ss.Optimize() had a logic bug; augment ss.CleanInput()

configatlaunch.go

@@ -211,9 +211,11 @@ func ConfigAtLaunch() {
 			"projurl":    PROJURL,
 			"vmodel":     Config.VectorModel,
 			"workers":    Config.WorkerCount}
+
 		t := template.Must(template.New("").Parse(HELPTEXTTEMPLATE))
+
 		var b bytes.Buffer
-		if err := t.Execute(&b, m); err != nil {
+		if ee := t.Execute(&b, m); ee != nil {
 			msg(FAIL7, MSGCRIT)
 		}
 		fmt.Println(styleoutput(coloroutput(b.String())))

fyi/README.md

@@ -158,13 +158,13 @@ self-test with vectors can be deceptive because `-wc` flag will not override con
       43 unique files.                              
        0 files ignored.
 
-github.com/AlDanial/cloc v 1.98  T=0.04 s (1087.2 files/s, 489430.3 lines/s)
+github.com/AlDanial/cloc v 1.98  T=0.04 s (1046.2 files/s, 472237.8 lines/s)
 -------------------------------------------------------------------------------
 Language                     files          blank        comment           code
 -------------------------------------------------------------------------------
-Go                              43           2889           3193          13275
+Go                              43           2901           3204          13305
 -------------------------------------------------------------------------------
-SUM:                            43           2889           3193          13275
+SUM:                            43           2901           3204          13305
 -------------------------------------------------------------------------------
 
 ```

go.mod

@@ -6,8 +6,8 @@ require (
 	github.com/e-gun/nlp v0.0.0-20230418221101-577c2209ffcc
 	github.com/e-gun/tsnemp v0.0.0
 	github.com/e-gun/wego v0.0.11
-	github.com/go-echarts/go-echarts/v2 v2.3.1
-	github.com/google/uuid v1.3.1
+	github.com/go-echarts/go-echarts/v2 v2.3.2
+	github.com/google/uuid v1.4.0
 	github.com/gorilla/websocket v1.5.0
 	github.com/jackc/pgx/v5 v5.4.3
 	github.com/json-iterator/go v1.1.12
@@ -23,7 +23,7 @@ require (
 	github.com/e-gun/sparse v0.0.0-20230418220937-07063da15582 // indirect
 	github.com/felixge/fgprof v0.9.3 // indirect
 	github.com/golang-jwt/jwt v3.2.2+incompatible // indirect
-	github.com/google/pprof v0.0.0-20230926050212-f7f687d19a98 // indirect
+	github.com/google/pprof v0.0.0-20231101202521-4ca4178f5c7a // indirect
 	github.com/inconshreveable/mousetrap v1.1.0 // indirect
 	github.com/jackc/pgpassfile v1.0.0 // indirect
 	github.com/jackc/pgservicefile v0.0.0-20221227161230-091c0ba34f0a // indirect

rt-getters.go

@@ -355,15 +355,16 @@ func RtGetJSSearchlist(c echo.Context) error {
 
 	m := message.NewPrinter(language.English)
 	sl := SessionIntoSearchlist(sess)
-	tw := 0
+
+	totalwords := 0
 
 	var wkk []string
 	for _, a := range sl.Inc.Authors {
 		for _, w := range AllAuthors[a].WorkList {
 			ct := WORKTMPL
 			cf := m.Sprintf(ct, AllAuthors[a].Cleaname, AllWorks[w].Title, AllWorks[w].WdCount)
 			wkk = append(wkk, cf)
-			tw += AllWorks[w].WdCount
+			totalwords += AllWorks[w].WdCount
 		}
 	}
 
@@ -372,20 +373,20 @@ func RtGetJSSearchlist(c echo.Context) error {
 		ct := WORKTMPL
 		cf := m.Sprintf(ct, thiswk.MyAu().Cleaname, thiswk.Title, thiswk.WdCount)
 		wkk = append(wkk, cf)
