/* read_mask_file
*
* Given an open file pointer, read the first token of the
* file and return it as *ret_mask. It must contain only
* '0' or '1' characters.
*
* Returns: eslOK on success.
*/
int
read_mask_file(char *filename, char *errbuf, char **ret_mask, int *ret_mask_len)
{
ESL_FILEPARSER *efp = NULL;
char *mask = NULL;
char *tok;
int toklen;
int n;
int status;
if (esl_fileparser_Open(filename, NULL, &efp) != eslOK) ESL_XFAIL(eslFAIL, errbuf, "failed to open %s in read_mask_file\n", filename);
esl_fileparser_SetCommentChar(efp, '#');
if((status = esl_fileparser_GetToken(efp, &tok, &toklen)) != eslOK) ESL_XFAIL(eslFAIL, errbuf, "failed to read a single token from %s\n", filename);
ESL_ALLOC(mask, sizeof(char) * (toklen+1));
for(n = 0; n < toklen; n++) {
if((tok[n] == '0') || (tok[n] == '1')) {
mask[n] = tok[n];
}
else { ESL_XFAIL(eslFAIL, errbuf, "read a non-0 and non-1 character (%c) in the mask file %s\n", tok[n], filename); }
}
mask[n] = '\0';
*ret_mask = mask;
*ret_mask_len = n;
esl_fileparser_Close(efp);
return eslOK;
ERROR:
if (efp) esl_fileparser_Close(efp);
if (mask) free(mask);
return status;
}
/* Function: esl_hyperexp_Read()
*
* Purpose: Reads hyperexponential parameters from an open <e>.
* which is an <ESL_FILEPARSER> tokenizer for an open stream.
*
* The first token is <K>, the number of mixture components.
* The second token is <mu>, the x offset shared by all components.
* Then for each mixture component <k=1..K>, it reads
* a mixture coefficient <q[k]> and a decay parameter
* <lambda[k]>.
*
* The <2K+2> data tokens must occur in this order, but
* they can be grouped into any number of lines, because the
* parser ignores line breaks.
*
* Anything after a <\#> character on a line is a comment, and
* is ignored.
*
* Returns: <eslOK> on success, and <ret_hxp> points to a new <ESL_HYPEREXP>
* object.
* <eslEFORMAT> on "normal" parse failure caused by a bad file
* format that's likely the user's fault.
*
* Throws: <eslEMEM> if allocation of the new <ESL_HYPEREXP> fails.
*
*
* FIXME: All our mixture models (esl_dirichlet, for example) should be
* reconciled w/ identical interfaces & behaviour.
*/
int
esl_hyperexp_Read(ESL_FILEPARSER *e, ESL_HYPEREXP **ret_hxp)
{
ESL_HYPEREXP *hxp = NULL;
char *tok;
int status = eslOK;
int nc;
int k;
double sum;
esl_fileparser_SetCommentChar(e, '#');
if ((status = esl_fileparser_GetToken(e, &tok, NULL)) != eslOK) goto ERROR;
nc = atoi(tok);
if (nc < 1) {
sprintf(e->errbuf, "Expected # of components K >= 1 as first token");
goto ERROR;
}
if ((hxp = esl_hyperexp_Create(nc)) == NULL) return eslEMEM; /* percolation */
if ((status = esl_fileparser_GetToken(e, &tok, NULL)) != eslOK) goto ERROR;
hxp->mu = atof(tok);
for (k = 0; k < hxp->K; k++)
{
if ((status = esl_fileparser_GetToken(e, &tok, NULL)) != eslOK) goto ERROR;
hxp->q[k] = atof(tok);
if ((status = esl_fileparser_GetToken(e, &tok, NULL)) != eslOK) goto ERROR;
hxp->lambda[k] = atof(tok);
if (hxp->q[k] < 0. || hxp->q[k] > 1.) {
sprintf(e->errbuf, "Expected a mixture coefficient q[k], 0<=q[k]<=1");
goto ERROR;
}
if (hxp->lambda[k] <= 0.) {
sprintf(e->errbuf, "Expected a lambda parameter, lambda>0");
goto ERROR;
}
}
sum = esl_vec_DSum(hxp->q, hxp->K);
if (fabs(sum-1.0) > 0.05) {
sprintf(e->errbuf, "Expected mixture coefficients to sum to 1");
goto ERROR;
}
esl_vec_DNorm(hxp->q, hxp->K);
*ret_hxp = hxp;
return eslOK;
ERROR:
esl_hyperexp_Destroy(hxp);
return eslEFORMAT;
}
/* multifetch:
* given a file containing lines with one name or key per line;
* parse the file line-by-line;
* if we have an SSI index available, retrieve the HMMs by key
* as we see each line;
* else, without an SSI index, store the keys in a hash, then
* read the entire HMM file in a single pass, outputting HMMs
* that are in our keylist.
