8ba095f8ff62a04a522daf2a62e74968a59d21f2 galt Wed Jun 12 14:27:47 2013 -0700 cleanup debug stuff diff --git src/hg/lib/hgFind.c src/hg/lib/hgFind.c index eb79c9e..8fc4c01 100644 --- src/hg/lib/hgFind.c +++ src/hg/lib/hgFind.c @@ -1420,43 +1420,41 @@ if (grepIndexRoot != NULL && hfsSetting != NULL) { char buf[1024]; safef(buf, sizeof(buf), "%s/%s/%s.%s", grepIndexRoot, db, table, suffix); if (fileExists(buf)) return cloneString(buf); } return NULL; } static struct slName *genbankGrepQuery(char *indexFile, char *table, char *key) /* grep -i key indexFile, return a list of ids (first word of each line). */ { -//verbose(1,"genbankGrepQuery table=[%s] key=[%s]\n", table, key); // DEBUG REMOVE char *extraOptions = ""; if (sameString(table, "author")) extraOptions = "-w"; return doGrepQuery(indexFile, table, key, extraOptions); } static struct slName *genbankSqlFuzzyQuery(struct sqlConnection *conn, char *table, char *key) /* Perform a fuzzy sql search for %key% in table.name; return list of * corresponding table.id's. */ { -//verbose(1,"genbankSqlFuzzyQuery table=[%s] key=[%s]\n", table, key); // DEBUG REMOVE struct slName *idList = NULL, *idEl = NULL; if (!isTooCommon(table, key)) { struct sqlResult *sr; char **row; char query[256]; sqlSafef(query, sizeof(query), "select id,name from %s where name like '%%%s%%'", table, key); sr = sqlGetResult(conn, query); while ((row = sqlNextRow(sr)) != NULL) { touppers(row[1]); if (keyIsPrefix(key, row[1])) { idEl = newSlName(row[0]); @@ -1472,59 +1470,57 @@ char *tables[], int tableCount) /* Return TRUE if all tables have a readable genbank index file. */ { int i; for (i=0; i < tableCount; i++) if (! getGenbankGrepIndex(db, hfs, tables[i], "idName")) return FALSE; return TRUE;; } static void findHitsToTables(char *db, struct hgFindSpec *hfs, char *key, char *tables[], int tableCount, struct hash **retHash, struct slName **retList) /* Return all unique accessions that match any table. */ { -//verbose(1,"findHitsToTables db=[%s] key=[%s]\n", db, key); // DEBUG REMOVE struct slName *list = NULL, *el; struct hash *hash = newHash(0); struct sqlConnection *conn = hAllocConn(db); struct sqlResult *sr; char **row; char query[256]; char *field; int i; for (i = 0; i<tableCount; ++i) { struct slName *idList = NULL, *idEl; char *grepIndexFile = NULL; /* I'm doing this query in two steps in C rather than * in one step in SQL just because it somehow is much * faster this way (like 100x faster) when using mySQL. */ field = tables[i]; if (!hTableExists(db, field)) continue; if ((grepIndexFile = getGenbankGrepIndex(db, hfs, field, "idName")) != NULL) idList = genbankGrepQuery(grepIndexFile, field, key); else idList = genbankSqlFuzzyQuery(conn, field, key); for (idEl = idList; idEl != NULL; idEl = idEl->next) { /* don't check srcDb to exclude refseq for compat with older tables */ - //verbose(1,"findHitsToTables field=[%s] idEl->name=[%s]\n", field, idEl->name); // DEBUG REMOVE sqlSafef(query, sizeof(query), "select acc, organism from gbCdnaInfo where %s = %s " " and type = 'mRNA'", field, idEl->name); sr = sqlGetResult(conn, query); while ((row = sqlNextRow(sr)) != NULL) { char *acc = row[0]; /* will use this later to distinguish xeno mrna */ int organismID = sqlUnsigned(row[1]); if (!isRefSeqAcc(acc) && !hashLookup(hash, acc)) { el = newSlName(acc); slAddHead(&list, el); hashAddInt(hash, acc, organismID); @@ -2024,39 +2020,37 @@ accEl->name); sr = sqlGetResult(conn, query); while ((row = sqlNextRow(sr)) != NULL) { struct refLink *rl = refLinkLoad(row); slAddHead(pList, rl); } sqlFreeResult(&sr); } } static boolean findRefGenes(char *db, struct hgFindSpec *hfs, char *spec, struct hgPositions *hgp) /* Look up refSeq genes in table. */ { -//verbose(1,"findRefGenes db=[%s] spec=[%s]\n", db, spec); // DEBUG REMOVE struct sqlConnection *conn = hAllocConn(db); struct dyString *ds = newDyString(256); struct refLink *rlList = NULL, *rl; boolean gotRefLink = hTableExists(db, "refLink"); boolean found = FALSE; char *specNoVersion = cloneString(spec); (void) chopPrefix(specNoVersion); -//verbose(1,"findRefGenes specNoVersion=[%s]\n", specNoVersion); // DEBUG REMOVE if (gotRefLink) { if (startsWith("NM_", specNoVersion) || startsWith("NR_", specNoVersion) || startsWith("XM_", specNoVersion)) { sqlDyStringPrintf(ds, "select * from refLink where mrnaAcc = '%s'", specNoVersion); addRefLinks(conn, ds, &rlList); } else if (startsWith("NP_", specNoVersion) || startsWith("XP_", specNoVersion)) { sqlDyStringPrintf(ds, "select * from refLink where protAcc = '%s'", specNoVersion); addRefLinks(conn, ds, &rlList); } else if (isUnsignedInt(specNoVersion)) { sqlDyStringPrintf(ds, "select * from refLink where locusLinkId = %s", @@ -2581,36 +2575,34 @@ char *table) { if (relativeFlag) hUserAbort("Sorry, range spec (\":%d-%d\") is not supported for %s.", relStart+1, relEnd, table); } #endif static boolean searchSpecial(char *db, struct hgFindSpec *hfs, char *term, struct hgPositions *hgp, boolean relativeFlag, int relStart, int relEnd, boolean *retFound) /* Handle searchTypes for which we have special code. Return true if * we have special code. Set retFind according to whether we find term. */ { -//verbose(1,"searchSpecial db=[%s] term=[%s]\n", db, term); // DEBUG REMOVE boolean isSpecial = TRUE; boolean found = FALSE; char *upcTerm = cloneString(term); touppers(upcTerm); -//verbose(1,"searchSpecial hfs->searchType=[%s]\n", hfs->searchType); // DEBUG REMOVE if (sameString(hfs->searchType, "knownGene")) { if (gotFullText(db)) found = findKnownGeneFullText(db, term, hgp); else /* NOTE, in a few months (say by April 1 2006) get rid of else -JK */ { if (!found && hTableExists(db, "kgAlias")) found = findKgGenesByAlias(db, term, hgp); if (!found && hTableExists(db, "kgProtAlias")) found = findKgGenesByProtAlias(db, term, hgp); if (!found) found = findKnownGene(db, term, hgp, hfs->searchTable); } } else if (sameString(hfs->searchType, "refGene")) @@ -2648,39 +2640,37 @@ } else if (sameString(hfs->searchType, "mrnaKeyword")) { found = findMrnaKeys(db, hfs, upcTerm, hgp); } else if (sameString(hfs->searchType, "sgdGene")) { found = findYeastGenes(db, term, hgp); } else { isSpecial = FALSE; } *retFound = found; freeMem(upcTerm); -//verbose(1,"searchSpecial done isSpecial=%d\n", isSpecial); // DEBUG REMOVE return(isSpecial); } static struct slPair *getXrefTerms(char *db, struct hgFindSpec *hfs, char *term) /* Search xrefTable for xrefQuery with term. Return all matching names. */ { -//verbose(1,"getXrefTerms db=[%s] term=[%s]\n", db, term); // DEBUG REMOVE struct slPair *xrefList = NULL, *xrefPtr = NULL; struct sqlConnection *conn = hAllocConn(db); struct sqlResult *sr = NULL; char **row; boolean isFuzzy = sameWord(hfs->searchMethod, "fuzzy"); // TODO wonder if we could re-work this better to get to upstream sql creation and // then be able to avoid this complexity:? // hfs->refTable sometimes contains a comma-separated table list // but we do not have control over the original sql since it is in trackDb.ra // example from human/hg19/trackDb.ra // xrefTable kgXref, ucscRetroInfo5 // xrefQuery select ucscRetroInfo5.name, spDisplayID from %s where spDisplayID like '%s%%' and kgName = kgID @@ -2727,31 +2717,30 @@ { int sz; va_list args; va_start(args, format); sz = vatruncatef(buf, size, format, args); va_end(args); } static boolean doQuery(char *db, struct hgFindSpec *hfs, char *xrefTerm, char *term, struct hgPositions *hgp, boolean relativeFlag, int relStart, int relEnd, boolean multiTerm) /* Perform a query as specified in hfs, assuming table existence has been * checked and xref'ing has been taken care of. */ { -//verbose(1,"doQuery term=[%s]\n", term); // DEBUG REMOVE struct slName *tableList = hSplitTableNames(db, hfs->searchTable); struct slName *tPtr = NULL; struct hgPosTable *table = NULL; struct hgPos *pos = NULL; struct sqlConnection *conn = hAllocConn(db); struct sqlResult *sr = NULL; char **row = NULL; char *termPrefix = hgFindSpecSetting(hfs, "termPrefix"); char *paddingStr = hgFindSpecSetting(hfs, "padding"); int padding = isEmpty(paddingStr) ? 