static int
bindkey(KEYMAP **mapp, const char *fname, KCHAR *keys, int kcount)
{
KEYMAP *curmap = *mapp;
KEYMAP *pref_map = NULL;
PF funct;
int c;
if (fname == NULL)
funct = rescan;
else if (((funct = name_function(fname)) == NULL) ?
(pref_map = name_map(fname)) == NULL : funct == NULL) {
ewprintf("[No match: %s]", fname);
return (FALSE);
}
while (--kcount) {
if (doscan(curmap, c = *keys++, &curmap) != NULL) {
if (remap(curmap, c, NULL, NULL) != TRUE)
return (FALSE);
/*
* XXX - Bizzarreness. remap creates an empty KEYMAP
* that the last key is supposed to point to.
*/
curmap = ele->k_prefmap;
}
}
(void)doscan(curmap, c = *keys, NULL);
return (remap(curmap, c, funct, pref_map));
}
int methyltrain_main(int argc, char** argv)
{
parse_methyltrain_options(argc, argv);
omp_set_num_threads(opt::num_threads);
Fast5Map name_map(opt::reads_file);
// copy the input model into a new type that will hold the trained models
const PoreModelMap& input_models = PoreModelSet::get_models(opt::initial_model_type);
assert(!input_models.empty());
for(auto model_iter : input_models) {
PoreModel model_copy = model_iter.second;
model_copy.type = opt::trained_model_type;
PoreModelSet::insert_model(model_copy.type, model_copy);
}
// Set the alphabet for this run to be the alphabet for the first model
assert(!PoreModelSet::get_models(opt::initial_model_type).empty());
mtrain_alphabet = PoreModelSet::get_models(opt::initial_model_type).begin()->second.pmalphabet;
for(size_t round = 0; round < opt::num_training_rounds; round++) {
fprintf(stderr, "Starting round %zu\n", round);
train_one_round(name_map, round);
if(opt::write_models) {
write_models(PoreModelSet::get_models(opt::trained_model_type), round);
}
}
return EXIT_SUCCESS;
}
void set_critical_spline_constants(Fluid *pFluid)
{
// This is a map of fluid name to CriticalSplineStruct_T instance
// Admittedly this is a bit hacky, but it is easier to update the
// file separately and at compile times the constants will be pulled into
// this block.
// The list of spline parameters is generated by the function rebuild_CriticalSplineConstants_T in FluidClass.cpp
std::pair<std::string, CriticalSplineStruct_T> name_map_datad[] = {
#include "CriticalSplineConstants_T.h"
};
//Now actually construct the map
std::map<std::string, CriticalSplineStruct_T> name_map(name_map_datad, name_map_datad + sizeof name_map_datad / sizeof name_map_datad[0]);
std::map<std::string,CriticalSplineStruct_T>::iterator it;
// Try to find using the map using the official name of the fluid
it = name_map.find(pFluid->get_name());
// If it is found the iterator will not be equal to end
if (it != name_map.end() )
{
pFluid->CriticalSpline_T = (*it).second;
}
else
{
// Didn't work, just filling with infinity
pFluid->CriticalSpline_T = CriticalSplineStruct_T(_HUGE,_HUGE,_HUGE,_HUGE,_HUGE);
}
}
/*
* Do the input for local-set-key, global-set-key and define-key
* then call remap to do the work.
*/
static int
dobind(KEYMAP *curmap, const char *p, int unbind)
{
KEYMAP *pref_map = NULL;
PF funct;
char bprompt[80], *bufp, *pep;
int c, s, n;
if (macrodef) {
/*
* Keystrokes aren't collected. Not hard, but pretty useless.
* Would not work for function keys in any case.
*/
ewprintf("Can't rebind key in macro");
return (FALSE);
}
if (inmacro) {
for (s = 0; s < maclcur->l_used - 1; s++) {
if (doscan(curmap, c = CHARMASK(maclcur->l_text[s]), &curmap)
!= NULL) {
if (remap(curmap, c, NULL, NULL)
!= TRUE)
return (FALSE);
}
}
(void)doscan(curmap, c = maclcur->l_text[s], NULL);
maclcur = maclcur->l_fp;
} else {
n = strlcpy(bprompt, p, sizeof(bprompt));
if (n >= sizeof(bprompt))
n = sizeof(bprompt) - 1;
pep = bprompt + n;
for (;;) {
ewprintf("%s", bprompt);
pep[-1] = ' ';
pep = getkeyname(pep, sizeof(bprompt) -
(pep - bprompt), c = getkey(FALSE));
if (doscan(curmap, c, &curmap) != NULL)
break;
*pep++ = '-';
*pep = '\0';
}
}
if (unbind)
funct = rescan;
else {
if ((bufp = eread("%s to command: ", bprompt, sizeof(bprompt),
EFFUNC | EFNEW, bprompt)) == NULL)
return (ABORT);
else if (bufp[0] == '\0')
return (FALSE);
if (((funct = name_function(bprompt)) == NULL) ?
