Changeset 3389 for trunk

Timestamp:

2002-08-23T10:10:18+12:00 (22 years ago)

Author:

kjdon

Message:

this now sends back real results

File:

• trunk/gsdl3/src/java/org/greenstone/gsdl3/service/PhindService.java (modified) (5 diffs)

trunk/gsdl3/src/java/org/greenstone/gsdl3/service/PhindService.java

@@ -19 +19 @@
 package org.greenstone.gsdl3.service;
 
+import org.greenstone.gsdl3.util.*;
+
+import org.greenstone.mgpp.*;
 import org.w3c.dom.Document; 
 import org.w3c.dom.Node; 
@@ -24 +27 @@
 import org.w3c.dom.Text; 
 
+import java.util.Vector;
+import java.util.HashMap;
 /** 
  * PhindService - the phind phrase browsing service
@@ -32 +37 @@
 public class PhindService 
     extends ServiceModule {
-
     
-    protected Element processService(String name, Element request) {
-
-    if (!name.equals("PhindApplet")) {
-        System.err.println("PhindService:you have asked for a non-existant service - "+name+"!");
-        return null;
-    }
-    // create dummy response
-    Element res = doc_.createElement("response");
-    res.setAttribute("from", "PhindApplet");
-    Element data = doc_.createElement("content");
-    Text t = doc_.createTextNode("this is the results for a phind request");
-    data.appendChild(t);
-    
-    res.appendChild(data);
-
-    return res;
-    
+    private MGPPWrapper mgpp_src_=null;
+    private String basepath_ = null;
+    public PhindService() {
+    mgpp_src_ = new MGPPWrapper();
+    // set up the default params
+    mgpp_src_.setQueryLevel("Document");
+    mgpp_src_.setReturnLevel("Document");
+    mgpp_src_.setMaxDocs(5);
+    mgpp_src_.setStem(false);
+    mgpp_src_.setCase(true);
     }
     /** configure the service module
@@ -73 +70 @@
     f.setAttribute("name", "PhindApplet");
 
-    // add in teh applet info for the phind applet
-    String app_info = "<applet CODEBASE='/gsdl3/lib/java' CODE='org.greenstone.applet.phind.Phind.class' ARCHIVE='phind.jar' WIDTH='500' HEIGHT='400'><PARAM NAME='library'     VALUE='/gsdl3/library'/> <PARAM NAME='phindcgi' VALUE='/gsdl3/library?a=b&amp;sa=phind'/> <PARAM NAME='collection'   VALUE='mgppdemo'/> <PARAM NAME='classifier'   VALUE='1'/>  <PARAM NAME='orientation'  VALUE='vertical'/> <PARAM NAME='depth'        VALUE='2'/> <PARAM NAME='resultorder' VALUE='L,l,E,e,D,d'/> <PARAM NAME='backdrop'     VALUE='/gsdl3/interfaces/default/images/phindbg1.jpg'/><PARAM NAME='fontsize'     VALUE='10'/> <PARAM NAME='blocksize'    VALUE='10'/>The Phind java applet.</applet>";
+    // add in the applet info for the phind applet
+    // need to make this dynamic - library names etc
+    String app_info = "<applet CODEBASE='/gsdl3/lib/java' CODE='org.greenstone.applet.phind.Phind.class' ARCHIVE='phind.jar' WIDTH='500' HEIGHT='400'><PARAM NAME='library' VALUE='/gsdl3/library'/> <PARAM NAME='phindcgi' VALUE='/gsdl3/library?a=a&amp;sa=r&amp;sn=Phind'/> <PARAM NAME='collection'   VALUE='mgppdemo'/> <PARAM NAME='classifier' VALUE='1'/>  <PARAM NAME='orientation'  VALUE='vertical'/> <PARAM NAME='depth' VALUE='2'/> <PARAM NAME='resultorder' VALUE='L,l,E,e,D,d'/> <PARAM NAME='backdrop' VALUE='/gsdl3/interfaces/default/images/phindbg1.jpg'/><PARAM NAME='fontsize' VALUE='10'/> <PARAM NAME='blocksize'    VALUE='10'/>The Phind java applet.</applet>";
 
