2 * Jalview - A Sequence Alignment Editor and Viewer ($$Version-Rel$$)
3 * Copyright (C) $$Year-Rel$$ The Jalview Authors
5 * This file is part of Jalview.
7 * Jalview is free software: you can redistribute it and/or
8 * modify it under the terms of the GNU General Public License
9 * as published by the Free Software Foundation, either version 3
10 * of the License, or (at your option) any later version.
12 * Jalview is distributed in the hope that it will be useful, but
13 * WITHOUT ANY WARRANTY; without even the implied warranty
14 * of MERCHANTABILITY or FITNESS FOR A PARTICULAR
15 * PURPOSE. See the GNU General Public License for more details.
17 * You should have received a copy of the GNU General Public License
18 * along with Jalview. If not, see <http://www.gnu.org/licenses/>.
19 * The Jalview Authors are detailed in the 'AUTHORS' file.
23 import jalview.datamodel.Sequence;
24 import jalview.datamodel.SequenceI;
25 import jalview.jsdev.RegExp;
26 import jalview.jsdev.api.RegExpInterface;
28 import java.util.Hashtable;
30 //import com.stevesoft.pat.Regex;
32 public class GroupUrlLink
34 public class UrlStringTooLongException extends Exception
36 public UrlStringTooLongException(int lng)
43 public String toString()
45 return "Generated url is estimated to be too long (" + urlLength
51 * Helper class based on the UrlLink class which enables URLs to be
52 * constructed from sequences or IDs associated with a group of sequences. URL
53 * definitions consist of a pipe separated string containing a <label>|<url
54 * construct>|<separator character>[|<sequence separator character>]. The url
55 * construct includes regex qualified tokens which are replaced with seuqence
56 * IDs ($SEQUENCE_IDS$) and/or seuqence regions ($SEQUENCES$) that are
57 * extracted from the group. See <code>UrlLink</code> for more information
58 * about the approach, and the original implementation. Documentation to come.
59 * Note - groupUrls can be very big!
61 private String url_prefix, target, label;
64 * these are all filled in order of the occurence of each token in the url
67 private String url_suffix[], separators[], regexReplace[];
69 private String invalidMessage = null;
72 * tokens that can be replaced in the URL.
74 private static String[] tokens;
77 * position of each token (which can appear once only) in the url
82 * contains tokens in the order they appear in the URL template.
84 private String[] mtch;
90 { "SEQUENCEIDS", "SEQUENCES", "DATASETID" };
95 * test for GroupURLType bitfield (with default tokens)
97 public static final int SEQUENCEIDS = 1;
100 * test for GroupURLType bitfield (with default tokens)
102 public static final int SEQUENCES = 2;
105 * test for GroupURLType bitfield (with default tokens)
107 public static final int DATASETID = 4;
109 // private int idseg = -1, seqseg = -1;
112 * parse the given linkString of the form '<label>|<url>|separator
113 * char[|optional sequence separator char]' into parts. url may contain a
114 * string $SEQUENCEIDS<=optional regex=>$ where <=optional regex=> must be of
115 * the form =/<perl style regex>/=$ or $SEQUENCES<=optional regex=>$ or
116 * $SEQUENCES<=optional regex=>$.