(pref_map = name_map(bprompt)) == NULL : funct == NULL) {
ewprintf("[No match]");
return (FALSE);
}
}
return (remap(curmap, c, funct, pref_map));
}
int
help_help(int f, int n)
{
KEYMAP *kp;
PF funct;
if ((kp = name_map("help")) == NULL)
return (FALSE);
ewprintf("a b c: ");
do {
funct = doscan(kp, getkey(FALSE), NULL);
} while (funct == NULL || funct == help_help);
if (macrodef && macrocount < MAXMACRO)
macro[macrocount - 1].m_funct = funct;
return ((*funct)(f, n));
}
int consensus_main(int argc, char** argv)
{
parse_consensus_options(argc, argv);
omp_set_num_threads(opt::num_threads);
Fast5Map name_map(opt::reads_file);
// Parse the window string
// Replace ":" and "-" with spaces to make it parseable with stringstream
std::replace(opt::window.begin(), opt::window.end(), ':', ' ');
std::replace(opt::window.begin(), opt::window.end(), '-', ' ');
const int WINDOW_LENGTH = 10000;
const int WINDOW_OVERLAP = 200;
std::stringstream parser(opt::window);
std::string contig;
int start_window_id;
int end_window_id;
parser >> contig >> start_window_id >> end_window_id;
FILE* out_fp = NULL;
if(!opt::output_file.empty()) {
out_fp = fopen(opt::output_file.c_str(), "w");
} else {
out_fp = stdout;
}
for(int window_id = start_window_id; window_id < end_window_id; ++window_id) {
int start_base = window_id * WINDOW_LENGTH;
int end_base = start_base + WINDOW_LENGTH + WINDOW_OVERLAP;
std::string window_consensus = call_consensus_for_window(name_map, contig, start_base, end_base);
fprintf(out_fp, ">%s:%d\n%s\n", contig.c_str(), window_id, window_consensus.c_str());
}
if(out_fp != stdout) {
fclose(out_fp);
}
return 0;
}
/* ARGSUSED */
int
redefine_key(int f, int n)
{
static char buf[48];
char tmp[32], *bufp;
KEYMAP *mp;
(void)strlcpy(buf, "Define key map: ", sizeof(buf));
if ((bufp = eread(buf, tmp, sizeof(tmp), EFNEW)) == NULL)
return (ABORT);
else if (bufp[0] == '\0')
return (FALSE);
(void)strlcat(buf, tmp, sizeof(buf));
if ((mp = name_map(tmp)) == NULL) {
ewprintf("Unknown map %s", tmp);
return (FALSE);
}
if (strlcat(buf, "key: ", sizeof(buf)) >= sizeof(buf))
return (FALSE);
return (dobind(mp, buf, FALSE));
}
void dot_graph(const binspector_analyzer_t::structure_map_t& struct_map,
const std::string& starting_struct,
const boost::filesystem::path& output_root)
{
boost::filesystem::path output_filename(output_root / "graph.gv");
boost::filesystem::ofstream output(output_filename);
dot_graph_t graph;
assert (output);
graph.name_m = "G";
graph.type_m = "digraph";
//graph.setting_vector_m.push_back("node [shape=box]");
//graph.setting_vector_m.push_back("rankdir=TB");
//graph.setting_vector_m.push_back("packMode=clust");
graph.setting_vector_m.push_back("outputMode=nodesfirst");
adobe::name_t starting_name(starting_struct.c_str());
std::string starting_id(name_map(starting_name));
dot_node_t node;
node.attribute_map_m["label"] = starting_name.c_str();
node.id_m = starting_id;
graph.node_vector_m.push_back(node);
graph_struct(struct_map,
starting_name,
starting_id,
binspector_analyzer_t::typedef_map_t(),
0,
graph);
format_graph(graph, output);
std::cerr << "File written to " << output_filename.string() << '\n';
}
/*
* excline - run a line from a load file or eval-expression. If FKEYS is
* defined, duplicate functionality of dobind so function key values don't
* have to fit in type char.