     Document dom = converter_.getDOM(app_info);
@@ -85 +83 @@
     }
 
+    protected Element processService(String name, Element request) {
+    
+    if (!name.equals("PhindApplet")) {
+        System.err.println("PhindService:you have asked for a non-existant service - "+name+"!");
+        return null;
+    }
+    Element param_elem = (Element)GSXML.getChildByTagName(request, "paramList");
+    HashMap params = GSXML.extractParams(param_elem);
+    
+    long first_e = Long.parseLong((String)params.get("pfe"));
+    long last_e = Long.parseLong((String)params.get("ple"));
+    long first_l = Long.parseLong((String)params.get("pfl"));
+    long last_l = Long.parseLong((String)params.get("pll"));
+    long first_d = Long.parseLong((String)params.get("pfd"));
+    long last_d = Long.parseLong((String)params.get("pld"));
+    
+    long phrase;
+    String phrase_str = (String)params.get("ppnum");
+    if (phrase_str == null || phrase_str.equals("")) {
+        phrase=0;
+    } else {
+        phrase = Long.parseLong(phrase_str);
+    }
+    String word = (String)params.get("pptext");
+    String phind_index = (String)params.get("pc");
+    // the location of the mgpp database files
+    basepath_ = GSFile.phindBasePath(site_home_, collection_name_, phind_index);
+
+    // the result element
+    Element result = doc_.createElement("response");
+    String from = GSPath.appendLink(collection_name_, "PhindApplet");
+    result.setAttribute("from", from);
+    result.setAttribute("type", "query");
+
+    // applet result info must be in appletInfo element
+    Element applet_data = doc_.createElement("appletData");
+    result.appendChild(applet_data);
+    Element phind_data = doc_.createElement("phindData");
+    applet_data.appendChild(phind_data);
+
+
+    // if we dont know the phrase number, look it up
+    if (phrase == 0) {
+        if (word==null || word.equals("")) {
+        Element error = phindError("no word or phrase");
+        phind_data.appendChild(error);
+        return result;
+        }
+        phrase = findPhraseNumberFromWord( word);
+        System.out.println("phind, term number for "+word+" is "+phrase);
+    }
+    if (phrase==0) {
+        // the word is not in the collection
+        // return a phind error string
+        Element error = phindError("the term "+word+" is not in the collection");
+        phind_data.appendChild(error);
+        return result;
+    }
+    
+    // get the phrase data into the phind_data node
+    getPhraseData(phind_data, phrase, first_l, last_l,
+              first_e, last_e,  first_d, last_d);
+    return result;
+    
+    
+    }// processService
+    
+    protected long findPhraseNumberFromWord(String word) {
+
+    // set the mgpp index data - we are looking up pword
+    mgpp_src_.loadIndexData(basepath_, "pword");
+
+    mgpp_src_.runQuery(word);
+
+    MGPPQueryResult res = mgpp_src_.getQueryResult();
+    Vector docs = res.getDocs();
+    if (docs.size()==0) {
+        // phrase not found
+        return 0;
+    }
+    MGPPDocInfo doc = (MGPPDocInfo)docs.firstElement();
+    return doc.num_;
+    }
+
+    protected boolean getPhraseData(Element phind_data,
+                    long phrase, long first_l, long last_l,
+                    long first_e, long last_e, long first_d, 
+                    long last_d) {
+
+    String record = mgpp_src_.getDocument(basepath_, "pdata", "Document",
+                          phrase);
+    if (record.equals("")) {
+        Element error = phindError("somethings gone wrong - we haven't got a record for phrase number "+phrase);
+        phind_data.appendChild(error);
+        return false;
+    }
+    
+    System.out.println("record="+record);
+    // parse the record - its in gordons cryptic form
+    // ":word:tf:ef:df:el:dl:lf:ll"
+    // el: e,e,e
+    // dl: d;f,d;f,
+    // lf and ll may be null
+    // l: type,dest, dest; type,dest,dest
+
+    // ignore everything up to and including first colon (has
+    // <Document>3505: at the start)
+    record = record.substring(record.indexOf(':')+1);
+
+    // split on ':'
+    String [] fields = record.split(":");
+    String word = fields[0];
+    String tf = fields[1];
+    String ef = fields[2];
+    String df = fields[3];
+
+    
+    String expansions = fields[4];
+    String documents = fields[5];
+    String lf = "0";
+    String linklist = "";
+    if (fields.length > 7) {// have thesaurus stuff
+        lf =fields[6];
+        linklist = fields[7];
+    }
+    
+    // the phindData attributes and phrase
+    phind_data.setAttribute("id", Long.toString(phrase));
+    phind_data.setAttribute("df", df);
+    phind_data.setAttribute("ef", ef);
+    phind_data.setAttribute("lf", lf);
+    phind_data.setAttribute("tf", tf);
+    GSXML.createTextElement(doc_, "phrase", word);
+
+    addExpansionList(phind_data, expansions, word, ef, first_e, last_e);
+    addDocumentList(phind_data, documents, word, df, first_d, last_d);
+    if (!lf.equals("0")) {
+        System.out.println("adding thesaurus stuff");
+        addThesaurusList(phind_data, linklist, word, lf, first_l, last_l);
+    }
+    return true;
+    }
+
+    protected boolean addExpansionList( Element phind_data, String record, 
+                       String word,
+                       String freq,
+                       long first, long last) {
+
+    Element expansion_list = doc_.createElement("expansionList");
+    phind_data.appendChild(expansion_list);
+    expansion_list.setAttribute("length", freq);
+    expansion_list.setAttribute("start", Long.toString(first));
+    expansion_list.setAttribute("end", Long.toString(last));
+
+    // get the list of strings
+    String [] expansions = record.split(",");
+    int length = expansions.length;
+    if (length < last) last = length;
+    for (long i = first; i < last; i++) {
+        long num  = Long.parseLong(expansions[(int)i]);
+        Element expansion = getExpansion( num, word);
+        expansion.setAttribute("num", Long.toString(i));
+        expansion_list.appendChild(expansion);
+    }
+    return true;
+    }
+    