120 public GroupUrlLink(String link)
122 int sep = link.indexOf("|");
123 segs = new int[tokens.length];
125 for (int i = 0; i < segs.length; i++)
127 if ((segs[i] = link.indexOf("$" + tokens[i])) > -1)
132 // expect at least one token
135 invalidMessage = "Group URL string must contain at least one of ";
136 for (int i = 0; i < segs.length; i++)
138 invalidMessage += " '$" + tokens[i] + "[=/regex=/]$'";
143 int[] ptok = new int[ntoks + 1];
144 String[] tmtch = new String[ntoks + 1];
145 mtch = new String[ntoks];
146 for (int i = 0, t = 0; i < segs.length; i++)
151 tmtch[t++] = tokens[i];
154 ptok[ntoks] = link.length();
155 tmtch[ntoks] = "$$$$$$$$$";
156 QuickSort.sortInt(ptok, tmtch);
157 for (int i = 0; i < ntoks; i++)
159 mtch[i] = tmtch[i]; // TODO: check order is ascending
162 * replaces the specific code below {}; if (psqids > -1 && pseqs > -1) { if
163 * (psqids > pseqs) { idseg = 1; seqseg = 0;
165 * ptok = new int[] { pseqs, psqids, link.length() }; mtch = new String[] {
166 * "$SEQUENCES", "$SEQUENCEIDS" }; } else { idseg = 0; seqseg = 1; ptok =
167 * new int[] { psqids, pseqs, link.length() }; mtch = new String[] {
168 * "$SEQUENCEIDS", "$SEQUENCES" }; } } else { if (psqids != -1) { idseg = 0;
169 * ptok = new int[] { psqids, link.length() }; mtch = new String[] {
170 * "$SEQUENCEIDS" }; } else { seqseg = 0; ptok = new int[] { pseqs,
171 * link.length() }; mtch = new String[] { "$SEQUENCES" }; } }
175 // first get the label and target part before the first |
179 p = link.indexOf("|", sep + 1);
180 } while (p > sep && p < ptok[0]);
181 // Assuming that the URL itself does not contain any '|' symbols
182 // sep now contains last pipe symbol position prior to any regex symbols
183 label = link.substring(0, sep);
184 if (label.indexOf("|") > -1)
186 // | terminated database name / www target at start of Label
187 target = label.substring(0, label.indexOf("|"));
189 else if (label.indexOf(" ") > 2)
191 // space separated Label - matches database name
192 target = label.substring(0, label.indexOf(" "));
198 // Now Parse URL : Whole URL string first
199 url_prefix = link.substring(sep + 1, ptok[0]);
200 url_suffix = new String[mtch.length];
201 regexReplace = new String[mtch.length];
202 // and loop through tokens
203 for (int pass = 0; pass < mtch.length; pass++)
205 int mlength = 3 + mtch[pass].length();
206 if (link.indexOf("$" + mtch[pass] + "=/") == ptok[pass]
207 && (p = link.indexOf("/=$", ptok[pass] + mlength)) > ptok[pass]
210 // Extract Regex and suffix
211 if (ptok[pass + 1] < p + 3)
213 // tokens are not allowed inside other tokens - e.g. inserting a
214 // $sequences$ into the regex match for the sequenceid
215 invalidMessage = "Token regexes cannot contain other regexes (did you terminate the $"
216 + mtch[pass] + " regex with a '/=$' ?";
219 url_suffix[pass] = link.substring(p + 3, ptok[pass + 1]);
220 regexReplace[pass] = link.substring(ptok[pass] + mlength, p);
223 RegExpInterface rg = RegExp.perlCode("/"
224 + regexReplace[pass] + "/");
227 invalidMessage = "Invalid Regular Expression : '"
228 + regexReplace[pass] + "'\n";
230 } catch (Exception e)
232 invalidMessage = "Invalid Regular Expression : '"
233 + regexReplace[pass] + "'\n";
238 regexReplace[pass] = null;
239 // verify format is really correct.
240 if ((p = link.indexOf("$" + mtch[pass] + "$")) == ptok[pass])
242 url_suffix[pass] = link.substring(p + mtch[pass].length() + 2,
247 invalidMessage = "Warning: invalid regex structure (after '"
248 + mtch[0] + "') for URL link : " + link;
253 separators = new String[url_suffix.length];
254 String suffices = url_suffix[url_suffix.length - 1], lastsep = ",";
255 // have a look in the last suffix for any more separators.
256 while ((p = suffices.indexOf('|')) > -1)
258 separators[pass] = suffices.substring(p + 1);
261 // trim the original suffix string
262 url_suffix[url_suffix.length - 1] = suffices.substring(0, p);
266 lastsep = (separators[pass - 1] = separators[pass - 1].substring(0,
269 suffices = separators[pass];
274 lastsep = separators[pass - 1];
276 // last separator is always used for all the remaining separators
277 while (pass < separators.length)
279 separators[pass++] = lastsep;
284 * @return the url_suffix
286 public String getUrl_suffix()
288 return url_suffix[url_suffix.length - 1];
292 * @return the url_prefix
294 public String getUrl_prefix()
302 public String getTarget()
310 public String getLabel()
316 * @return the sequence ID regexReplace
318 public String getIDRegexReplace()
320 return _replaceFor(tokens[0]);
323 private String _replaceFor(String token)
325 for (int i = 0; i < mtch.length; i++)
327 if (segs[i] > -1 && mtch[i].equals(token))
329 return regexReplace[i];
336 * @return the sequence ID regexReplace
338 public String getSeqRegexReplace()
340 return _replaceFor(tokens[1]);
344 * @return the invalidMessage
346 public String getInvalidMessage()
348 return invalidMessage;
352 * Check if URL string was parsed properly.
