Enable zstd compression format

.gitignore

@@ -39,4 +39,4 @@ fastp
 
 # Test Output
 *.json
-*.html
+*.html

Makefile

@@ -4,8 +4,8 @@ DIR_OBJ := ./obj
 
 PREFIX ?= /usr/local
 BINDIR ?= $(PREFIX)/bin
-INCLUDE_DIRS ?=
-LIBRARY_DIRS ?=
+INCLUDE_DIRS ?= 
+LIBRARY_DIRS ?= 
 
 SRC := $(wildcard ${DIR_SRC}/*.cpp)
 OBJ := $(patsubst %.cpp,${DIR_OBJ}/%.o,$(notdir ${SRC}))
@@ -16,7 +16,7 @@ BIN_TARGET := ${TARGET}
 
 CXX ?= g++
 CXXFLAGS := -std=c++11 -pthread -g -O3 -MD -MP -I. -I${DIR_INC} $(foreach includedir,$(INCLUDE_DIRS),-I$(includedir)) ${CXXFLAGS}
-LIBS := -lisal -ldeflate -lhwy -lpthread
+LIBS := -lisal -ldeflate -lzstd -lhwy -lpthread
 
 UNAME_S := $(shell uname -s)
 ifeq ($(UNAME_S),Linux)
@@ -27,10 +27,7 @@ else
   FIND_STATIC = $(firstword $(foreach d,$(LIBRARY_DIRS),$(wildcard $(d)/lib$(1).a)) $(wildcard /usr/local/lib/lib$(1).a /opt/homebrew/lib/lib$(1).a))
   STATIC_LIBS :=
   DYNAMIC_LIBS :=
-  $(foreach lib,isal deflate hwy,\
-    $(if $(call FIND_STATIC,$(lib)),\
-      $(eval STATIC_LIBS += $(call FIND_STATIC,$(lib))),\
-      $(eval DYNAMIC_LIBS += -l$(lib))))
+  $(foreach lib,isal deflate zstd hwy,    $(if $(call FIND_STATIC,$(lib)),      $(eval STATIC_LIBS += $(call FIND_STATIC,$(lib))),      $(eval DYNAMIC_LIBS += -l$(lib))))
   LD_FLAGS := $(foreach librarydir,$(LIBRARY_DIRS),-L$(librarydir)) $(STATIC_LIBS) $(DYNAMIC_LIBS) -lpthread $(LD_FLAGS)
 endif
 

README.md

@@ -76,6 +76,7 @@ If you use fastp in your work, you can cite fastp as:  *Shifu Chen. fastp 1.0: A
 11. support long reads (data from PacBio / Nanopore devices).
 12. support reading from STDIN and writing to STDOUT
 13. support interleaved input
+14. support reading Zstandard-compressed FASTQ/FASTA files (`.zst` / `.zstd`)
 14. support ultra-fast FASTQ-level deduplication
 15. ...
 
@@ -182,6 +183,7 @@ make -j INCLUDE_DIRS=/opt/homebrew/include LIBRARY_DIRS=/opt/homebrew/lib
 * for PE data, you should also specify read2 input by `-I` or `--in2`, and specify read2 output by `-O` or `--out2`.
 * if you don't specify the output file names, no output files will be written, but the QC will still be done for both data before and after filtering.
 * the output will be gzip-compressed if its file name ends with `.gz`
+* the input can be gzip-compressed (`.gz`) or Zstandard-compressed (`.zst`, `.zstd`); compression is auto-detected from the extension
 ## output to STDOUT
 `fastp` supports streaming the passing-filter reads to STDOUT, so that it can be passed to other compressors like `bzip2`, or be passed to aligners like `bwa` and `bowtie2`.
 * specify `--stdout` to enable this mode to stream output to STDOUT

src/fastqreader.cpp

@@ -50,13 +50,24 @@ FastqReader::FastqReader(string filename, bool hasQuality, bool phred64){
 	mHasNoLineBreakAtEnd = false;
 	mGzipInputUsedBytes = 0;
 	mReadPool = NULL;
+	mUseZstd = false;
+	mZstdFinished = false;
+	mZstdStream = NULL;
+	mZstdInput.src = NULL;
+	mZstdInput.size = 0;
+	mZstdInput.pos = 0;
+	mZstdInputUsedBytes = 0;
 	init();
 }
 
 FastqReader::~FastqReader(){
 	close();
 	delete[] mFastqBuf;
 	delete[] mGzipInputBuffer;
+	if(mZstdStream){
+		ZSTD_freeDStream(mZstdStream);
+		mZstdStream = NULL;
+	}
 }
 
 bool FastqReader::hasNoLineBreakAtEnd() {
@@ -69,11 +80,13 @@ void FastqReader::setReadPool(ReadPool* rp) {
 
 
 bool FastqReader::bufferFinished() {
-	if(mZipped) {
-		return eof() && mGzipState.avail_in == 0;
-	} else {
+	if(!mZipped)
 		return eof();
-	}
+
+	if(mUseZstd)
+		return mZstdFinished && mZstdInput.pos == mZstdInput.size;
+
+	return eof() && mGzipState.avail_in == 0;
 }
 
 void FastqReader::readToBufIgzip(){
@@ -139,10 +152,67 @@ void FastqReader::readToBufIgzip(){
 	}
 }
 
