("Error writting record");
fclose(buffer);
}
if ((record_index_ % records_per_block_) == 0) {
@ -101,24 +101,24 @@ namespace twml {
int BlockFormatWriter::flush() {
// Flush the records in the buffer to outputfile
FILE *buffer = fopen(temp_file_name_, "r");
std::fstream buffer(temp_file_name_, std::ios::in);
if (buffer) {
fseek(buffer, 0, SEEK_END);
int64_t block_size = ftell(buffer);
fseek(buffer, 0, SEEK_SET);
buffer.seekg(0, buffer.end);
int64_t block_size = buffer.tellg();
buffer.seekg(0, buffer.beg);
if (fwrite(_marker, sizeof(_marker), 1, outputfile_) != 1) return 1;
if (outputfile_.write(reinterpret_cast<const char*>(_marker), sizeof(_marker)).fail() != 1) return 1;
if (write_int(outputfile_, block_size)) return 1;
uint8_t buff[4096];
std::uint8_t buff[4096];
while (1) {
size_t n = fread(buff, 1, sizeof(buff), buffer);
buffer.read(reinterpret_cast<char*>(buff), sizeof(buff));
std::streamsize n = buffer.gcount();
if (n) {
size_t x = fwrite(buff, 1, n, outputfile_);
if (x != n) return 1;
outputfile_.write(reinterpret_cast<const char*>(buff), n);
if (outputfile_.fail()) return 1;
}
if (n != sizeof(buff)) break;
}
fclose(buffer);
// Remove the buffer
if (remove(temp_file_name_)) return 1;
}