354 * @return boolean - if false then <code>getInvalidMessage</code> returns an
357 public boolean isValid()
359 return invalidMessage == null;
363 * return one or more URL strings by applying regex to the given idstring
366 * array of id strings to pass to service
368 * array of seq strings to pass to service
369 * @param onlyIfMatches
370 * - when true url strings are only made if regex is defined and
371 * matches for all qualified tokens in groupURL - TODO: consider if
372 * onlyIfMatches is really a useful parameter!
373 * @return null or Object[] { int[] { number of seqs substituted},boolean[] {
374 * which seqs were substituted }, StringBuffer[] { substituted lists
375 * for each token }, String[] { url } }
376 * @throws UrlStringTooLongException
378 public Object[] makeUrls(String[] idstrings, String[] seqstrings,
379 String dsstring, boolean onlyIfMatches)
380 throws UrlStringTooLongException
382 Hashtable rstrings = replacementArgs(idstrings, seqstrings, dsstring);
383 return makeUrls(rstrings, onlyIfMatches);
387 * gathers input into a hashtable
394 private Hashtable replacementArgs(String[] idstrings,
395 String[] seqstrings, String dsstring)
397 Hashtable rstrings = new Hashtable();
398 rstrings.put(tokens[0], idstrings);
399 rstrings.put(tokens[1], seqstrings);
400 rstrings.put(tokens[2], new String[]
402 if (idstrings.length != seqstrings.length)
404 throw new Error(MessageManager.getString("error.idstring_seqstrings_only_one_per_sequence"));
409 public Object[] makeUrls(Hashtable repstrings, boolean onlyIfMatches)
410 throws UrlStringTooLongException
412 return makeUrlsIf(true, repstrings, onlyIfMatches);
421 * @return URL stub objects ready to pass to constructFrom
422 * @throws UrlStringTooLongException
424 public Object[] makeUrlStubs(String[] ids, String[] seqstr,
425 String string, boolean b) throws UrlStringTooLongException
427 Hashtable rstrings = replacementArgs(ids, seqstr, string);
428 Object[] stubs = makeUrlsIf(false, rstrings, b);
432 { stubs[0], stubs[1], rstrings, new boolean[]
435 // TODO Auto-generated method stub
440 * generate the URL for the given URL stub object array returned from
444 * @return URL string.
445 * @throws UrlStringTooLongException
447 public String constructFrom(Object[] stubs)
448 throws UrlStringTooLongException
450 Object[] results = makeUrlsIf(true, (Hashtable) stubs[2],
451 ((boolean[]) stubs[3])[0]);
452 return ((String[]) results[3])[0];
456 * conditionally generate urls or stubs for a given input.
458 * @param createFullUrl
459 * set to false if you only want to test if URLs would be generated.
461 * @param onlyIfMatches
462 * @return null if no url is generated. Object[] { int[] { number of matches
463 * seqs }, boolean[] { which matched }, (if createFullUrl also has
464 * StringBuffer[] { segment generated from inputs that is used in URL
465 * }, String[] { url })}
466 * @throws UrlStringTooLongException
468 protected Object[] makeUrlsIf(boolean createFullUrl,
469 Hashtable repstrings, boolean onlyIfMatches)
470 throws UrlStringTooLongException
474 // prepare string arrays in correct order to be assembled into URL input
475 String[][] idseq = new String[mtch.length][]; // indexed by pass
476 int mins = 0, maxs = 0; // allowed two values, 1 or n-sequences.