*/
int
excline(char *line)
{
PF fp;
struct line *lp, *np;
int status, c, f, n;
char *funcp, *tmp;
char *argp = NULL;
long nl;
#ifdef FKEYS
int bind;
KEYMAP *curmap;
#define BINDARG 0 /* this arg is key to bind (local/global set key) */
#define BINDNO 1 /* not binding or non-quoted BINDARG */
#define BINDNEXT 2 /* next arg " (define-key) */
#define BINDDO 3 /* already found key to bind */
#define BINDEXT 1 /* space for trailing \0 */
#else /* FKEYS */
#define BINDEXT 0
#endif /* FKEYS */
lp = NULL;
if (macrodef || inmacro) {
ewprintf("Not now!");
return (FALSE);
}
f = 0;
n = 1;
funcp = skipwhite(line);
if (*funcp == '\0')
return (TRUE); /* No error on blank lines */
line = parsetoken(funcp);
if (*line != '\0') {
*line++ = '\0';
line = skipwhite(line);
if (ISDIGIT(*line) || *line == '-') {
argp = line;
line = parsetoken(line);
}
}
if (argp != NULL) {
f = FFARG;
nl = strtol(argp, &tmp, 10);
if (*tmp != '\0')
return (FALSE);
if (nl >= INT_MAX || nl <= INT_MIN)
return (FALSE);
n = (int)nl;
}
if ((fp = name_function(funcp)) == NULL) {
ewprintf("Unknown function: %s", funcp);
return (FALSE);
}
#ifdef FKEYS
if (fp == bindtokey || fp == unbindtokey) {
bind = BINDARG;
curmap = fundamental_map;
} else if (fp == localbind || fp == localunbind) {
bind = BINDARG;
curmap = curbp->b_modes[curbp->b_nmodes]->p_map;
} else if (fp == redefine_key)
bind = BINDNEXT;
else
bind = BINDNO;
#endif /* FKEYS */
/* Pack away all the args now... */
if ((np = lalloc(0)) == FALSE)
return (FALSE);
np->l_fp = np->l_bp = maclcur = np;
while (*line != '\0') {
argp = skipwhite(line);
if (*argp == '\0')
break;
line = parsetoken(argp);
if (*argp != '"') {
if (*argp == '\'')
++argp;
if ((lp = lalloc((int) (line - argp) + BINDEXT)) ==
NULL) {
status = FALSE;
goto cleanup;
}
bcopy(argp, ltext(lp), (int)(line - argp));
#ifdef FKEYS
/* don't count BINDEXT */
lp->l_used--;
if (bind == BINDARG)
bind = BINDNO;
#endif /* FKEYS */
} else {
/* quoted strings are special */
++argp;
#ifdef FKEYS
if (bind != BINDARG) {
#endif /* FKEYS */
lp = lalloc((int)(line - argp) + BINDEXT);
if (lp == NULL) {
status = FALSE;
goto cleanup;
}
lp->l_used = 0;
#ifdef FKEYS
} else
key.k_count = 0;
#endif /* FKEYS */
while (*argp != '"' && *argp != '\0') {
if (*argp != '\\')
c = *argp++;
else {
switch (*++argp) {
case 't':
case 'T':
c = CCHR('I');
break;
case 'n':
case 'N':
c = CCHR('J');
break;
case 'r':
case 'R':
c = CCHR('M');
break;
case 'e':
case 'E':
c = CCHR('[');
break;
case '^':
/*
* split into two statements
* due to bug in OSK cpp
*/
c = CHARMASK(*++argp);
c = ISLOWER(c) ?
