Clean built-in couple formatter prose

caption_metadata_routes.py

@@ -78,6 +78,21 @@ def couple_clothing_sentence(clothing: str, clean_text: Callable[[Any], str]) ->
     return f"They wear {clothing}"
 
 
+def couple_subject_sentence(
+    subject: str,
+    ages: str,
+    cap_first: Callable[[str], str],
+    clean_age_phrase: Callable[[str], str],
+) -> str:
+    subject = cap_first(subject or "adult couple")
+    ages = clean_age_phrase(ages)
+    if ages:
+        return f"{subject}, {ages}"
+    if subject.lower() == "adult couple":
+        return subject
+    return f"{subject} are adults"
+
+
 def single_from_row_result(
     request: CaptionMetadataRouteRequest,
     deps: CaptionMetadataRouteDependencies,
@@ -176,9 +191,7 @@ def couple_from_row_result(
     camera_scene = deps.clean_text(row.get("camera_scene_directive"))
     style = deps.field_row_value(row, "style") if keep_style else ""
 
-    parts = [f"{deps.cap_first(subject)} are adults"]
-    if ages:
-        parts.append(f"The age detail is {deps.clean_age_phrase(ages)}")
+    parts = [couple_subject_sentence(subject, ages, deps.cap_first, deps.clean_age_phrase)]
     if body:
         parts.append(f"Their body types are {body}")
     if clothing:

krea_normal_formatter.py

@@ -54,6 +54,19 @@ def _couple_clothing_phrase(item: str, clean: Callable[[Any], str]) -> str:
     return f"The couple wears {item}"
 
 
+def _cap_first(text: str) -> str:
+    text = str(text or "").strip()
+    return f"{text[:1].upper()}{text[1:]}" if text else ""
+
+
+def _couple_subject_phrase(subject: str, ages: str) -> str:
+    subject = _cap_first(subject or "adult couple")
+    ages = str(ages or "").strip()
+    if ages:
+        return f"{subject}, {ages}"
+    return subject
+
+
 def format_normal_row_result(
     request: KreaNormalRowRequest,
     deps: KreaNormalRowDependencies,
@@ -98,8 +111,7 @@ def format_normal_row_result(
         ages = deps.age_detail_phrase(deps.row_value(row, "age", ("Ages",)) or row.get("age_band"))
         body = deps.row_value(row, "body", ("Body types",)) or deps.clean(row.get("body_type"))
         parts = [
-            f"An adult couple: {subject}, all visibly adult",
-            f"Age detail: {ages}" if ages else "",
+            _couple_subject_phrase(subject, ages),
             f"Body types: {body}" if body else "",
             _couple_clothing_phrase(item, deps.clean) if item else "",
             f"The pose is {pose}" if pose else "",

tools/prompt_smoke.py

@@ -638,6 +638,63 @@ def smoke_builtin_single() -> None:
     _expect_formatter_outputs(row, "builtin_single_woman", target="single")
 
 
+def smoke_builtin_couple_metadata() -> None:
+    row = _prompt_row(name="builtin_couple", category="couple", subcategory="random", seed=1003, men_count=0)
+    _expect(row.get("source") == "built_in_generator", "builtin couple row should come from built-in generator")
+    _expect(row.get("subject_type") == "couple", "builtin couple row lost normalized subject_type")
+    _expect(row.get("women_count") + row.get("men_count") == row.get("person_count"), "builtin couple row lost count consistency")
+    _expect(row.get("person_count") == 2, "builtin couple row lost normalized person count")
+    subject = _expect_text("builtin_couple.subject_phrase", row.get("subject_phrase"), 5)
+    age = _expect_text("builtin_couple.age_band", row.get("age_band"), 8)
+    body = _expect_text("builtin_couple.body_type", row.get("body_type"), 5)
+    scene = _expect_text("builtin_couple.scene_text", row.get("scene_text"), 12)
+    item = _expect_text("builtin_couple.item", row.get("item"), 8)
+    pose = _expect_text("builtin_couple.pose", row.get("pose"), 8)
+    expression = _expect_text("builtin_couple.expression", row.get("expression"), 8)
+    composition = _expect_text("builtin_couple.composition", row.get("composition"), 8)
+
+    metadata_only = dict(row)
+    metadata_only["prompt"] = ""
+    metadata_only["caption"] = ""
+    metadata = _json(metadata_only)
+    krea = krea_formatter.format_krea2_prompt("", metadata_json=metadata, target="auto")
+    sdxl = sdxl_formatter.format_sdxl_prompt(
+        "",
+        metadata_json=metadata,
+        target="auto",
+        trigger=SdxlTrigger,
+        prepend_trigger=True,
+    )
+    caption, caption_method = caption_naturalizer.naturalize_caption(
+        "",
+        metadata_json=metadata,
+        target="auto",
+        trigger=Trigger,
+        include_trigger=True,
+    )
+
+    clean_age = caption_text_policy.clean_age_phrase(age)
+    krea_prompt = str(krea.get("krea_prompt", ""))
+    sdxl_prompt = str(sdxl.get("sdxl_prompt", "")).lower()
+    _expect("metadata(couple)" in krea.get("method", ""), "Krea built-in couple route did not use metadata")
+    _expect(subject.capitalize() in krea_prompt and clean_age in krea_prompt, "Krea built-in couple route lost subject or age")
+    _expect("all visibly adult" not in krea_prompt.lower(), "Krea built-in couple route kept redundant adult noise")
+    _expect("Age detail:" not in krea_prompt, "Krea built-in couple route kept age helper label")
+    for value in (body, scene, item, pose, expression, composition):
+        _expect(value in krea_prompt, f"Krea built-in couple route lost metadata value: {value}")
+
+    _expect("2women" in sdxl_prompt or "1woman, 1man" in sdxl_prompt or "2men" in sdxl_prompt, "SDXL built-in couple route lost count tags")
+    for value in (body, scene, item, pose, expression, composition):
+        for tag in sdxl_tag_policy.split_tag_text(value):
+            _expect(tag.lower() in sdxl_prompt, f"SDXL built-in couple route lost metadata tag: {tag}")
+
+    _expect(caption_method.endswith("metadata(couple)"), "Caption built-in couple route did not use metadata")
+    _expect(clean_age in caption and scene in caption, "Caption built-in couple route lost age or scene")
+    _expect("The age detail is" not in caption, "Caption built-in couple route kept age helper sentence")
+    for value in (body, item, pose, expression, composition):
+        _expect(value in caption, f"Caption built-in couple route lost metadata value: {value}")
+
+
 def smoke_camera_scene_single() -> None:
     row = _prompt_row(
         name="camera_scene_single",
@@ -8022,6 +8079,7 @@ def smoke_node_profile_filter_registration() -> None:
 
 SMOKE_CASES: list[tuple[str, Callable[[], None]]] = [
     ("builtin_single_woman", smoke_builtin_single),
+    ("builtin_couple_metadata", smoke_builtin_couple_metadata),
     ("camera_scene_single", smoke_camera_scene_single),
     ("row_camera_policy", smoke_row_camera_policy),
     ("config_route_location_theme", smoke_config_route_location_theme),