95030f73f067d8e8f9221a09c1811784607e4e72 markd Fri Jul 10 17:51:29 2020 -0700 split dynamic server genomeDataPrefix into genome and genomeDataDir. More verbose, but clearer diff --git src/gfServer/gfServer.c src/gfServer/gfServer.c index a189186..8b11e35 100644 --- src/gfServer/gfServer.c +++ src/gfServer/gfServer.c @@ -86,44 +86,49 @@ " gfServer protQuery host port probe.fa\n" " To query a server with translated DNA sequence:\n" " gfServer transQuery host port probe.fa\n" " To query server with PCR primers:\n" " gfServer pcr host port fPrimer rPrimer maxDistance\n" " To process one probe fa file against a .2bit format genome (not starting server):\n" " gfServer direct probe.fa file(s).2bit\n" " To test PCR without starting server:\n" " gfServer pcrDirect fPrimer rPrimer file(s).2bit\n" " To figure out usage level:\n" " gfServer status host port\n" " To get input file list:\n" " gfServer files host port\n" " To generate a precomputed index:\n" " gfServer index gfidx file(s)\n" - " where the files are .2bit or .nib format files. Separate indexes must be created\n" - " for untranslated and translated queries. These can be used with a persistent server\n" - " as with 'start -indexFile or a dynamic server. They must follow the naming convention for\n" - " for dynamic servers.\n" - " To run a dynamic server (usually called by xinet):\n" + " where the files are .2bit or .nib format files. Separate indexes are\n" + " be created for untranslated and translated queries. These can be used\n" + " with a persistent server as with 'start -indexFile or a dynamic server.\n" + " They must follow the naming convention for for dynamic servers.\n" + " To run a dynamic server (usually called by xinetd):\n" " gfServer dynserver rootdir\n" - " The root directory must contain directories for each genome with the twobit and index\n" - " files following the convention:\n" - " $rootdir/$containingDirs/$genome/$genome.2bit\n" - " $rootdir/$containingDirs/$genome/$genome.untrans.gfidx\n" - " $rootdir/$containingDirs/$genome/$genome.trans.gfidx\n" + " Data files for genomes are found relative to the root directory.\n" + " Queries are made using the prefix of the file path relative to the root\n" + " directory. The files $genome.2bit, $genome.untrans.gfidx, and\n" + " $genome.trans.gfidx are required. Typically the structure will be in\n" + " the form:\n" + " $rootdir/$genomeDataDir/$genome.2bit\n" + " $rootdir/$genomeDataDir/$genome.untrans.gfidx\n" + " $rootdir/$genomeDataDir/$genome.untrans.gfidx\n" + " in this case, one would call gfClient with \n" + " -genome=$genome -genomeDataDir=$genomeDataDir\n" " Where the contain directories are optional.\n" " The -perSeqMax functionality can be implemented by creating a file\n" - " $rootdir/$genome/$genome.perseqmax\n" + " $rootdir/$genomeDataDir/$genome.perseqmax\n" "\n" "options:\n" " -tileSize=N Size of n-mers to index. Default is 11 for nucleotides, 4 for\n" " proteins (or translated nucleotides).\n" " -stepSize=N Spacing between tiles. Default is tileSize.\n" " -minMatch=N Number of n-mer matches that trigger detailed alignment.\n" " Default is 2 for nucleotides, 3 for proteins.\n" " -maxGap=N Number of insertions or deletions allowed between n-mers.\n" " Default is 2 for nucleotides, 0 for proteins.\n" " -trans Translate database to protein in 6 frames. Note: it is best\n" " to run this on RepeatMasked data in this case.\n" " -log=logFile Keep a log file that records server requests.\n" " -seqLog Include sequences in log file (not logged with -syslog).\n" " -ipLog Include user's IP in log file (not logged with -syslog).\n" " -debugLog Include debugging info in log file.