+    protected Element getExpansion(long phrase_num, 
+                   String orig_phrase) {
+    
+    // look up the phrase in the pdata thingy
+    String record = mgpp_src_.getDocument(basepath_, "pdata", "Document",
+                          phrase_num);
+
+    if (record ==null || record.equals("")) return null;
+
+    // ignore everything up to and including first colon 
+    record = record.substring(record.indexOf(':')+1);
+
+    String [] fields = record.split(":");
+    String phrase = fields[0];
+    String tf = fields[1];
+    //String ef = fields[2]; dont use this
+    String df = fields[3];
+
+    Element expansion = doc_.createElement("expansion");
+    expansion.setAttribute("tf", tf);
+    expansion.setAttribute("df", df);
+    expansion.setAttribute("id", Long.toString(phrase_num));
+
+    // get teh suffix and prefix
+    String [] ends = splitPhraseOnWord(phrase, orig_phrase);
+    if (!ends[0].equals("")) {
+        expansion.appendChild(GSXML.createTextElement(doc_, "prefix", ends[0]));
+    }
+    if (!ends[1].equals("")) {
+        expansion.appendChild(GSXML.createTextElement(doc_, "suffix", ends[1]));
+    }
+
+    return expansion;
+
+    }
+
+    protected boolean addDocumentList(Element phind_data, String record, 
+                      String word,
+                      String freq,
+                      long first, long last) {
+
+    Element document_list = doc_.createElement("documentList");
+    phind_data.appendChild(document_list);
+    document_list.setAttribute("length", freq);
+    document_list.setAttribute("start", Long.toString(first));
+    document_list.setAttribute("end", Long.toString(last));
+
+    // get the list of doc,freq
+    String [] doc_freqs = record.split(";");
+    int length = doc_freqs.length;
+    if (length<last) last=length;
+
+    for (long i = first; i < last; i++) {
+        String doc_elem = doc_freqs[(int)i];
+        int p = doc_elem.indexOf(',');
+        long doc_num;
+        String doc_freq;
+        if (p == -1) { // there is no freq in the record
+        doc_num =Long.parseLong(doc_elem);
+        doc_freq = "1";
+        } else {
+        doc_num = Long.parseLong(doc_elem.substring(0,p));
+        doc_freq = doc_elem.substring(p+1);
+        }
+        Element document = getDocument( doc_num);
+        document.setAttribute("freq", doc_freq);
+        document.setAttribute("num", Long.toString(i));
+        document_list.appendChild(document);
+    }
+
+    
+    return true;
+    }
+
+
+    protected Element getDocument(long doc_num) {
+    
+    // look up the phrase in the docs thingy
+    String record = mgpp_src_.getDocument(basepath_, "docs", "Document",
+                          doc_num);
+    
+    if (record ==null || record.equals("")) return null;
+    System.out.println("doc record:"+record);
+    
+    // ignore everything up to and including first \t
+    record = record.substring(record.indexOf('\t')+1);
+
+    String [] fields = record.split("\t");
+    String hash = fields[0];
+    String title = fields[1];
+
+    Element d = doc_.createElement("document");
+    d.setAttribute("hash", hash);
+    d.appendChild(GSXML.createTextElement(doc_, "title", title));
+    
+    return d;
+
+    }
+    protected boolean addThesaurusList(Element phind_data, String record,
+                       String word,
+                       String freq,
+                       long first, long last) {
+
+
+    Element thesaurus_list = doc_.createElement("thesaurusList");
+    phind_data.appendChild(thesaurus_list);
+    thesaurus_list.setAttribute("length", freq);
+    thesaurus_list.setAttribute("start", Long.toString(first));
+    thesaurus_list.setAttribute("end", Long.toString(last));
+    
+    System.out.println("record for thesaurus="+record);
+
+    // get the list of type,dest,dest
+    String [] links = record.split(";");
+    int length = links.length;
+    long index = 0;
+    for (int i = 0; i < length; i++) { // go through the entries
+        String link_info = links[(int)i];
+        String [] items = link_info.split(",");
+        // the first entry is teh type
+        String type = items[0];
+        for (int j = 1; j<items.length; j++, index++) {
+        if (index >= first && index < last) { // only output the ones we want
+            long phrase = Long.parseLong(items[j]);
+            Element t = getThesaurus(phrase);
+            t.setAttribute("type", type);
+            thesaurus_list.appendChild(t);
+        }
+        }
+    }
+
+    return true;
+    }
+
+    protected Element getThesaurus(long phrase_num) {
+
+    // look up the phrase in the pdata thingy
+    String record = mgpp_src_.getDocument(basepath_, "pdata", "Document",
+                          phrase_num);
+
+    if (record ==null || record.equals("")) return null;
+
+    // ignore everything up to and including first colon 
+    record = record.substring(record.indexOf(':')+1);
+
+    String [] fields = record.split(":");
+    String phrase = fields[0];
+    String tf = fields[1];
+    //String ef = fields[2]; dont use this
+    String df = fields[3];
+
+    Element thesaurus = doc_.createElement("thesaurus");
+    thesaurus.setAttribute("tf", tf);
+    thesaurus.setAttribute("df", df);
+    thesaurus.setAttribute("id", Long.toString(phrase_num));
+    thesaurus.appendChild(GSXML.createTextElement(doc_, "phrase", phrase));
+    return thesaurus;
+
+    }
+
+    /** returns an array of two elements - the prefix and the suffix*/
+    protected String [] splitPhraseOnWord(String phrase, String word) {
+    
+    if (word.equals("")) {
+        
+        String [] res =  {phrase, ""};
+        return res;
+    }
+    // use 2 so that we only split on the first occurrance. trailing empty strings should be included 
+    String [] result = phrase.split(word, 2);
+    if (result.length !=2) {
+        System.out.println("didn't get two substrings!!");
+    }
+    return result;
+    
+    }
+
+    protected Element phindError(String message) {
+    Element e = doc_.createElement("phindError");
+    Text t = doc_.createTextNode(message);
+    e.appendChild(t);
+    return e;
+    }
     