*
* Note that with an SSI index, you get the HMMs in the order they
* appear in the <keyfile>, but without an SSI index, you get HMMs in
* the order they occur in the HMM file.
*/
static void
multifetch(ESL_GETOPTS *go, FILE *ofp, char *keyfile, P7_HMMFILE *hfp)
{
ESL_KEYHASH *keys = esl_keyhash_Create();
ESL_FILEPARSER *efp = NULL;
ESL_ALPHABET *abc = NULL;
P7_HMM *hmm = NULL;
int nhmm = 0;
char *key;
int keylen;
int keyidx;
int status;
if (esl_fileparser_Open(keyfile, NULL, &efp) != eslOK) p7_Fail("Failed to open key file %s\n", keyfile);
esl_fileparser_SetCommentChar(efp, '#');
while (esl_fileparser_NextLine(efp) == eslOK)
{
if (esl_fileparser_GetTokenOnLine(efp, &key, &keylen) != eslOK)
p7_Fail("Failed to read HMM name on line %d of file %s\n", efp->linenumber, keyfile);
status = esl_key_Store(keys, key, &keyidx);
if (status == eslEDUP) p7_Fail("HMM key %s occurs more than once in file %s\n", key, keyfile);
if (hfp->ssi != NULL) { onefetch(go, ofp, key, hfp); nhmm++; }
}
if (hfp->ssi == NULL)
{
while ((status = p7_hmmfile_Read(hfp, &abc, &hmm)) != eslEOF)
{
if (status == eslEOD) p7_Fail("read failed, HMM file %s may be truncated?", hfp->fname);
else if (status == eslEFORMAT) p7_Fail("bad file format in HMM file %s", hfp->fname);
else if (status == eslEINCOMPAT) p7_Fail("HMM file %s contains different alphabets", hfp->fname);
else if (status != eslOK) p7_Fail("Unexpected error in reading HMMs from %s", hfp->fname);
if (esl_key_Lookup(keys, hmm->name, &keyidx) == eslOK ||
((hmm->acc) && esl_key_Lookup(keys, hmm->acc, &keyidx) == eslOK))
{
p7_hmmfile_WriteASCII(ofp, -1, hmm);
nhmm++;
}
p7_hmm_Destroy(hmm);
}
}
if (ofp != stdout) printf("\nRetrieved %d HMMs.\n", nhmm);
if (abc != NULL) esl_alphabet_Destroy(abc);
esl_keyhash_Destroy(keys);
esl_fileparser_Close(efp);
return;
}
/* multifetch:
* given a file containing lines with one name or key per line;
* parse the file line-by-line;
* if we have an SSI index available, retrieve the MSAs by key
* as we see each line;
* else, without an SSI index, store the keys in a hash, then
* read the entire MSA file in a single pass, outputting MSAs
* that are in our keylist.
*
* Note that with an SSI index, you get the MSAs in the order they
* appear in the <keyfile>, but without an SSI index, you get MSAs in
* the order they occur in the MSA file.
*/
static void
multifetch(ESL_GETOPTS *go, FILE *ofp, int outfmt, char *keyfile, ESLX_MSAFILE *afp)
{
ESL_KEYHASH *keys = esl_keyhash_Create();
ESL_FILEPARSER *efp = NULL;
ESL_MSA *msa = NULL;
int nali = 0;
char *key;
int keylen;
int keyidx;
int status;
if (esl_fileparser_Open(keyfile, NULL, &efp) != eslOK)
esl_fatal("Failed to open key file %s\n", keyfile);
esl_fileparser_SetCommentChar(efp, '#');
while (esl_fileparser_NextLine(efp) == eslOK)
{
if (esl_fileparser_GetTokenOnLine(efp, &key, &keylen) != eslOK)
esl_fatal("Failed to read MSA name on line %d of file %s\n", efp->linenumber, keyfile);
status = esl_keyhash_Store(keys, key, keylen, &keyidx);
if (status == eslEDUP) esl_fatal("MSA key %s occurs more than once in file %s\n", key, keyfile);
if (afp->ssi) { onefetch(go, ofp, outfmt, key, afp); nali++; }
}
if (! afp->ssi)
{
while ((status = eslx_msafile_Read(afp, &msa)) != eslEOF)
{
if (status != eslOK) eslx_msafile_ReadFailure(afp, status);
nali++;
if (msa->name == NULL)
esl_fatal("Every alignment in file must have a name to be retrievable. Failed to find name of alignment #%d\n", nali);
if ( (esl_keyhash_Lookup(keys, msa->name, -1, NULL) == eslOK) ||
(msa->acc != NULL && esl_keyhash_Lookup(keys, msa->acc, -1, NULL) == eslOK))
eslx_msafile_Write(ofp, msa, outfmt);
esl_msa_Destroy(msa);
}
}
if (ofp != stdout) printf("\nRetrieved %d alignments.\n", nali);
esl_keyhash_Destroy(keys);
esl_fileparser_Close(efp);
return;
}
/* Function: esl_paml_ReadE()
* Incept: SRE, Fri Jul 9 09:27:24 2004 [St. Louis]
*
* Purpose: Read an amino acid rate matrix in PAML format from stream
* <fp>. Return it in two pieces: the symmetric E
* exchangeability matrix in <E>, and the stationary
* probability vector $\pi$ in <pi>.