\n" @@ -1048,74 +1053,73 @@ buf[msgLen] = '\0'; logError("%s", buf); printf("Error: %s\n", buf); } static void dynReadBytes(char *buf, int bufSize) /* read pending bytes */ { int readSize = read(STDIN_FILENO, buf, bufSize-1); if (readSize < 0) errAbort("EOF from client"); buf[readSize] = '\0'; } static void dynReadCommand(char **commandRet, int *qsizeRet, boolean *isTransRet, - char **genomeNameRet, char **dynGenomeDirRet) + char **genomeRet, char **genomeDataDirRet) /* read query request from stdin, same as server expect includes database * Format for query commands: - * signature+command qsize dynGenomeDir + * signature+command qsize genome genomeDataDir * Formats for info command: - * signature+command dynGenomeDir + * signature+command genome genomeDataDir */ { char buf[256]; dynReadBytes(buf, sizeof(buf)); logDebug("query: %s", buf); if (!startsWith(gfSignature(), buf)) errAbort("query does not start with signature, got '%s'", buf); -char *words[5]; +char *words[6]; int numWords = chopByWhite(buf, words, ArraySize(words)); if (numWords == 0) errAbort("empty command"); char *command = buf + strlen(gfSignature()); *commandRet = cloneString(command); *isTransRet = sameString("protQuery", command) || sameString("transQuery", command) || sameString("transInfo", command); if (sameString("query", command) || sameString("protQuery", command) || sameString("transQuery", command)) { - if (numWords != 3) - errAbort("expected 3 words in query command, got %d", numWords); + if (numWords != 4) + errAbort("expected 4 words in query command, got %d", numWords); *qsizeRet = atoi(words[1]); - *dynGenomeDirRet = cloneString(words[2]); + *genomeRet = cloneString(words[2]); + *genomeDataDirRet = cloneString(words[3]); } else if (sameString("untransInfo", command) || sameString("transInfo", command)) { - if (numWords != 2) - errAbort("expected 2 words in query command, got %d", numWords); + if (numWords != 3) + errAbort("expected 3 words in query command, got %d", numWords); *qsizeRet = 0; - *dynGenomeDirRet = cloneString(words[1]); + *genomeRet = cloneString(words[1]); + *genomeDataDirRet = cloneString(words[2]); } else errAbort("invalid command '%s'", command); - -// parse genomeName out of directory -*genomeNameRet = cloneString(findTail(*dynGenomeDirRet, '/')); } static struct dnaSeq* dynReadQuerySeq(int qSize, boolean isTrans, boolean queryIsProt) /* read the DNA sequence from the query, filtering junk */ { struct dnaSeq *seq; AllocVar(seq); seq->size = qSize; seq->dna = needLargeMem(qSize+1); if (gfReadMulti(STDIN_FILENO, seq->dna, qSize) != qSize) errAbort("read of %d bytes of query sequence failed", qSize); seq->dna[qSize] = '\0'; if (queryIsProt) { @@ -1125,122 +1129,121 @@ else { seq->size = dnaFilteredSize(seq->dna); dnaFilter(seq->dna, seq->dna); } int maxSize = (isTrans ? maxAaSize : maxNtSize); if (seq->size > maxSize) { seq->size = maxSize; seq->dna[maxSize] = 0; } return seq; } -static void dynGetDataFiles(char *rootDir, char *genomeName, char *dynGenomeDir, +static void dynGetDataFiles(char *rootDir, char *genome, char *genomeDataDir, boolean isTrans, char gfIdxFile[PATH_LEN], struct hash **perSeqMaxHashRet) /* get paths for sequence files to handle requests and validate they exist */ { char seqFile[PATH_LEN]; -safef(seqFile, PATH_LEN, "%s/%s/%s.2bit", rootDir, dynGenomeDir, genomeName); +safef(seqFile, PATH_LEN, "%s/%s/%s.2bit", rootDir, genomeDataDir, genome); if (!fileExists(seqFile)) - errAbort("sequence file for %s does not exist: %s", genomeName, seqFile); + errAbort("sequence file for %s does not exist: %s", genome, seqFile); -safef(gfIdxFile, PATH_LEN, "%s/%s/%s.%s.gfidx", rootDir, dynGenomeDir, genomeName, isTrans ? "trans" : "untrans"); +safef(gfIdxFile, PATH_LEN, "%s/%s/%s.