 }
+
+
+    /*
+    // CREATE dummy response
+    Element res = doc_.createElement("response");
+    res.setAttribute("from", "PhindApplet");
+    Element data = doc_.createElement("service");
+    Element app_data = doc_.createElement("appletData");
+    data.appendChild(app_data);
+    String phind_info ="<phindData id='2507' tf='19424' ef='1632' df='1843' lf='0'><phrase>FOREST</phrase><expansionList length='1632' start='0' end='10'><expansion num='0' id='177648' tf='2162' df='519'><suffix>MANAGEMENT</suffix></expansion> <expansion num='1' id='177531' tf='1958' df='566'><suffix>PRODUCTS</suffix></expansion> <expansion num='2' id='177469' tf='1328' df='532'><suffix>RESOURCES</suffix></expansion> <expansion num='3' id='177773' tf='943' df='177'><suffix>GENETIC</suffix></expansion> <expansion num='4' id='177335' tf='736' df='258'><prefix>SUSTAINABLE</prefix></expansion> </expansionList><documentList length='1843' start='0' end='10'><document num='0' hash='HASH011fb8a7d8bf781ab3cbb087' freq='363'><title>FO-edu List of Countries 0</title></document><document num='1' hash='HASH27ae41229eb0636849a5be' freq='344' ><title>FO-edu List of Countries 1</title></document><document num='2' hash='HASH0187ef85c9dbf5bf132ea1d1' freq='263'><title>FO-edu List of Countries 2</title></document><document num='3' hash='HASH0125ec9ef67960446f471280' freq='238'><title>FO-edu List of Countries 3</title></document><document num='4' hash='HASH67087f7717eb35050ce1ac' freq='213'><title>FO-edu List of Countries 4</title></document></documentList><thesaurusList><thesaurus num='3' id='36506' tf='0' df='0' type='RT'><phrase>FRANCOPHONE</phrase></thesaurus></thesaurusList></phindData>";
+    
+    Node t = converter_.getDOM(phind_info).getDocumentElement();
+    app_data.appendChild(doc_.importNode(t, true));
+    
+    res.appendChild(data);
+
+    return res;
+    */