* Caller provides the memory for both <E> and <pi>. <E>
* is a $20 \times 20$ matrix allocated as
* <esl_dmatrix_Create(20, 20)>. <pi> is an array with
* space for at least 20 doubles.
*
* The <E> matrix is symmetric for off-diagonal elements:
* $E_{ij} = E_{ij}$ for $i \neq j$. The on-diagonal
* elements $E_{ii}$ are not valid and should not be
* accessed. (They are set to zero.)
* The rate matrix will later be obtained from <E>
* and <pi> as
* $Q_{ij} = E_{ij} \pi_j$ for $i \neq j$
* and
* $Q_{ii} = -\sum_{j \neq i} Q_{ij}$
* then scaled to units of one
* substitution/site; see <esl_ratemx_E2Q()> and
* <esl_ratemx_ScaleTo()>.
*
* Data file format: First 190 numbers are a
* lower-triangular matrix E of amino acid
* exchangeabilities $E_{ij}$. Next 20 numbers are the
* amino acid frequencies $\pi_i$. Remainder of the
* datafile is ignored.
*
* The alphabet order in the matrix and the frequency
* vector is assumed to be "ARNDCQEGHILKMFPSTWYV"
* (alphabetical by three-letter code), which appears to be
* PAML's default order. This is transformed to Easel's
* "ACDEFGHIKLMNPQRSTVWY" (alphabetical by one-letter code)
* in the $E_{ij}$ and $\pi_i$ that are returned.
*
* Args: fp - open datafile for reading.
* E - RETURN: E matrix of amino acid exchangeabilities e_ij,
* symmetric (E_ij = E_ji),
* in Easel amino acid alphabet order A..Y.
* Caller provides appropriately allocated space.
* pi - RETURN: \pi_i vector of amino acid frequencies,
* in Easel amino acid alphabet order A..Y.
* Caller provides appropriately allocated space.
*
* Returns: <eslOK> on success.
* Returns <eslEOF> on premature end of file (parse failed), in which
* case the contents of <E> and <pi> are undefined.
*
* Throws: <eslEMEM> on internal allocation failure,
* and the contents of <E> and <pi> are undefined.
*
* Xref: STL8/p.56.
*/
int
esl_paml_ReadE(FILE *fp, ESL_DMATRIX *E, double *pi)
{
int status;
ESL_FILEPARSER *efp = NULL;
char *tok;
int i,j;
char *pamlorder = "ARNDCQEGHILKMFPSTWYV";
char *eslorder = "ACDEFGHIKLMNPQRSTVWY";
int perm[20];
if ((status = esl_dmatrix_SetZero(E)) != eslOK) goto ERROR;
esl_vec_DSet(pi, 20, 0.);
if ((efp = esl_fileparser_Create(fp)) == NULL) goto ERROR;
if ((status = esl_fileparser_SetCommentChar(efp, '#')) != eslOK) goto ERROR;
/* Construct the alphabet permutation we need.
