Merge branch 'fix/g07' into fix/integrated

# Conflicts:
#	internal/ipc/api.go
#	internal/ipc/client.go
#	internal/llm/client.go

internal/router/feeds.go

@@ -16,14 +16,25 @@ type FeedQuery struct {
 	Category string
 }
 
-// feedNouns — the words that make a question about the feeds themselves.
+// feedNouns — the words that name the feeds themselves. One of these is enough,
+// with an ask, to make the turn a feed question.
 var feedNouns = []string{
 	"лента", "ленте", "ленты", "лентах", "лентам",
 	"новости", "новостей", "новостях", "новостям",
-	"новое", "нового", "новенького",
 	"feed", "feeds", "news", "headlines",
 }
 
+// vagueNouns — the newness words that are NOT about the feeds by themselves.
+//
+// "что нового?" is the most common opener in the language and it is a greeting,
+// not a request for headlines. It used to match here, so the shipping daemon —
+// which has no feeds block — answered "я пока не читаю ленты, они не настроены",
+// a configuration status in reply to hello. With feeds on it answered "в лентах
+// пока ничего нового", which is no better. A vague noun claims the turn only
+// when the utterance narrows it: a named topic ("что нового по технологиям"), or
+// a feed noun somewhere in it ("что нового в лентах").
+var vagueNouns = []string{"новое", "нового", "новенького", "new"}
+
 // newnessMarkers — the "что нового" half. "нового" alone is in feedNouns
 // because it carries the question on its own ("что нового?"); a bare "лента"
 // needs the ask, which is what askMarkers below is for.
@@ -33,13 +44,15 @@ var askMarkers = []string{
 }
 
 // ParseFeedQuery reports whether an utterance asks what is new in the feeds, and
-// which topic if it names one after "по"/"о"/"про"/"about".
+// which topic if it names one after "по"/"об"/"про"/"about".
 //
-// Both a feed noun and an ask are required. "у меня новая лента в инстаграме" is
-// a statement and must not be read as a request to recite headlines.
+// A feed noun and an ask are required. "у меня новая лента в инстаграме" is a
+// statement and must not be read as a request to recite headlines. A vague
+// newness word counts as the noun only when a topic is named — see vagueNouns
+// for why the bare "что нового?" must fall through.
 func ParseFeedQuery(text string) (FeedQuery, bool) {
 	toks := planTokens(text)
-	noun, ask := false, false
+	noun, vague, ask := false, false, false
 	for _, t := range toks {
 		for _, n := range feedNouns {
 			if t == n {
@@ -47,6 +60,12 @@ func ParseFeedQuery(text string) (FeedQuery, bool) {
 				break
 			}
 		}
+		for _, n := range vagueNouns {
+			if t == n {
+				vague = true
+				break
+			}
+		}
 		for _, a := range askMarkers {
 			if t == a {
 				ask = true
@@ -54,24 +73,34 @@ func ParseFeedQuery(text string) (FeedQuery, bool) {
 			}
 		}
 	}
-	if !noun || !ask {
+	if !ask {
 		return FeedQuery{}, false
 	}
-	return FeedQuery{Category: feedCategory(toks)}, true
+	cat := feedCategory(toks)
+	if !noun && !(vague && cat != "") {
+		return FeedQuery{}, false
+	}
+	return FeedQuery{Category: cat}, true
 }
 
 // categoryPreps — the prepositions a topic follows. Russian marks the topic with
 // a preposition ("по технологиям", "про политику"), so the word after one is the
 // category; there is no stemming here, and the match against the configured
 // category is a prefix comparison for exactly that reason.
-var categoryPreps = map[string]bool{"по": true, "о": true, "об": true, "про": true, "about": true, "on": true}
+//
+// "о" is not in the list. It is one rune and it turns up as filler, a typo and
+// half of "о'кей", so any utterance carrying a stray "о" produced a category of
+// whatever word came next and she answered "по этой теме в лентах пока ничего"
+// to a question that named no theme. "об" and "про" carry the same meaning and
+// cannot be mistaken for anything else.
+var categoryPreps = map[string]bool{"по": true, "об": true, "про": true, "about": true, "on": true}
 
 func feedCategory(toks []string) string {
 	for i, t := range toks {
 		if categoryPreps[t] && i+1 < len(toks) {
 			next := toks[i+1]
 			// "по новостям" names no topic, it repeats the noun.
-			for _, n := range feedNouns {
+			for _, n := range append(append([]string{}, feedNouns...), vagueNouns...) {
 				if next == n {
 					return ""
 				}
@@ -82,12 +111,16 @@ func feedCategory(toks []string) string {
 	return ""
 }
 
-// CategoryMatches reports whether a note's text plausibly belongs to the
-// category he named. Russian inflects the topic ("технологиям" vs the configured
+// CategoryMatches reports whether a feed note's own category tag is the one he
+// named. Russian inflects the topic ("технологиям" vs the configured
 // "технологии"), and there is no stemmer in this repo, so the comparison is on a
 // common prefix — long enough that "полит" and "погод" stay apart, short enough
 // to survive a case ending.
-func CategoryMatches(text, category string) bool {
+//
+// tag is the note's stored category (rss.NoteCategory), NOT the whole note. It
+// used to be the whole note, which meant "что нового про погоду" matched any
+// tech headline whose link happened to contain "pogod".
+func CategoryMatches(tag, category string) bool {
 	if category == "" {
 		return true
 	}
@@ -95,7 +128,7 @@ func CategoryMatches(text, category string) bool {
 	if stem == "" {
 		return false
 	}
-	return strings.Contains(strings.ToLower(text), stem)
+	return strings.Contains(strings.ToLower(tag), stem)
 }
 
 // categoryStem cuts a word down to the part inflection leaves alone. 5 runes is

internal/router/feeds_test.go

@@ -9,13 +9,19 @@ func TestParseFeedQuery(t *testing.T) {
 		category string
 	}{
 		{"что нового в лентах?", true, ""},
-		{"что нового?", true, ""},
+		{"что нового по технологиям", true, "технологиям"},
 		{"какие новости?", true, ""},
 		{"что нового по технологиям?", true, "технологиям"},
 		{"расскажи новости про политику", true, "политику"},
 		{"что нового по новостям", true, ""},
 		{"what's new in the feeds?", true, ""},
 		{"any news about kubernetes", true, "kubernetes"},
+		// "что нового?" is a greeting. Claiming it made the shipping daemon
+		// answer hello with "я пока не читаю ленты — они не настроены".
+		{"что нового?", false, ""},
+		{"ну что нового", false, ""},
+		// A stray "о" is not a topic marker.
+		{"что нового в лентах, о боже", true, ""},
 		// Statements, not requests.
 		{"у меня новая лента в инстаграме", false, ""},
 		{"новости меня утомили", false, ""},
@@ -36,12 +42,16 @@ func TestParseFeedQuery(t *testing.T) {
 
 func TestCategoryMatches(t *testing.T) {
 	// The inflected form he says must match the form the config spells.
-	if !CategoryMatches("Новый релиз [технологии]", "технологиям") {
+	if !CategoryMatches("технологии", "технологиям") {
 		t.Error("inflected category did not match")
 	}
-	if CategoryMatches("Новый релиз [технологии]", "политику") {
+	if CategoryMatches("технологии", "политику") {
 		t.Error("unrelated category matched")
 	}
+	// The tag, not the note. The link in a tech headline is not a weather report.
+	if CategoryMatches("технологии", "погоду") {
+		t.Error("a tech note matched a weather question")
+	}
 	if !CategoryMatches("anything", "") {
 		t.Error("an empty category must match everything")
 	}