477 for (int i = 0; i < mtch.length; i++)
479 idseq[i] = (String[]) repstrings.get(mtch[i]);
480 if (idseq[i].length >= 1)
482 if (mins == 0 && idseq[i].length == 1)
488 maxs = idseq[i].length;
492 if (maxs != idseq[i].length)
494 throw new Error(MessageManager.formatMessage("error.cannot_have_mixed_length_replacement_vectors",
495 new String[]{(mtch[i]), Integer.valueOf(idseq[i].length).toString(),Integer.valueOf(maxs).toString()}));
501 throw new Error(MessageManager.getString("error.cannot_have_zero_length_vector_replacement_strings"));
504 // iterate through input, collating segments to be inserted into url
505 StringBuffer matched[] = new StringBuffer[idseq.length];
506 // and precompile regexes
507 RegExpInterface[] rgxs = new RegExpInterface[matched.length];
508 for (pass = 0; pass < matched.length; pass++)
510 matched[pass] = new StringBuffer();
511 if (regexReplace[pass] != null)
513 rgxs[pass] = RegExp.perlCode("/"
514 + regexReplace[pass] + "/");
521 // tot up the invariant lengths for this url
522 int urllength = url_prefix.length();
523 for (pass = 0; pass < matched.length; pass++)
525 urllength += url_suffix[pass].length();
528 // flags to record which of the input sequences were actually used to
531 boolean[] thismatched = new boolean[maxs];
533 for (int sq = 0; sq < maxs; sq++)
535 // initialise flag for match
536 thismatched[sq] = false;
537 StringBuffer[] thematches = new StringBuffer[rgxs.length];
538 for (pass = 0; pass < rgxs.length; pass++)
540 thematches[pass] = new StringBuffer(); // initialise - in case there are
543 // if a regex is provided, then it must match for all sequences in all
544 // tokens for it to be considered.
545 if (idseq[pass].length <= sq)
547 // no more replacement strings to try for this token
550 if (rgxs[pass] != null)
552 RegExpInterface rg = rgxs[pass];
554 // concatenate all matches of re in the given string!
555 while (rg.searchFrom(idseq[pass][sq], rematchat))
557 rematchat = rg.matchedTo();
558 thismatched[sq] |= true;
559 urllength += rg.charsMatched(); // count length
560 if ((urllength + 32) > Platform.getMaxCommandLineLength())
562 throw new UrlStringTooLongException(urllength);
567 continue; // don't bother making the URL replacement text.
569 // do we take the cartesian products of the substituents ?
570 int ns = rg.numSubs();
573 thematches[pass].append(rg.stringMatched());// take whole regex
576 * else if (ns==1) { // take only subgroup match return new String[]
577 * { rg.stringMatched(1), url_prefix+rg.stringMatched(1)+url_suffix
580 // deal with multiple submatch case - for moment we do the simplest
581 // - concatenate the matched regions, instead of creating a complete
582 // list for each alternate match over all sequences.
583 // TODO: specify a 'replace pattern' - next refinement
588 * for (int s = 0; s <= rg.numSubs(); s++) {
589 * System.err.println("Sub " + s + " : " + rg.matchedFrom(s) +
590 * " : " + rg.matchedTo(s) + " : '" + rg.stringMatched(s) + "'");
593 // try to collate subgroup matches
594 StringBuffer subs = new StringBuffer();
595 // have to loop through submatches, collating them at top level
600 if (s + 1 <= ns && rg.matchedToI(s) > -1
601 && rg.matchedToI(s + 1) > -1
602 && rg.matchedToI(s + 1) < rg.matchedToI(s))
604 // s is top level submatch. search for submatches enclosed by
607 StringBuffer rmtch = new StringBuffer();
608 while (r <= ns && rg.matchedToI(r) <= rg.matchedToI(s))
610 if (rg.matchedFromI(r) > -1)
612 rmtch.append(rg.stringMatchedI(r));
616 if (rmtch.length() > 0)
618 subs.append(rmtch); // simply concatenate
624 if (rg.matchedFromI(s) > -1)
626 subs.append(rg.stringMatchedI(s)); // concatenate
631 thematches[pass].append(subs);
637 // are we only supposed to take regex matches ?