* perm[i] -> original row/column i goes to row/column perm[i]
*/
for (i = 0; i < 20; i++)
perm[i] = (int) (strchr(eslorder, pamlorder[i]) - eslorder);
/* Read the s_ij matrix data in, permuting as we go. */
for (i = 1; i < 20; i++)
for (j = 0; j < i; j++)
{
if ((status = esl_fileparser_GetToken(efp, &tok, NULL)) != eslOK) goto ERROR;
E->mx[perm[i]][perm[j]] = atof(tok);
E->mx[perm[j]][perm[i]] = E->mx[perm[i]][perm[j]];
}
/* Read the pi_i vector in, permuting as we read. */
for (i = 0; i < 20; i++)
{
if ((status = esl_fileparser_GetToken(efp, &tok, NULL)) != eslOK) goto ERROR;
pi[perm[i]] = atof(tok);
}
esl_fileparser_Destroy(efp);
return eslOK;
ERROR:
if (efp != NULL) esl_fileparser_Destroy(efp);
return status;
}
static void
read_tabfile(ESL_GETOPTS *go, char *tabfile, ESL_KEYHASH *kh, ESL_DMATRIX *D)
{
ESL_FILEPARSER *efp = NULL;
int nline = 0;
int vfield = esl_opt_GetInteger(go, "-v");
int qfield = esl_opt_GetInteger(go, "-q");
int tfield = esl_opt_GetInteger(go, "-t");
char *tok;
int toklen;
int ntok;
double value;
int qidx, tidx;
if (esl_fileparser_Open(tabfile, NULL, &efp) != eslOK) esl_fatal("File open failed");
esl_fileparser_SetCommentChar(efp, '#');
esl_dmatrix_Set(D, eslINFINITY);
while (esl_fileparser_NextLine(efp) == eslOK)
{
nline++;
ntok = 0;
qidx = tidx = -1;
value = eslNaN;
while (esl_fileparser_GetTokenOnLine(efp, &tok, &toklen) == eslOK)
{
ntok++;
if (ntok == vfield) value = atof(tok);
if (ntok == qfield && esl_keyhash_Lookup(kh, tok, toklen, &qidx) != eslOK) esl_fatal("failed to find query key %s", tok);
if (ntok == tfield && esl_keyhash_Lookup(kh, tok, toklen, &tidx) != eslOK) esl_fatal("failed to find target key %s", tok);
}
if (qidx == -1) esl_fatal("Failed to find query name on line %d (looking for field %d)\n", nline, qfield);
if (tidx == -1) esl_fatal("Failed to find target name on line %d (looking for field %d)\n", nline, tfield);
if (isnan(value)) esl_fatal("Failed to find value on line %d (looking for field %d)\n", nline, vfield);
D->mx[qidx][tidx] = value;
if (D->mx[tidx][qidx] == eslINFINITY) D->mx[tidx][qidx] = value;
}
esl_fileparser_Close(efp);
}
/* Function: esl_hyperexp_ReadFile()
*
* Purpose: Convenience wrapper around <esl_hyperexp_Read()> that takes
* a filename as an argument, instead of an open <ESL_FILEPARSER>.
*
* This lets you quickly read an object from a file, but it
* limits your ability to deal gracefully and flexibly with
* 'normal' errors like 'file not found' or 'bad file format'.
* Here, all errors are fatal.
*
* Returns: <eslOK> on success.
*
* Throws: <eslEMEM> on an allocation failure.
*
* <eslEFORMAT> on any parse error. Diagnostic information is
* unavailable, because the <ESL_FILEPARSER> that's holding
* that information is internal to this function.
*
* <eslENOTFOUND> on any failure to open the file.
*/
int
esl_hyperexp_ReadFile(char *filename, ESL_HYPEREXP **ret_hxp)
{
FILE *fp;
ESL_FILEPARSER *e;
int status;
if ((fp = fopen(filename, "r")) == NULL)
ESL_EXCEPTION(eslENOTFOUND, "file not found");
if ((e = esl_fileparser_Create(fp)) == NULL) {
fclose(fp);
ESL_EXCEPTION(eslEMEM, "failed to create fileparser");
}
esl_fileparser_SetCommentChar(e, '#');
status = esl_hyperexp_Read(e, ret_hxp);
esl_fileparser_Destroy(e);
fclose(fp);
return status;
}
static void
multifetch_subseq(ESL_GETOPTS *go, FILE *ofp, char *gdffile, ESL_SQFILE *sqfp)
{
ESL_FILEPARSER *efp = NULL;
char *newname;
char *s;
int n1, n2;
int start, end;
char *source;
if (esl_fileparser_Open(gdffile, NULL, &efp) != eslOK) esl_fatal("Failed to open key file %s\n", gdffile);
esl_fileparser_SetCommentChar(efp, '#');
while (esl_fileparser_NextLine(efp) == eslOK)
{
if (esl_fileparser_GetTokenOnLine(efp, &newname, &n1) != eslOK)
esl_fatal("Failed to read subseq name on line %d of file %s\n", efp->linenumber, gdffile);
if (esl_fileparser_GetTokenOnLine(efp, &s, NULL) != eslOK)
esl_fatal("Failed to read start coord on line %d of file %s\n", efp->linenumber, gdffile);
start = atoi(s);
if(start <= 0)
esl_fatal("Read invalid start coord %d on line %d of file %s (must be positive integer)\n", start, efp->linenumber, gdffile);