CCHR(TOUPPER(c)) : CCHR(c);
break;
case '0':
case '1':
case '2':
case '3':
case '4':
case '5':
case '6':
case '7':
c = *argp - '0';
if (argp[1] <= '7' &&
argp[1] >= '0') {
c <<= 3;
c += *++argp - '0';
if (argp[1] <= '7' &&
argp[1] >= '0') {
c <<= 3;
c += *++argp
- '0';
}
}
break;
#ifdef FKEYS
case 'f':
case 'F':
c = *++argp - '0';
if (ISDIGIT(argp[1])) {
c *= 10;
c += *++argp - '0';
}
c += KFIRST;
break;
#endif /* FKEYS */
default:
c = CHARMASK(*argp);
break;
}
argp++;
}
#ifdef FKEYS
if (bind == BINDARG)
key.k_chars[key.k_count++] = c;
else
#endif /* FKEYS */
lp->l_text[lp->l_used++] = c;
}
if (*line)
line++;
}
#ifdef FKEYS
switch (bind) {
case BINDARG:
bind = BINDDO;
break;
case BINDNEXT:
lp->l_text[lp->l_used] = '\0';
if ((curmap = name_map(lp->l_text)) == NULL) {
ewprintf("No such mode: %s", lp->l_text);
status = FALSE;
free(lp);
goto cleanup;
}
free(lp);
bind = BINDARG;
break;
default:
#endif /* FKEYS */
lp->l_fp = np->l_fp;
lp->l_bp = np;
np->l_fp = lp;
np = lp;
#ifdef FKEYS
}
#endif /* FKEYS */
}
#ifdef FKEYS
switch (bind) {
default:
ewprintf("Bad args to set key");
status = FALSE;
break;
case BINDDO:
if (fp != unbindtokey && fp != localunbind) {
lp->l_text[lp->l_used] = '\0';
status = bindkey(&curmap, lp->l_text, key.k_chars,
key.k_count);
} else
status = bindkey(&curmap, NULL, key.k_chars,
key.k_count);
break;
case BINDNO:
#endif /* FKEYS */
inmacro = TRUE;
maclcur = maclcur->l_fp;
status = (*fp)(f, n);
inmacro = FALSE;
#ifdef FKEYS
}
#endif /* FKEYS */
cleanup:
lp = maclcur->l_fp;
while (lp != maclcur) {
np = lp->l_fp;
free(lp);
lp = np;
}
free(lp);
return (status);
}
/*
return the short name for a component of a full name
*/
char *pvfs_short_name_component(struct pvfs_state *pvfs, const char *name)
{
return name_map(pvfs->mangle_ctx, name, true, true);
}
int methyltest_main(int argc, char** argv)
{
parse_methyltest_options(argc, argv);
omp_set_num_threads(opt::num_threads);
Fast5Map name_map(opt::reads_file);
ModelMap models = read_models_fofn(opt::models_fofn, mtest_alphabet);
// Open the BAM and iterate over reads
// load bam file
htsFile* bam_fh = sam_open(opt::bam_file.c_str(), "r");
assert(bam_fh != NULL);
// load bam index file
std::string index_filename = opt::bam_file + ".bai";
hts_idx_t* bam_idx = bam_index_load(index_filename.c_str());
assert(bam_idx != NULL);
// read the bam header
bam_hdr_t* hdr = sam_hdr_read(bam_fh);
// load reference fai file
faidx_t *fai = fai_load(opt::genome_file.c_str());
hts_itr_t* itr;
// If processing a region of the genome, only emit events aligned to this window
int clip_start = -1;
int clip_end = -1;
if(opt::region.empty()) {
// TODO: is this valid?
itr = sam_itr_queryi(bam_idx, HTS_IDX_START, 0, 0);
} else {
fprintf(stderr, "Region: %s\n", opt::region.c_str());
itr = sam_itr_querys(bam_idx, hdr, opt::region.c_str());
hts_parse_reg(opt::region.c_str(), &clip_start, &clip_end);
}
#ifndef H5_HAVE_THREADSAFE
if(opt::num_threads > 1) {
fprintf(stderr, "You enabled multi-threading but you do not have a threadsafe HDF5\n");
fprintf(stderr, "Please recompile nanopolish's built-in libhdf5 or run with -t 1\n");
exit(1);
}
#endif
// Initialize writers
OutputHandles handles;
handles.site_writer = fopen(std::string(opt::bam_file + ".methyltest.sites.bed").c_str(), "w");
handles.read_writer = fopen(std::string(opt::bam_file + ".methyltest.reads.tsv").c_str(), "w");
handles.strand_writer = fopen(std::string(opt::bam_file + ".methyltest.strand.tsv").c_str(), "w");
// Write a header to the reads.tsv file
fprintf(handles.read_writer, "name\tsum_ll_ratio\tn_cpg\tcomplement_model\ttags\n");
// strand header
fprintf(handles.strand_writer, "name\tsum_ll_ratio\tn_cpg\tmodel\n");
// Initialize iteration
std::vector<bam1_t*> records(opt::batch_size, NULL);
for(size_t i = 0; i < records.size(); ++i) {
records[i] = bam_init1();
}
int result;
size_t num_reads_processed = 0;
size_t num_records_buffered = 0;
Progress progress("[methyltest]");
do {
assert(num_records_buffered < records.size());
// read a record into the next slot in the buffer
result = sam_itr_next(bam_fh, itr, records[num_records_buffered]);
num_records_buffered += result >= 0;
// realign if we've hit the max buffer size or reached the end of file