0 : atoi(paddingStr); boolean found = FALSE; char *description = NULL; char buf[2048]; if (isNotEmpty(termPrefix) && startsWith(termPrefix, term)) @@ -2824,53 +2813,51 @@ } if (table != NULL) slReverse(&table->posList); sqlFreeResult(&sr); hFreeConn(&conn); slFreeList(&tableList); return(found); } boolean hgFindUsingSpec(char *db, struct hgFindSpec *hfs, char *term, struct hgPositions *hgp, boolean relativeFlag, int relStart, int relEnd, boolean multiTerm) /* Perform the search described by hfs on term. If successful, put results * in hgp and return TRUE. (If not, don't modify hgp.) */ { -//verbose(1,"hgFindUsingSpec db=[%s] term=[%s]\n", db, term); // DEBUG REMOVE struct slPair *xrefList = NULL, *xrefPtr = NULL; boolean found = FALSE; if (hfs == NULL || term == NULL || hgp == NULL) errAbort("NULL passed to hgFindUsingSpec.\n"); if (strlen(term)<2 && ! (sameString(hfs->searchName, "knownGene") || sameString(hfs->searchName, "flyBaseGeneSymbolOneLetter"))) return FALSE; if (isNotEmpty(hfs->termRegex) && ! regexMatchNoCase(term, hfs->termRegex)) return(FALSE); if (! hTableOrSplitExists(db, hfs->searchTable)) return(FALSE); if (isNotEmpty(hfs->searchType) && searchSpecial(db, hfs, term, hgp, relativeFlag, relStart, relEnd, &found)) return(found); -//verbose(1,"hgFindUsingSpec hfs->xrefTable=[%s]\n", hfs->xrefTable); // DEBUG REMOVE if (isNotEmpty(hfs->xrefTable)) { struct sqlConnection *conn = hAllocConn(db); // NOTE hfs->xrefTable can sometimes contain a comma-separated table list, // rather than just a single table. char *tables = replaceChars(hfs->xrefTable, ",", " "); boolean exists = sqlTablesExist(conn, tables); hFreeConn(&conn); freeMem(tables); if (! exists) return(FALSE); xrefList = getXrefTerms(db, hfs, term); } else @@ -3095,32 +3082,31 @@ } else relEnd = atoi(stripCommas(term+substrs[3].rm_so)); if (canonicalSpec || gbrowserSpec || lengthSpec) relStart--; if (lengthSpec) relEnd += relStart; if (relStart > relEnd) { int tmp = relStart; relStart = relEnd; relEnd = tmp; } relativeFlag = TRUE; } -//term = sqlEscapeString(term); // term is pre-escaped ! // DEBUG REMOVE -term = cloneString(term); // DEBUG KEEP? +term = cloneString(term); // because hgOfficialChromName mangles it if (hgOfficialChromName(db, term) != NULL) // this mangles the term { char *chrom; int start, end; hgParseChromRange(db, term, &chrom, &start, &end); if (relativeFlag) { int chromSize = end; end = start + relEnd; start = start + relStart; if (end > chromSize) end = chromSize; if (start < 0) @@ -3128,32 +3114,31 @@ } singlePos(hgp, "Chromosome Range", NULL, "chromInfo", originalTerm, "", chrom, start, end); } else { struct hgFindSpec *shortList = NULL, *longList = NULL; struct hgFindSpec *hfs; boolean done = FALSE; // Disable singleBaseSpec for any term that is not hgOfficialChromName // because that mangles legitimate IDs that are [A-Z]:[0-9]+. if (singleBaseSpec) { singleBaseSpec = relativeFlag = FALSE; - //term = sqlEscapeString(originalTerm); // DEBUG REMOVE - term = cloneString(originalTerm); // DEBUG KEEP + term = cloneString(originalTerm); // restore original term relStart = relEnd = 0; } if (!trackHubDatabase(db)) hgFindSpecGetAllSpecs(db, &shortList, &longList); for (hfs = shortList; hfs != NULL; hfs = hfs->next) { if (hgFindUsingSpec(db, hfs, term, hgp, relativeFlag, relStart, relEnd, multiTerm)) { done = TRUE; if (! hgFindSpecSetting(hfs, "semiShortCircuit")) break; } }