if (esl_fileparser_GetTokenOnLine(efp, &s, NULL) != eslOK)
esl_fatal("Failed to read end coord on line %d of file %s\n", efp->linenumber, gdffile);
end = atoi(s);
if(end < 0)
esl_fatal("Read invalid end coord %d on line %d of file %s (must be positive integer, or 0 for full length)\n", end, efp->linenumber, gdffile);
if (esl_fileparser_GetTokenOnLine(efp, &source, &n2) != eslOK)
esl_fatal("Failed to read source seq name on line %d of file %s\n", efp->linenumber, gdffile);
onefetch_subseq(go, ofp, sqfp, newname, source, start, end);
}
esl_fileparser_Close(efp);
}
static void
read_keyfile(ESL_GETOPTS *go, char *keyfile, ESL_KEYHASH *kh)
{
ESL_FILEPARSER *efp = NULL;
int nline = 0;
char *tok = NULL;
int toklen;
int status;
if (esl_fileparser_Open(keyfile, NULL, &efp) != eslOK) esl_fatal("File open failed");
esl_fileparser_SetCommentChar(efp, '#');
while (esl_fileparser_NextLine(efp) == eslOK)
{
nline++;
if (esl_fileparser_GetTokenOnLine(efp, &tok, &toklen) != eslOK) esl_fatal("No token found on line %d", nline);
status = esl_keyhash_Store(kh, tok, toklen, NULL);
if (status == eslEDUP) esl_fatal("Saw key %s twice: keys must be unique", tok);
else if (status != eslOK) esl_fatal("unknown error in storing key %s\n", tok);
}
esl_fileparser_Close(efp);
}
int
main(int argc, char **argv)
{
ESL_GETOPTS *go = NULL; /* application configuration */
char *seqfile = NULL; /* sequence file name */
char *maskfile = NULL; /* mask coordinate file name */
int infmt = eslSQFILE_UNKNOWN; /* format code for seqfile */
int outfmt = eslSQFILE_FASTA; /* format code for output seqs */
ESL_SQFILE *sqfp = NULL; /* open sequence file */
ESL_FILEPARSER *maskefp = NULL; /* open mask coord file */
FILE *ofp = NULL; /* output stream for masked seqs */
char *source = NULL; /* name of current seq to mask */
char *p1, *p2; /* pointers used in parsing */
int64_t start, end; /* start, end coord for masking */
int64_t i, j, pos; /* coords in a sequence */
int64_t overmask; /* # of extra residues to mask */
ESL_SQ *sq = esl_sq_Create(); /* current sequence */
int do_fetching;
int do_lowercase;
int maskchar;
int status; /* easel return code */
/****************************************************************************
* Parse command line
****************************************************************************/
go = esl_getopts_Create(options);
if (esl_opt_ProcessCmdline(go, argc, argv) != eslOK) cmdline_failure(argv[0], "Failed to parse command line: %s\n", go->errbuf);
if (esl_opt_VerifyConfig(go) != eslOK) cmdline_failure(argv[0], "Error in configuration: %s\n", go->errbuf);
if (esl_opt_GetBoolean(go, "-h") ) cmdline_help (argv[0], go);
if (esl_opt_ArgNumber(go) != 2) cmdline_failure(argv[0], "Incorrect number of command line arguments.\n");
do_fetching = esl_opt_GetBoolean(go, "-R");
do_lowercase = esl_opt_GetBoolean(go, "-l");
overmask = (esl_opt_IsOn(go, "-x") ? esl_opt_GetInteger(go, "-x") : 0);
maskchar = (esl_opt_IsOn(go, "-m") ? esl_opt_GetChar(go, "-m") : 'X');
seqfile = esl_opt_GetArg(go, 1);
maskfile = esl_opt_GetArg(go, 2);
/* Open the <seqfile>: text mode, not digital */
if (esl_opt_GetString(go, "--informat") != NULL) {
infmt = esl_sqio_EncodeFormat(esl_opt_GetString(go, "--informat"));
if (infmt == eslSQFILE_UNKNOWN) cmdline_failure(argv[0], "%s is not a valid input sequence file format for --informat");
}
status = esl_sqfile_Open(seqfile, infmt, NULL, &sqfp);
if (status == eslENOTFOUND) cmdline_failure(argv[0], "Sequence file %s not found.\n", seqfile);
else if (status == eslEFORMAT) cmdline_failure(argv[0], "Format of file %s unrecognized.\n", seqfile);
else if (status == eslEINVAL) cmdline_failure(argv[0], "Can't autodetect stdin or .gz.\n");
else if (status != eslOK) cmdline_failure(argv[0], "Open failed, code %d.\n", status);
if(do_fetching)
{
status = esl_sqfile_OpenSSI(sqfp, NULL);
if (status == eslEFORMAT) cmdline_failure(argv[0], "SSI index is in incorrect format\n");
else if (status == eslERANGE) cmdline_failure(argv[0], "SSI index is in 64-bit format and we can't read it\n");