if(num_records_buffered == records.size() || result < 0) {
#pragma omp parallel for
for(size_t i = 0; i < num_records_buffered; ++i) {
bam1_t* record = records[i];
size_t read_idx = num_reads_processed + i;
if( (record->core.flag & BAM_FUNMAP) == 0) {
calculate_methylation_for_read(models, name_map, fai, hdr, record, read_idx, handles);
}
}
num_reads_processed += num_records_buffered;
num_records_buffered = 0;
}
} while(result >= 0);
assert(num_records_buffered == 0);
progress.end();
// cleanup records
for(size_t i = 0; i < records.size(); ++i) {
bam_destroy1(records[i]);
}
// cleanup
fclose(handles.site_writer);
fclose(handles.read_writer);
fclose(handles.strand_writer);
sam_itr_destroy(itr);
bam_hdr_destroy(hdr);
fai_destroy(fai);
sam_close(bam_fh);
hts_idx_destroy(bam_idx);
return EXIT_SUCCESS;
}
int scorereads_main(int argc, char** argv)
{
parse_scorereads_options(argc, argv);
omp_set_num_threads(opt::num_threads);
Fast5Map name_map(opt::reads_file);
ModelMap models;
if (!opt::models_fofn.empty())
models = read_models_fofn(opt::models_fofn);
// Open the BAM and iterate over reads
// load bam file
htsFile* bam_fh = sam_open(opt::bam_file.c_str(), "r");
assert(bam_fh != NULL);
// load bam index file
std::string index_filename = opt::bam_file + ".bai";
hts_idx_t* bam_idx = bam_index_load(index_filename.c_str());
assert(bam_idx != NULL);
// read the bam header
bam_hdr_t* hdr = sam_hdr_read(bam_fh);
// load reference fai file
faidx_t *fai = fai_load(opt::genome_file.c_str());
hts_itr_t* itr;
// If processing a region of the genome, only emit events aligned to this window
int clip_start = -1;
int clip_end = -1;
if(opt::region.empty()) {
// TODO: is this valid?
itr = sam_itr_queryi(bam_idx, HTS_IDX_START, 0, 0);
} else {
fprintf(stderr, "Region: %s\n", opt::region.c_str());
itr = sam_itr_querys(bam_idx, hdr, opt::region.c_str());
hts_parse_reg(opt::region.c_str(), &clip_start, &clip_end);
}
#ifndef H5_HAVE_THREADSAFE
if(opt::num_threads > 1) {
fprintf(stderr, "You enabled multi-threading but you do not have a threadsafe HDF5\n");
fprintf(stderr, "Please recompile nanopolish's built-in libhdf5 or run with -t 1\n");
exit(1);
}
#endif
// Initialize iteration
std::vector<bam1_t*> records(opt::batch_size, NULL);
for(size_t i = 0; i < records.size(); ++i) {
records[i] = bam_init1();
}
int result;
size_t num_reads_realigned = 0;
size_t num_records_buffered = 0;
do {
assert(num_records_buffered < records.size());
// read a record into the next slot in the buffer
result = sam_itr_next(bam_fh, itr, records[num_records_buffered]);
num_records_buffered += result >= 0;
// realign if we've hit the max buffer size or reached the end of file
if(num_records_buffered == records.size() || result < 0) {
#pragma omp parallel for schedule(dynamic)
for(size_t i = 0; i < num_records_buffered; ++i) {
bam1_t* record = records[i];
size_t read_idx = num_reads_realigned + i;
if( (record->core.flag & BAM_FUNMAP) == 0) {
//load read
std::string read_name = bam_get_qname(record);
std::string fast5_path = name_map.get_path(read_name);
SquiggleRead sr(read_name, fast5_path);
// TODO: early exit when have processed all of the reads in readnames
if (!opt::readnames.empty() &&
std::find(opt::readnames.begin(), opt::readnames.end(), read_name) == opt::readnames.end() )
continue;
for(size_t strand_idx = 0; strand_idx < NUM_STRANDS; ++strand_idx) {
std::vector<EventAlignment> ao = alignment_from_read(sr, strand_idx, read_idx,
models, fai, hdr,
record, clip_start, clip_end);
if (ao.size() == 0)
continue;
// Update pore model based on alignment
if ( opt::calibrate )
recalibrate_model(sr, strand_idx, ao, false);
double score = model_score(sr, strand_idx, fai, ao, 500);
if (score > 0)
continue;
#pragma omp critical(print)
std::cout << read_name << " " << ( strand_idx ? "complement" : "template" )
<< " " << sr.pore_model[strand_idx].name << " " << score << std::endl;
}
}
}
num_reads_realigned += num_records_buffered;
num_records_buffered = 0;
}
} while(result >= 0);
// cleanup records
for(size_t i = 0; i < records.size(); ++i) {
bam_destroy1(records[i]);
}
// cleanup
sam_itr_destroy(itr);
bam_hdr_destroy(hdr);
fai_destroy(fai);
sam_close(bam_fh);
hts_idx_destroy(bam_idx);
return 0;
}
