Merge the talk-fixture run and the JSON escape fix (V-44)

docs/evals/2026-08-05-talk-fixture-resident.md

@@ -0,0 +1,60 @@
+# Talk fixture against the resident model, 2026-08-05
+
+Vikunja #44 step 1. `MAVEN_LLM_URL=http://127.0.0.1:8899 make eval-phrasing`,
+Qwen3-1.7B-UD-Q4_K_XL on the host, no workstation in the run. The fixture holds
+36 cases now, against 27 when the bakeoff measured it. So the old score is not
+a column in this table.
+
+## Result
+
+| | before the escape fix | after |
+|---|---|---|
+| talk, passes every check | 2/36 (5.6%) | 25/36 (69.4%) |
+| failed generations | 31 | 0 |
+| by path: chat | 0/9 | 4/9 |
+| by path: knowledge | 1/9 | 6/9 |
+| by path: query | 0/9 | 9/9 |
+| by path: reply | 1/9 | 6/9 |
+| feminine | 5/36 | 36/36 |
+| address | 5/36 | 33/36 |
+| ontopic | 2/36 | 28/36 |
+| p50 latency | 3.05s | 2.97s |
+| nudges (15 cases) | 15/15 | 15/15 |
+
+## What the 31 errors were
+
+Not the model. `escapeRawControls` in `internal/phraser/llmphraser.go`, added
+for #537 to repair a raw newline written inside a string, escaped the whole
+object. Qwen3-1.7B pretty-prints: it opens `{` and writes three newlines before
+the first key. Those newlines became a literal backslash-n, which is legal
+nowhere outside a string, so the object stopped parsing and `parseResponseMood`
+reported `errBrokenJSON`.
+
+The comment said escaping unconditionally could not turn valid JSON into
+anything else, because JSON permits no control character outside a string. It
+permits three. Newline, tab and return are whitespace between tokens, and that
+is what pretty-printing is made of.
+
+Every chat reply and every knowledge answer the resident model wrote was being
+discarded for a stub line. The nudge path never showed it, because the nudge
+prompt gets compact JSON back.
+
+## The 11 that still fail
+
+Eight are `ontopic`, three are `address`.
+
+The address failures are all plural imperatives written to a formal listener:
+`держите`, `уточните`, `попробуйте`. Feminine self-reference held in all 36,
+which is the half #122 is training for. So the persona gap the CPT is aimed at
+is now the address half, not the gender half.
+
+The ontopic failures are the resident model answering next to the question
+rather than in it. `chat-joke` describes crying dolls instead of telling one,
+`know-hiccups` calls hiccups an icon, `know-boil-egg` answers about an omelette.
+`chat-about-me` answers "Я - записка", which is the same confabulation the
+bakeoff recorded.
+
+## Not measured here
+
+The workstation. Every number above is the homesrv floor. `make eval-phrasing`
+points at one URL, so a gemma-4-12b column needs its own run.

internal/phraser/broken_json_test.go

@@ -91,6 +91,39 @@ func TestEscapeRawControlsLeavesValidJSONAlone(t *testing.T) {
 	}
 }
 
+// Pretty-printed JSON is what Qwen3-1.7B writes: it opens the object and puts
+// three newlines before the first key. Escaping those structural newlines made
+// the object unparseable, so 31 of 36 conversational cases in the talk fixture
+// answered from the stub (Vikunja #44, measured 2026-08-05).
+func TestParseResponseMoodReadsPrettyPrintedJSON(t *testing.T) {
+	raw := "{\n\n\n \"response\": \"Хорошо настроение.\",\n  \"mood\": \"neutral\"\n}"
+	text, mood, err := parseResponseMood(raw)
+	if err != nil {
+		t.Fatalf("err = %v, want nil — this is valid JSON, not a failed generation", err)
+	}
+	if want := "Хорошо настроение."; text != want {
+		t.Errorf("response = %q, want %q", text, want)
+	}
+	if mood != "neutral" {
+		t.Errorf("mood = %q, want neutral", mood)
+	}
+}
+
+// Both at once: structural newlines outside the strings, a raw one inside.
+func TestParseResponseMoodRepairsInsideAndKeepsOutside(t *testing.T) {
+	raw := "{\n\t\"response\": \"первая\nвторая\",\n\t\"mood\": \"tired\"\n}"
+	text, mood, err := parseResponseMood(raw)
+	if err != nil {
+		t.Fatalf("err = %v, want nil", err)
+	}
+	if want := "первая\nвторая"; text != want {
+		t.Errorf("response = %q, want %q", text, want)
+	}
+	if mood != "tired" {
+		t.Errorf("mood = %q, want tired", mood)
+	}
+}
+
 // A tab and a bare control byte take the same path as the newline.
 func TestParseResponseMoodRepairsOtherControls(t *testing.T) {
 	raw := "{\"response\": \"таб\tи \x01байт\", \"mood\": \"tired\"}"

internal/phraser/llmphraser.go

@@ -1155,27 +1155,51 @@ func parseResponseMood(raw string) (response, mood string, err error) {
 // second line, for the paths that send no grammar at all — NoGrammar, and any
 // remote model whose server ignores one. A raw newline is the shape that was
 // measured; the rest of the range is here because the same argument covers it.
-// Only characters outside a string are affected in principle, and JSON permits
+//
-// none of this range outside a string either, so escaping unconditionally
+// Inside a string only. The first version escaped the whole object on the
-// cannot turn valid JSON into anything else.
+// argument that JSON permits no control character outside a string either, so
+// rewriting one could not do harm. That argument is wrong: JSON permits a
+// newline, a tab and a return BETWEEN tokens, which is what pretty-printing is.
+// Qwen3-1.7B pretty-prints — it opens `{` and writes three newlines before the
+// first key — and escaping those into a literal backslash-n broke every reply
+// it wrote. Measured 2026-08-05 on the talk fixture: 31 of 36 conversational
+// cases came back as errBrokenJSON and answered from the stub (Vikunja #44).
 func escapeRawControls(s string) string {
 	if !strings.ContainsFunc(s, func(r rune) bool { return r < 0x20 }) {
 		return s
 	}
 	var b strings.Builder
 	b.Grow(len(s) + 8)
+	inString := false
+	escaped := false
 	for _, r := range s {
 		switch {
+		case escaped:
+			// The character after a backslash is the model's own escape and is
+			// already whatever it meant to write.
+			escaped = false
+			b.WriteRune(r)
+			continue
+		case inString && r == '\\':
+			escaped = true
+			b.WriteRune(r)
+			continue
+		case r == '"':
+			inString = !inString
+			b.WriteRune(r)
+			continue
+		}
+		switch {
+		case !inString || r >= 0x20:
+			b.WriteRune(r)
 		case r == '\n':
 			b.WriteString(`\n`)
 		case r == '\r':
 			b.WriteString(`\r`)
 		case r == '\t':
 			b.WriteString(`\t`)
-		case r < 0x20:
-			fmt.Fprintf(&b, `\u%04x`, r)
 		default:
-			b.WriteRune(r)
+			fmt.Fprintf(&b, `\u%04x`, r)
 		}
 	}
 	return b.String()