else if (status != eslOK) cmdline_failure(argv[0], "Failed to open SSI index\n");
}
/* Open the <maskfile> */
if (esl_fileparser_Open(maskfile, NULL, &maskefp) != eslOK)
cmdline_failure(argv[0], "Failed to open mask coordinate file %s\n", maskfile);
esl_fileparser_SetCommentChar(maskefp, '#');
/* Open the output file, if any */
if (esl_opt_GetString(go, "-o") != NULL)
{
if ((ofp = fopen(esl_opt_GetString(go, "-o"), "w")) == NULL)
cmdline_failure(argv[0], "Failed to open output file %s\n", esl_opt_GetString(go, "-o"));
}
else ofp = stdout;
/****************************************************************************
* Main loop over lines in <maskfile>
****************************************************************************/
/* Read one data line at a time from the <maskfile>;
* parse into data fields <seqname> <start> <end>
*/
while (esl_fileparser_NextLine(maskefp) == eslOK)
{
/* First field is sequence name */
if (esl_fileparser_GetTokenOnLine(maskefp, &source, NULL) != eslOK)
esl_fatal("Failed to read source seq name on line %d of file %s\n", maskefp->linenumber, maskfile);
/* Get the sequence */
if (do_fetching)
{ /* If the <seqfile> is SSI indexed, try to reposition it and read <source> seq by random access */
status = esl_sqio_Fetch(sqfp, source, sq);
if (status == eslENOTFOUND) esl_fatal("seq %s not found in SSI index for file %s\n", source, sqfp->filename);
else if (status == eslEINVAL) esl_fatal("No SSI index or can't reposition in file %s\n", sqfp->filename);
else if (status == eslEFORMAT) esl_fatal("Parse failed:\n%s\n", esl_sqfile_GetErrorBuf(sqfp));
else if (status != eslOK) esl_fatal("Unexpected failure in fetching %s from file %s\n", source, sqfp->filename);
}
else
{ /* else, assume we're reading sequentially; <sqfile> and <maskfile> have seqs in same order */
status = esl_sqio_Read(sqfp, sq);
if (status == eslEOF) esl_fatal("File %s ended prematurely; didn't find %s\n", sqfp->filename, source);
else if (status == eslEFORMAT) esl_fatal("Parse failed:\n%s\n", esl_sqfile_GetErrorBuf(sqfp));
else if (status != eslOK) esl_fatal("Unexpected error reading sequence file %s\n", sqfp->filename);
if ((strcmp(sq->name, source) != 0) && (strcmp(sq->acc, source) != 0))
esl_fatal("Sequences in <sqfile> and <maskfile> aren't in same order; try -R");
}
/* If we're masking by lowercase, first make sure everything's uppercase */
if (do_lowercase)
for (pos = 0; pos < sq->n; pos++)
if (isalpha(sq->seq[pos]))
sq->seq[pos] = toupper(sq->seq[pos]);
/* Next two fields are <start>, <end> for the masking */
/* possible future extension: wrap loop around this, enable multiple masked regions */
if (esl_fileparser_GetTokenOnLine(maskefp, &p1, NULL) != eslOK)
esl_fatal("Failed to read start coord on line %d of file %s\n", maskefp->linenumber, maskfile);
start = strtoll(p1, &p2, 0) - 1;
if (esl_fileparser_GetTokenOnLine(maskefp, &p2, NULL) != eslOK)
esl_fatal("Failed to read end coord on line %d of file %s\n", maskefp->linenumber, maskfile);
end = strtoll(p2, &p1, 0) - 1;
/* Do the masking */
if (esl_opt_GetBoolean(go, "-r")) /* Reverse masking */
{ /* leave start..end unmasked; mask prefix 0..start-1, end+1..L-1 */
i = 0;
j = ESL_MIN(sq->n-1, start - 1 + overmask);
for (pos = i; pos <= j; pos++)
if (isalpha(sq->seq[pos]))
sq->seq[pos] = (do_lowercase ? tolower(sq->seq[pos]) : maskchar);
i = ESL_MAX(0, end + 1 - overmask);
j = sq->n-1;
for (pos = i; pos <= j; pos++)
if (isalpha(sq->seq[pos]))
sq->seq[pos] = (do_lowercase ? tolower(sq->seq[pos]) : maskchar);
}
else
{ /* normal: mask start..end */
i = ESL_MAX(0, start - overmask);
j = ESL_MIN(sq->n-1, end + overmask);
for (pos = i; pos <= j; pos++)
if (isalpha(sq->seq[pos]))
sq->seq[pos] = (do_lowercase ? tolower(sq->seq[pos]) : maskchar);
}
esl_sqio_Write(ofp, sq, outfmt, FALSE);
esl_sq_Reuse(sq);
}
esl_sq_Destroy(sq);
esl_fileparser_Close(maskefp);
esl_sqfile_Close(sqfp);
esl_getopts_Destroy(go);
if (ofp != stdout) fclose(ofp);
return 0;
}
/* Function: p7_bg_Read()
* Synopsis: Read background frequencies from a file.