640 thismatched[sq] |= true;
641 urllength += idseq[pass][sq].length(); // tot up length
644 thematches[pass] = new StringBuffer(idseq[pass][sq]); // take
647 // regardless - probably not a
650 * TODO: do some boilerplate trimming of the fields to make them
651 * sensible e.g. trim off any 'prefix' in the id string (see
652 * UrlLink for the below) - pre 2.4 Jalview behaviour if
653 * (idstring.indexOf("|") > -1) { idstring =
654 * idstring.substring(idstring.lastIndexOf("|") + 1); }
662 // check if we are going to add this sequence's results ? all token
663 // replacements must be valid for this to happen!
664 // (including single value replacements - eg. dataset name)
669 for (pass = 0; pass < matched.length; pass++)
671 if (idseq[pass].length > 1 && matched[pass].length() > 0)
673 matched[pass].append(separators[pass]);
675 matched[pass].append(thematches[pass]);
681 // finally, if any sequences matched, then form the URL and return
682 if (seqsmatched == 0 || (createFullUrl && matched[0].length() == 0))
684 // no matches - no url generated
687 // check if we are beyond the feasible command line string limit for this
689 if ((urllength + 32) > Platform.getMaxCommandLineLength())
691 throw new UrlStringTooLongException(urllength);
695 // just return the essential info about what the URL would be generated
699 { seqsmatched }, thismatched };
701 // otherwise, create the URL completely.
703 StringBuffer submiturl = new StringBuffer();
704 submiturl.append(url_prefix);
705 for (pass = 0; pass < matched.length; pass++)
707 submiturl.append(matched[pass]);
708 if (url_suffix[pass] != null)
710 submiturl.append(url_suffix[pass]);
716 { seqsmatched }, thismatched, matched, new String[]
717 { submiturl.toString() } };
723 * @return number of distinct sequence (id or seuqence) replacements predicted
726 public int getNumberInvolved(Object[] urlstub)
728 return ((int[]) urlstub[0])[0]; // returns seqsmatched from
729 // makeUrlsIf(false,...)
733 * get token types present in this url as a bitfield indicating presence of
734 * each token from tokens (LSB->MSB).
736 * @return groupURL class as integer
738 public int getGroupURLType()
741 for (int pass = 0; pass < tokens.length; pass++)
743 for (int i = 0; i < mtch.length; i++)
745 if (mtch[i].equals(tokens[pass]))
754 public String toString()
756 StringBuffer result = new StringBuffer();
757 result.append(label + "|" + url_prefix);
759 for (r = 0; r < url_suffix.length; r++)
762 result.append(mtch[r]);
763 if (regexReplace[r] != null)
766 result.append(regexReplace[r]);
770 result.append(url_suffix[r]);
772 for (r = 0; r < separators.length; r++)
775 result.append(separators[r]);
777 return result.toString();
781 * report stats about the generated url string given an input set
787 private static void testUrls(GroupUrlLink ul, String[][] idstring,
793 System.out.println("Created NO urls.");
797 System.out.println("Created a url from " + ((int[]) url[0])[0]
798 + "out of " + idstring[0].length + " sequences.");
799 System.out.println("Sequences that did not match:");
800 for (int sq = 0; sq < idstring[0].length; sq++)
802 if (!((boolean[]) url[1])[sq])
804 System.out.println("Seq " + sq + ": " + idstring[0][sq] + "\t: "
808 System.out.println("Sequences that DID match:");
809 for (int sq = 0; sq < idstring[0].length; sq++)
811 if (((boolean[]) url[1])[sq])
813 System.out.println("Seq " + sq + ": " + idstring[0][sq] + "\t: "
817 System.out.println("The generated URL:");
818 System.out.println(((String[]) url[3])[0]);