-		tw += thiswk.WdCount
+		totalwords += thiswk.WdCount
 	}
 
 	pattern := regexp.MustCompile(REG)
 	for _, p := range sl.Inc.Passages {
 		cit, count := searchlistpassages(pattern, p)
 		wkk = append(wkk, cit)
-		tw += count
+		totalwords += count
 	}
 
 	for _, p := range sl.Excl.Passages {
 		cit, count := searchlistpassages(pattern, p)
 		wkk = append(wkk, cit+"[EXCLUDED]")
-		tw -= count
+		totalwords -= count
 	}
 
 	if len(wkk) > MAXSEARCHINFOLISTLEN {
@@ -394,7 +395,7 @@ func RtGetJSSearchlist(c echo.Context) error {
 		wkk = append(wkk, m.Sprintf(SPILLOVER, diff))
 	}
 
-	wkk = append(wkk, m.Sprintf(SUMMARY, tw))
+	wkk = append(wkk, m.Sprintf(SUMMARY, totalwords))
 
 	ht := strings.Join(wkk, "<br>\n")
 	var j JSStruct

searchstructs.go

@@ -89,6 +89,21 @@ func (s *SearchStruct) CleanInput() {
 
 	s.Seeking = Purgechars(dropping, s.Seeking)
 	s.Proximate = Purgechars(dropping, s.Proximate)
+
+	// don't let BoxA be blank if BoxB is not
+	BoxA := s.Seeking == "" && s.LemmaOne == ""
+	NotBoxB := s.Proximate != "" || s.LemmaTwo != ""
+
+	if BoxA && NotBoxB {
+		if s.Proximate != "" {
+			s.Seeking = s.Proximate
+			s.Proximate = ""
+		}
+		if s.LemmaTwo != "" {
+			s.LemmaOne = s.LemmaTwo
+			s.LemmaTwo = ""
+		}
+	}
 }
 
 // SetType - set internal values via self-probe
@@ -115,15 +130,15 @@ func (s *SearchStruct) SetType() {
 		s.HasPhraseBoxB = true
 	}
 
-	if len(s.LemmaOne) != 0 {
+	if s.LemmaOne != "" {
 		s.HasLemmaBoxA = true
 		// accented line has "volat" in latin; and "uolo" will not find it
 		if isGreek.MatchString(s.LemmaOne) {
 			s.SrchColumn = "accented_line"
 		}
 	}
 
-	if len(s.LemmaTwo) != 0 {
+	if s.LemmaTwo != "" {
 		s.HasLemmaBoxB = true
 	}
 
@@ -258,8 +273,10 @@ func (s *SearchStruct) InclusionOverview(sessincl SearchIncExl) string {
 	return r
 }
 
-// Optimize - consider rewriting the search to make it faster
+// Optimize - think about rewriting the search to make it faster
 func (s *SearchStruct) Optimize() {
+	// only zero or one of the following should be true
+
 	// if BoxA has a lemma and BoxB has a phrase, it is almost certainly faster to search B, then A...
 	if s.HasLemmaBoxA && s.HasPhraseBoxB {
 		s.SwapPhraseAndLemma()
@@ -272,14 +289,16 @@ func (s *SearchStruct) Optimize() {
 		return
 	}
 
-	// a single word should be faster than a lemma
-	if s.HasLemmaBoxA && !s.HasPhraseBoxB {
+	// a single word should be faster than a lemma; but do not swap an empty string
+	if s.HasLemmaBoxA && !s.HasPhraseBoxB && s.Proximate != "" {
 		s.SwapWordAndLemma()
+		return
 	}
 
 	// consider looking for the string with more characters in it first
 	if len(s.Seeking) > 0 && len(s.Proximate) > 0 {
 		s.SearchQuickestFirst()
+		return
 	}
 }
 

vectorquerylda.go

@@ -78,7 +78,7 @@ func LDASearch(c echo.Context, srch SearchStruct) error {
 		LDAMSG = `Building LDA model for the current selections`
 		ESM1   = "<br>preparing the text for modeling"
 		ESM2   = "<br>building topic models"
-		ESM3   = "<br>using t-Distributed Stochastic Neighbor Embedding to build graph"
+		ESM3   = "<br>using t-Distributed Stochastic Neighbor Embedding to build graph (please be patient...)"
 	)
 	c.Response().After(func() { messenger.LogPaths("LDASearch()") })