/* Need to special-case some URLs. It's yucky to do this,
* but that's the only way this will work for some of the Drosophila
* result types. The assumption in the code is that there is one URL
* per type, but for us it depends on which database the hit is on. */
protected String getURLPrefix(FeatureProperty prop, SeqFeatureI f, String id) {
String defaultURL = prop.getURLString();
if (!(f instanceof FeaturePair)) return (defaultURL);
FeaturePair fp = (FeaturePair) f;
SeqFeatureI fs = (SeqFeatureI) fp.getRefFeature();
String database = fs.getDatabase();
String url = defaultURL;
/* Here we go--a bunch of special cases.
* It sucks that these are in the code--
* these should be in the style file or something! */
String nucleotide = "nucleotide";
String protein = "protein";
String ENTREZ_N =
"http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=search&db="
+ nucleotide
+ "&doptcmdl=GenBank&term=";
String ENTREZ_P =
"http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=search&db="
+ protein
+ "&doptcmdl=GenBank&term=";
// In most cases, the IDs for BLASTX Similarity to Fly seem to work better
// with the default SWALL URL; however, the AA ones seem not to be in SWALL
// and we have to look in Entrez. Yuck.
// if (prop.typeEquals("BLASTX Similarity to Fly") && id.startsWith("AA")) {
if (prop.getDisplayType().startsWith("BLASTX") && id.startsWith("AA")) {
url = ENTREZ_P;
} else if (prop.typeEquals("Rodent") && (database.indexOf("unigene") >= 0)) {
url = ENTREZ_N;
} else if (prop.typeEquals("Insect") && (database.indexOf("dbEST") >= 0)) {
url = ENTREZ_N;
}
// This isn't working right yet. The tRNA ids look like:
// gb|AE002593|AE002593.trna5-ArgTCG
// and we want just the part between the first two ||s.
else {
String acc;
if (prop.typeEquals("tRNA-result") && (database.indexOf("na_tRNA.dros") >= 0)) {
acc = id.substring(id.indexOf("|") + 1);
acc = acc.substring(0, acc.indexOf("|"));
} else if (prop.typeEquals("New Fly Sequence")) {
// These accs look like gi||gb|AY135216|AY135216
acc = id;
if (acc.indexOf("|") >= 0) acc = acc.substring(acc.lastIndexOf("|") + 1);
/* We've already appended the correct acc,
* but the URL constructor will automatically
* slap the uncorrected one at the end.
* The last # is so that when the whole ID is appended at the end, it
* won't have any effect. */
url = ENTREZ_N + acc + "#";
}
}
return url;
}
protected String getIdForURL(SeqFeatureI f) {
/* 02/06/2004: If user selected a whole result (a FeatureSet),
* neither the id nor the name is what we want.
* So just look at the first child of the FeatureSet
* (which is probably a FeaturePair). Is this bad? --NH */
if (f instanceof FeatureSet) {
f = ((FeatureSet) f).getFeatureAt(0);
}
String id;
if (f instanceof FeaturePair) {
FeaturePair fp = (FeaturePair) f;
SeqFeatureI sbjct = (SeqFeatureI) fp.getHitFeature();
id = sbjct.getName();
if (id.equals("") || id.equals("no_name")) id = f.getName();
// System.out.println("getIdForURL: for feature pair " + f.getName() + ", subject name =
// " + id); // DEL
} else {
id = getDisplayName(f);
// System.out.println("getIdForURL: for feature " + f.getName() + ", display name = " +
// id); // DEL
}
// Special case: BDGP EST IDs look like "LD18592.5prime" rather than "LD18592",
// which isn't what we want. Need to leave out the .[53]prime part.
// (Sima says there aren't any cases where we'd want to preserve that part.)
if (id.endsWith(".3prime") || id.endsWith(".5prime")) {
// System.out.println("Truncating EST/cDNA ID " + id + " to remove .[35]prime");
id = id.substring(0, id.lastIndexOf("."));
}
// Some ESTs IDs have :contig1 at the end
if (id.indexOf(":contig") > 0) {
id = id.substring(0, id.indexOf(":contig"));
// In r4.1, I see EST IDs like UNKNOWN_RE01983:contig1 and INVERTED_GH07123:contig1
if (id.indexOf("_") > 0) id = id.substring(id.indexOf("_") + 1);
}
// Some ESTs have _revcomp at the end, e.g. BE975849_revcomp
if (id.indexOf("_revcomp") > 0) id = id.substring(0, id.indexOf("_revcomp"));
return id;
}