827 public static void main(String argv[])
829 // note - JAL-1383 - these services are all dead
830 String[] links = new String[]
832 "EnVision2|IDS|http://www.ebi.ac.uk/enfin-srv/envision2/pages/linkin.jsf?workflow=Enfin%20Default%20Workflow&datasetName=linkInDatasetFromJalview&input=$SEQUENCEIDS$&inputType=0|,",
833 "EnVision2|Seqs|http://www.ebi.ac.uk/enfin-srv/envision2/pages/linkin.jsf?workflow=Enfin%20Default%20Workflow&datasetName=linkInDatasetFromJalview&input=$SEQUENCES$&inputType=1|,",
834 "EnVision2|IDS|http://www.ebi.ac.uk/enfin-srv/envision2/pages/linkin.jsf?workflow=Enfin%20Default%20Workflow&datasetName=$DATASETID$&input=$SEQUENCEIDS$&inputType=0|,",
835 "EnVision2|Seqs|http://www.ebi.ac.uk/enfin-srv/envision2/pages/linkin.jsf?workflow=Enfin%20Default%20Workflow&datasetName=$DATASETID$&input=$SEQUENCES$&inputType=1|,",
836 "EnVision2|IDS|http://www.ebi.ac.uk/enfin-srv/envision2/pages/linkin.jsf?workflow=$SEQUENCEIDS$&datasetName=linkInDatasetFromJalview&input=$SEQUENCEIDS$&inputType=0|,",
837 "EnVision2|Seqs|http://www.ebi.ac.uk/enfin-srv/envision2/pages/linkin.jsf?workflow=$SEQUENCEIDS$&datasetName=$DATASETID$&input=$SEQUENCES$&inputType=1|,",
838 "EnVision2 Seqs|http://www.ebi.ac.uk/enfin-srv/envision2/pages/linkin.jsf?workflow=Default&datasetName=JalviewSeqs$DATASETID$&input=$SEQUENCES=/([a-zA-Z]+)/=$&inputType=1|,",
839 "EnVision2 Seqs|http://www.ebi.ac.uk/enfin-srv/envision2/pages/linkin.jsf?workflow=Default&datasetName=JalviewSeqs$DATASETID$&input=$SEQUENCES=/[A-Za-z]+/=$&inputType=1|,"
841 * http://www.ebi.ac.uk/enfin-srv/envision2/pages/linkin.jsf?input=P38389,P38398
842 * &inputType=0&workflow=Enfin%20Default%20Workflow&datasetName=
843 * linkInDatasetFromPRIDE
847 SequenceI[] seqs = new SequenceI[]
848 { new Sequence("StupidLabel:gi|9234|pdb|102L|A",
849 "asdiasdpasdpadpwpadasdpaspdw"), };
850 String[][] seqsandids = formStrings(seqs);
851 for (int i = 0; i < links.length; i++)
853 GroupUrlLink ul = new GroupUrlLink(links[i]);
856 System.out.println("\n\n\n");
857 System.out.println("Link " + i + " " + links[i] + " : "
859 System.out.println(" pref : " + ul.getUrl_prefix());
860 System.out.println(" IdReplace : " + ul.getIDRegexReplace());
861 System.out.println(" SeqReplace : " + ul.getSeqRegexReplace());
862 System.out.println(" Suffixes : " + ul.getUrl_suffix());
865 .println("<insert input id and sequence strings here> Without onlyIfMatches:");
869 urls = ul.makeUrls(seqsandids[0], seqsandids[1], "mydataset",
871 testUrls(ul, seqsandids, urls);
872 } catch (UrlStringTooLongException ex)
874 System.out.println("too long exception " + ex);
877 .println("<insert input id and sequence strings here> With onlyIfMatches set:");
880 urls = ul.makeUrls(seqsandids[0], seqsandids[1], "mydataset",
882 testUrls(ul, seqsandids, urls);
883 } catch (UrlStringTooLongException ex)
885 System.out.println("too long exception " + ex);
890 System.err.println("Invalid URLLink : " + links[i] + " : "
891 + ul.getInvalidMessage());
897 * covenience method to generate the id and sequence string vector from a set
898 * of seuqences using each sequence's getName() and getSequenceAsString()
902 * @return String[][] {{sequence ids},{sequence strings}}
904 public static String[][] formStrings(SequenceI[] seqs)
906 String[][] idset = new String[2][seqs.length];
907 for (int i = 0; i < seqs.length; i++)
909 idset[0][i] = seqs[i].getName();
910 idset[1][i] = seqs[i].getSequenceAsString();
915 public void setLabel(String newlabel)
917 this.label = newlabel;