From 107e02b74a465f303021fdc5596333a29bc883db Mon Sep 17 00:00:00 2001
From: Jean-Francois Dockes <jfd@recoll.org>
Date: Tue, 21 Dec 2010 16:00:25 +0100
Subject: [PATCH] Gui search: make autophrase work with a query language query

---
 src/qtgui/ssearch_w.cpp  |  3 ++
 src/rcldb/searchdata.cpp | 68 ++++++++++++++++++++++++++++++++++++++--
 src/rcldb/searchdata.h   | 14 ++++++++-
 3 files changed, 82 insertions(+), 3 deletions(-)

diff --git a/src/qtgui/ssearch_w.cpp b/src/qtgui/ssearch_w.cpp
index f98dfa06..5fe8aaf5 100644
--- a/src/qtgui/ssearch_w.cpp
+++ b/src/qtgui/ssearch_w.cpp
@@ -139,6 +139,9 @@ void SSearch::startSimpleSearch()
 				 QString::fromAscii(reason.c_str()));
 	    return;
 	}
+	if (prefs.ssearchAutoPhrase) {
+	    sdata->maybeAddAutoPhrase();
+	}
     } else {
 	sdata = new Rcl::SearchData(Rcl::SCLT_OR);
 	if (sdata == 0) {
diff --git a/src/rcldb/searchdata.cpp b/src/rcldb/searchdata.cpp
index 0d8f3828..4c58a5b1 100644
--- a/src/rcldb/searchdata.cpp
+++ b/src/rcldb/searchdata.cpp
@@ -242,7 +242,7 @@ bool SearchData::toNativeQuery(Rcl::Db &db, void *d)
 	for (vector<string>::iterator it = exptps.begin(); 
 	     it != exptps.end(); it++) {
 	    string term = "T" + *it;
-	    LOGDEB(("Adding file type term: [%s]\n", term.c_str()));
+	    LOGDEB0(("Adding file type term: [%s]\n", term.c_str()));
 	    tq = tq.empty() ? Xapian::Query(term) : 
 		Xapian::Query(Xapian::Query::OP_OR, tq, Xapian::Query(term));
 	}
@@ -268,6 +268,70 @@ bool SearchData::toNativeQuery(Rcl::Db &db, void *d)
     return true;
 }
 
+
+bool SearchData::maybeAddAutoPhrase()
+{
+    LOGDEB0(("SearchData::maybeAddAutoPhrase()\n"));
+    if (!m_query.size()) {
+	LOGDEB2(("SearchData::maybeAddAutoPhrase: empty query\n"));
+	return false;
+    }
+
+    string field;
+    string words;
+    // Walk the clause list. If we find any non simple clause or different
+    // field names, bail out.
+    for (qlist_it_t it = m_query.begin(); it != m_query.end(); it++) {
+	SClType tp = (*it)->m_tp;
+	if (tp != SCLT_AND && tp != SCLT_OR) {
+	    LOGDEB2(("SearchData::maybeAddAutoPhrase: complex clause\n"));
+	    return false;
+	}
+	SearchDataClauseSimple *clp = 
+	    dynamic_cast<SearchDataClauseSimple*>(*it);
+	if (clp == 0) {
+	    LOGDEB2(("SearchData::maybeAddAutoPhrase: dyncast failed\n"));
+	    return false;
+	}
+	if (it == m_query.begin()) {
+	    field = clp->getfield();
+	} else {
+	    if (clp->getfield().compare(field)) {
+		LOGDEB2(("SearchData::maybeAddAutoPhrase: different fields\n"));
+		return false;
+	    }
+	}
+	if (!words.empty())
+	    words += " ";
+	words +=  clp->gettext();
+    }
+
+    if (words.find_first_of("\"*[]?") != string::npos && // has wildcards
+	TextSplit::countWords(words) <= 1) { // Just one word.
+	LOGDEB2(("SearchData::maybeAddAutoPhrase: wildcards or single word\n"));
+	return false;
+    }
+
+    SearchDataClauseDist *nclp = 
+	new SearchDataClauseDist(SCLT_PHRASE, words, 0, field);
+    if (m_tp == SCLT_OR) {
+	addClause(nclp);
+    } else {
+	// My type is AND. Change it to OR and insert two queries, one
+	// being the original query as a subquery, the other the
+	// phrase.
+	SearchData *sd = new SearchData(m_tp);
+	sd->m_query = m_query;
+	m_tp = SCLT_OR;
+	m_query.clear();
+	SearchDataClauseSub *oq = 
+	    new SearchDataClauseSub(SCLT_OR, RefCntr<SearchData>(sd));
+	addClause(oq);
+	addClause(nclp);
+    }
+    return true;
+}
+
 // Add clause to current list. OR lists cant have EXCL clauses.
 bool SearchData::addClause(SearchDataClause* cl)
 {
@@ -284,7 +348,7 @@ bool SearchData::addClause(SearchDataClause* cl)
 
 // Make me all new
 void SearchData::erase() {
-    LOGDEB(("SearchData::erase\n"));
+    LOGDEB0(("SearchData::erase\n"));
     m_tp = SCLT_AND;
     for (qlist_it_t it = m_query.begin(); it != m_query.end(); it++)
 	delete *it;
diff --git a/src/rcldb/searchdata.h b/src/rcldb/searchdata.h
index 9c0f9502..29a1f0cd 100644
--- a/src/rcldb/searchdata.h
+++ b/src/rcldb/searchdata.h
@@ -73,7 +73,10 @@ class SearchData {
 public:
     SearchData(SClType tp) 
         : m_tp(tp), m_haveDates(false), m_haveWildCards(false) 
-    {}
+    {
+	if (m_tp != SCLT_OR && m_tp != SCLT_AND) 
+	    m_tp = SCLT_OR;
+    }
     ~SearchData() {erase();}
 
     /** Make pristine */
@@ -91,6 +94,13 @@ public:
     /** We become the owner of cl and will delete it */
     bool addClause(SearchDataClause *cl);
 
+    /** If this is a simple query (one field only, no distance clauses),
+	add phrase made of query terms to query, so that docs containing the
+	user terms in order will have higher relevance. This must be called 
+	before toNativeQuery().
+    */
+    bool maybeAddAutoPhrase();
+
     /** Set/get top subdirectory for filtering results */
     void setTopdir(const string& t) {m_topdir = t;}
     string getTopdir() {return m_topdir;}
@@ -207,6 +217,8 @@ public:
     {
 	terms.insert(terms.end(), m_uterms.begin(), m_uterms.end());
     }
+    virtual const string& gettext() {return m_text;}
+    virtual const string& getfield() {return m_field;}
 protected:
     string  m_text;  // Raw user entry text.
     string  m_field; // Field specification if any