*
* Purpose: Read new background frequencies from file <bgfile>,
* overwriting the frequencies previously in the
* <P7_BG> object <bg>.
*
* Note that <bg> is already created by the caller, not
* created here. Also note that <p7_bg_Read()> only reads
* residue background frequencies used for the "null
* model", whereas a <P7_BG> object contains additional
* information for the bias filter and for the biased
* composition correction.
*
* Args: bgfile - file to read.
* bg - existing <P7_BG> object provided by the caller.
* errbuf - OPTIONAL: space for an error message, upon parse errors; or NULL.
*
* Returns: <eslOK> on success, and background frequencies in <bg>
* are overwritten.
*
* <eslENOTFOUND> if <bgfile> can't be opened for reading.
* <eslEFORMAT> if parsing of <bgfile> fails for some
* reason. In both cases, <errbuf> contains a
* user-directed error message upon return, including (if
* relevant) the file name <bgfile> and the line number on
* which an error was detected. <bg> is unmodified.
*
* Throws: <eslEMEM> on allocation failure; <bg> is unmodified,
* and <errbuf> is empty.
*/
int
p7_bg_Read(char *bgfile, P7_BG *bg, char *errbuf)
{
ESL_FILEPARSER *efp = NULL;
float *fq = NULL;
int n = 0;
char *tok;
int toklen;
int alphatype;
ESL_DSQ x;
int status;
if (errbuf) errbuf[0] = '\0';
status = esl_fileparser_Open(bgfile, NULL, &efp);
if (status == eslENOTFOUND) ESL_XFAIL(eslENOTFOUND, errbuf, "couldn't open bg file %s for reading", bgfile);
else if (status != eslOK) goto ERROR;
esl_fileparser_SetCommentChar(efp, '#');
/* First token is alphabet type: amino | DNA | RNA */
status = esl_fileparser_GetToken(efp, &tok, &toklen);
if (status == eslEOF) ESL_XFAIL(eslEFORMAT, errbuf, "premature end of file [line %d of bgfile %s]", efp->linenumber, bgfile);
else if (status != eslOK) goto ERROR;
alphatype = esl_abc_EncodeType(tok);
if (alphatype == eslUNKNOWN) ESL_XFAIL(eslEFORMAT, errbuf, "expected alphabet type but saw \"%s\" [line %d of bgfile %s]", tok, efp->linenumber, bgfile);
else if (alphatype != bg->abc->type) ESL_XFAIL(eslEFORMAT, errbuf, "bg file's alphabet is %s; expected %s [line %d, %s]", tok, esl_abc_DecodeType(bg->abc->type), efp->linenumber, bgfile);
ESL_ALLOC(fq, sizeof(float) * bg->abc->K);
esl_vec_FSet(fq, bg->abc->K, -1.0);
while ((status = esl_fileparser_NextLine(efp)) == eslOK)
{
status = esl_fileparser_GetTokenOnLine(efp, &tok, &toklen);
if (status == eslEOL) ESL_XFAIL(eslEFORMAT, errbuf, "premature end of file [line %d of bgfile %s", efp->linenumber, bgfile);
else if (status != eslOK) goto ERROR;
if (toklen != 1 || ! esl_abc_CIsCanonical(bg->abc, *tok))
ESL_XFAIL(eslEFORMAT, errbuf, "expected to parse a residue letter; saw %s [line %d of bgfile %s]", tok, efp->linenumber, bgfile);
x = esl_abc_DigitizeSymbol(bg->abc, *tok);
if (fq[x] != -1.0) ESL_XFAIL(eslEFORMAT, errbuf, "already parsed probability of %c [line %d of bgfile %s]", bg->abc->sym[x], efp->linenumber, bgfile);
n++;
status = esl_fileparser_GetTokenOnLine(efp, &tok, &toklen);
if (status == eslEOL) ESL_XFAIL(eslEFORMAT, errbuf, "premature end of file, expected a probability [line %d of bgfile %s]", efp->linenumber, bgfile);
else if (status != eslOK) goto ERROR;
if (! esl_str_IsReal(tok)) ESL_XFAIL(eslEFORMAT, errbuf, "expected a probability, saw %s [line %d of bgfile %s]", tok, efp->linenumber, bgfile);