%s.gfidx", rootDir, genomeDataDir, genome, isTrans ? "trans" : "untrans"); if (!fileExists(gfIdxFile)) - errAbort("gf index file for %s does not exist: %s", genomeName, gfIdxFile); + errAbort("gf index file for %s does not exist: %s", genome, gfIdxFile); char perSeqMaxFile[PATH_LEN]; -safef(perSeqMaxFile, PATH_LEN, "%s/%s/%s.perseqmax", rootDir, dynGenomeDir, genomeName); +safef(perSeqMaxFile, PATH_LEN, "%s/%s/%s.perseqmax", rootDir, genomeDataDir, genome); *perSeqMaxHashRet = NULL; if (fileExists(perSeqMaxFile)) { /* only the basename of the file is saved in the index */ char *slash = strrchr(seqFile, '/'); char *seqFiles[1] = {(slash != NULL) ? slash + 1 : seqFile}; *perSeqMaxHashRet = buildPerSeqMax(1, seqFiles, perSeqMaxFile); } } -static void dynamicServerQuery(char *command, int qSize, char *genomeName, - struct genoFindIndex *gfIdx, struct hash *perSeqMaxHash) +static void dynamicServerQuery(char *command, int qSize, struct genoFindIndex *gfIdx, struct hash *perSeqMaxHash) /* handle search queries */ { mustWriteFd(STDOUT_FILENO, "Y", 1); boolean queryIsProt = sameString(command, "protQuery"); struct dnaSeq* seq = dynReadQuerySeq(qSize, gfIdx->isTrans, queryIsProt); if (gfIdx->isTrans) { if (queryIsProt) transQuery(gfIdx->transGf, seq, STDOUT_FILENO); else transTransQuery(gfIdx->transGf, seq, STDOUT_FILENO); } else { dnaQuery(gfIdx->untransGf, seq, STDOUT_FILENO, perSeqMaxHash); } netSendString(STDOUT_FILENO, "end"); logDebug("query done"); } -static void dynamicServerInfo(char *command, char *genomeName, struct genoFindIndex *gfIdx) +static void dynamicServerInfo(char *command, struct genoFindIndex *gfIdx) /* handle one of the info commands */ { char buf[256]; struct genoFind *gf = gfIdx->isTrans ? gfIdx->transGf[0][0] : gfIdx->untransGf; sprintf(buf, "version %s", gfVersion); netSendString(STDOUT_FILENO, buf); sprintf(buf, "type %s", (gfIdx->isTrans ? "translated" : "nucleotide")); netSendString(STDOUT_FILENO, buf); sprintf(buf, "tileSize %d", gf->tileSize); netSendString(STDOUT_FILENO, buf); sprintf(buf, "stepSize %d", gf->stepSize); netSendString(STDOUT_FILENO, buf); sprintf(buf, "minMatch %d", gf->minMatch); netSendString(STDOUT_FILENO, buf); netSendString(STDOUT_FILENO, "end"); } static void dynamicServer(char* rootDir) /* dynamic server for inetd. Read query from stdin, open index, query, respond, exit. * only one query at a time */ { // make sure error is logged pushWarnHandler(dynWarnErrorVa); -char *command, *genomeName, *dynGenomeDir; +char *command, *genome, *genomeDataDir; int qSize; boolean isTrans; -dynReadCommand(&command, &qSize, &isTrans, &genomeName, &dynGenomeDir); -logInfo("dynserver: %s %s %s %s size=%d ", command, genomeName, dynGenomeDir, (isTrans ? "trans" : "untrans"), qSize); +dynReadCommand(&command, &qSize, &isTrans, &genome, &genomeDataDir); +logInfo("dynserver: %s %s %s %s size=%d ", command, genome, genomeDataDir, (isTrans ? "trans" : "untrans"), qSize); time_t startTime = clock1000(); char gfIdxFile[PATH_LEN]; struct hash *perSeqMaxHash = NULL; -dynGetDataFiles(rootDir, genomeName, dynGenomeDir, isTrans, gfIdxFile, &perSeqMaxHash); +dynGetDataFiles(rootDir, genome, genomeDataDir, isTrans, gfIdxFile, &perSeqMaxHash); logInfo("dynserver: index loading completed in %4.3f seconds", 0.001 * (clock1000() - startTime)); startTime = clock1000(); struct genoFindIndex *gfIdx = genoFindIndexLoad(gfIdxFile, isTrans); if (endsWith(command, "Info")) - dynamicServerInfo(command, genomeName, gfIdx); + dynamicServerInfo(command, gfIdx); else - dynamicServerQuery(command, qSize, genomeName, gfIdx, perSeqMaxHash); + dynamicServerQuery(command, qSize, gfIdx, perSeqMaxHash); logInfo("dynserver: %s completed in %4.3f seconds", command, 0.001 * (clock1000() - startTime)); } int main(int argc, char *argv[]) /* Process command line. */ { char *command; gfCatchPipes(); dnaUtilOpen(); optionInit(&argc, argv, optionSpecs); command = argv[1]; if (optionExists("trans")) { doTrans = TRUE;