lh3 / crlf Goto Github PK

CRLF, or Concise Run-Length Format, is a lightweight format to store run-length encoded strings, typically for small alphabets such as DNA. This repository contains the spec, a library and a tool to convert encodings. The library is implemented in two files crlf.h and crlf.c without any dependencies to other libraries.

The minimal code create a CRLF to stdout from a BWT string:

uint32_t dectab[256], l_BWT, i;
uint8_t *BWT; // $ACGTN encoded as 012345
crlf_t *crlf;
crlf_dectab_RL53(dectab); // generate the decoding table
crlf = crlf_create(0, 6, dectab, crlf_write_RL53, 0, 0);
for (i = 0; i < l_BWT; ++i)
    crlf_write(crlf, BWT[i], 1);
crlf_close(crlf);

The minimal code to read a CRLF from stdin:

int c;
uint64_t i, l;
crlf_t *crlf;
crlf = crlf_open(0);
while ((c = crlf_read(crlf, &l)) >= 0)
    for (i = 0; i < l; ++i)
        putchar("$ACGTN"[c]);
crlf_close(crlf);

Basic APIs (see crlf.h for details):

• crlf_create() creates a CRLF and writes the header, with a user-provided decoding table and a function pointer for encoding a run.

• crlf_open() opens an existing CRLF for reading.

• crlf_close() closes a CRLF.

• crlf_write() writes a run to CRLF. If this function is called consecutively on runs of the same symbol, these runs will be merged.

• crlf_read() reads a run from CRLF until it meets the next run of a different symbol from the current run.