+void FastqReader::readToBufZstd(){
+	mBufDataLen = 0;
+	if(mZstdFinished)
+		return;
+
+	ZSTD_outBuffer outBuf;
+	outBuf.dst = mFastqBuf;
+	outBuf.size = mGzipOutputBufferSize;
+	outBuf.pos = 0;
+
+	while(outBuf.pos == 0){
+		if(mZstdInput.pos == mZstdInput.size){
+			size_t readBytes = fread(mGzipInputBuffer, 1, mGzipInputBufferSize, mFile);
+			if(readBytes == 0){
+				if(eof()){
+					mZstdFinished = true;
+					break;
+				} else {
+					error_exit("zstd: read error on file: " + mFilename);
+				}
+			}
+			mZstdInput.src = mGzipInputBuffer;
+			mZstdInput.size = readBytes;
+			mZstdInput.pos = 0;
+			mZstdInputUsedBytes += readBytes;
+		}
+
+		size_t ret = ZSTD_decompressStream(mZstdStream, &outBuf, &mZstdInput);
+		if(ZSTD_isError(ret)){
+			error_exit("zstd: decompression error for file: " + mFilename + ", error: " + string(ZSTD_getErrorName(ret)));
+		}
+
+		if(ret == 0){
+			if(mZstdInput.pos < mZstdInput.size || !eof()){
+				size_t resetRet = ZSTD_initDStream(mZstdStream);
+				if(ZSTD_isError(resetRet)){
+					error_exit("zstd: failed to reset stream for file: " + mFilename + ", error: " + string(ZSTD_getErrorName(resetRet)));
+				}
+			} else {
+				mZstdFinished = true;
+			}
+		}
+
+		if(eof() && mZstdInput.pos == mZstdInput.size && ret != 0){
+			error_exit("zstd: unexpected eof found in file: " + mFilename);
+		}
+
+		if(mZstdFinished || outBuf.pos > 0)
+			break;
+	}
+
+	mBufDataLen = outBuf.pos;
+}
+
 void FastqReader::readToBuf() {
 	mBufDataLen = 0;
 	if(mZipped) {
-		readToBufIgzip();
+		if(mUseZstd)
+			readToBufZstd();
+		else
+			readToBufIgzip();
 	} else {
 		if(!eof())
 			mBufDataLen = fread(mFastqBuf, 1, FQ_BUF_SIZE, mFile);
@@ -173,6 +243,26 @@ void FastqReader::init(){
 		}
 		mZipped = true;
 	}
+	else if (ends_with(mFilename, ".zst") || ends_with(mFilename, ".zstd")){
+		mFile = fopen(mFilename.c_str(), "rb");
+		if(mFile == NULL) {
+			error_exit("Failed to open file: " + mFilename);
+		}
+		mZstdStream = ZSTD_createDStream();
+		if(mZstdStream == NULL) {
+			error_exit("zstd: failed to allocate decompressor for file: " + mFilename);
+		}
+		size_t ret = ZSTD_initDStream(mZstdStream);
+		if(ZSTD_isError(ret)){
+			error_exit("zstd: failed to init decompressor for file: " + mFilename + ", error: " + string(ZSTD_getErrorName(ret)));
+		}
+		mZipped = true;
+		mUseZstd = true;
+		mZstdFinished = false;
+		mZstdInput.src = mGzipInputBuffer;
+		mZstdInput.size = 0;
+		mZstdInput.pos = 0;
+	}
 	else {
 		if(mFilename == "/dev/stdin") {
 			mFile = stdin;
@@ -189,7 +279,10 @@ void FastqReader::init(){
 
 void FastqReader::getBytes(size_t& bytesRead, size_t& bytesTotal) {
 	if(mZipped) {
-		bytesRead = mGzipInputUsedBytes - mGzipState.avail_in;
+		if(mUseZstd)
+			bytesRead = mZstdInputUsedBytes - (mZstdInput.size - mZstdInput.pos);
+		else
+			bytesRead = mGzipInputUsedBytes - mGzipState.avail_in;
 	} else {
 		bytesRead = ftell(mFile);//mFile.tellg();
 	}
@@ -362,6 +455,22 @@ bool FastqReader::isZipFastq(string filename) {
 		return true;
 	else if (ends_with(filename, ".fa.gz"))
 		return true;
+	else if (ends_with(filename, ".fastq.zst"))
+		return true;
+	else if (ends_with(filename, ".fq.zst"))
+		return true;
+	else if (ends_with(filename, ".fastq.zstd"))
+		return true;
+	else if (ends_with(filename, ".fq.zstd"))
+		return true;
+	else if (ends_with(filename, ".fasta.zst"))
+		return true;
+	else if (ends_with(filename, ".fa.zst"))
+		return true;
+	else if (ends_with(filename, ".fasta.zstd"))
+		return true;
+	else if (ends_with(filename, ".fa.zstd"))
+		return true;
 	else
 		return false;
 }

src/fastqreader.h

@@ -36,6 +36,7 @@ SOFTWARE.
 #else
 #include "igzip_lib.h"
 #endif
+#include <zstd.h>
 #include "readpool.h"
 
 class FastqReader{
@@ -65,6 +66,7 @@ class FastqReader{
 	void clearLineBreaks(char* line);
 	void readToBuf();
 	void readToBufIgzip();
+	void readToBufZstd();
 	bool bufferFinished();
 
 private:
@@ -87,6 +89,11 @@ class FastqReader{
 	bool mHasQuality;
 	bool mPhred64;
     ReadPool* mReadPool;
+	bool mUseZstd;
+	bool mZstdFinished;
+	ZSTD_DStream* mZstdStream;
+	ZSTD_inBuffer mZstdInput;
+	size_t mZstdInputUsedBytes;
 
 };