fq[x] = atof(tok);
status = esl_fileparser_GetTokenOnLine(efp, &tok, &toklen);
if (status == eslOK) ESL_XFAIL(eslEFORMAT, errbuf, "extra unexpected data found [line %d of bgfile %s]", efp->linenumber, bgfile);
else if (status != eslEOL) goto ERROR;
}
if (status != eslEOF) goto ERROR;
if ( n != bg->abc->K)
ESL_XFAIL(eslEFORMAT, errbuf, "expected %d residue frequencies, but found %d in bgfile %s", bg->abc->K, n, bgfile);
if ( esl_FCompare(esl_vec_FSum(fq, bg->abc->K), 1.0, 0.001) != eslOK)
ESL_XFAIL(eslEFORMAT, errbuf, "residue frequencies do not sum to 1.0 in bgfile %s", bgfile);
/* all checking complete. no more error cases. overwrite bg with the new frequencies */
esl_vec_FNorm(fq, bg->abc->K);
esl_vec_FCopy(fq, bg->abc->K, bg->f);
free(fq);
esl_fileparser_Close(efp);
return eslOK;
ERROR:
if (fq) free(fq);
if (efp) esl_fileparser_Close(efp);
return status;
}
/* multifetch:
* given a file containing lines with one name or key per line;
* parse the file line-by-line;
* if we have an SSI index available, retrieve the seqs by key
* as we see each line;
* else, without an SSI index, store the keys in a hash, then
* read the entire seq file in a single pass, outputting seqs
* that are in our keylist.
*
* Note that with an SSI index, you get the seqs in the order they
* appear in the <keyfile>, but without an SSI index, you get seqs in
* the order they occur in the seq file.
*/
static void
multifetch(ESL_GETOPTS *go, FILE *ofp, char *keyfile, ESL_SQFILE *sqfp)
{
ESL_KEYHASH *keys = esl_keyhash_Create();
ESL_FILEPARSER *efp = NULL;
int nseq = 0;
int nkeys = 0;
char *key;
int keylen;
int keyidx;
int status;
if (esl_fileparser_Open(keyfile, NULL, &efp) != eslOK) esl_fatal("Failed to open key file %s\n", keyfile);
esl_fileparser_SetCommentChar(efp, '#');
while (esl_fileparser_NextLine(efp) == eslOK)
{
if (esl_fileparser_GetTokenOnLine(efp, &key, &keylen) != eslOK)
esl_fatal("Failed to read seq name on line %d of file %s\n", efp->linenumber, keyfile);
status = esl_keyhash_Store(keys, key, keylen, &keyidx);
if (status == eslEDUP) esl_fatal("seq key %s occurs more than once in file %s\n", key, keyfile);
/* if we have an SSI index, just fetch them as we go. */
if (sqfp->data.ascii.ssi != NULL) { onefetch(go, ofp, key, sqfp); nseq++; }
nkeys++;
}
/* If we don't have an SSI index, we haven't fetched anything yet; do it now. */
if (sqfp->data.ascii.ssi == NULL)
{
ESL_SQ *sq = esl_sq_Create();
while ((status = esl_sqio_Read(sqfp, sq)) == eslOK)
{
if ( (sq->name[0] != '\0' && esl_keyhash_Lookup(keys, sq->name, -1, NULL) == eslOK) ||
(sq->acc[0] != '\0' && esl_keyhash_Lookup(keys, sq->acc, -1, NULL) == eslOK))
{
if (esl_opt_GetBoolean(go, "-r") )
if (esl_sq_ReverseComplement(sq) != eslOK)
esl_fatal("Failed to reverse complement %s\n", sq->name);
esl_sqio_Write(ofp, sq, eslSQFILE_FASTA, FALSE);
nseq++;
}
esl_sq_Reuse(sq);
}
if (status == eslEFORMAT) esl_fatal("Parse failed (sequence file %s):\n%s\n",
sqfp->filename, esl_sqfile_GetErrorBuf(sqfp));
else if (status != eslEOF) esl_fatal("Unexpected error %d reading sequence file %s",
status, sqfp->filename);
esl_sq_Destroy(sq);
}
if (nkeys != nseq) esl_fatal("Tried to retrieve %d keys, but only retrieved %d sequences\n", nkeys, nseq);
if (ofp != stdout) printf("\nRetrieved %d sequences.\n", nseq);
esl_keyhash_Destroy(keys);
esl_fileparser_Close(efp);
return;
}
