SlabUploader/backend/tests/test_inference.py

"""S7 inference tests (METADATA-105/106, 45, 55, 70, SETTINGS-45)."""

import base64
import json
import os
import shutil
import tempfile

import pytest
from fastapi.testclient import TestClient

import backend.db as db_mod


@pytest.fixture
def env(monkeypatch):
    tmp = tempfile.mkdtemp(prefix="inf_")
    os.environ["SLAB_DB_PATH"] = os.path.join(tmp, "i.db")
    os.environ["SLAB_IMAGES_DIR"] = tmp
    os.environ["WOO_BASE_URL"] = "https://x.com"
    os.environ["SLAB_AES_KEY"] = base64.b64encode(b"0123456789abcdef0123456789abcdef").decode()
    db_mod._reset_for_tests()
    from backend.crypto import encrypt

    conn = db_mod.init_db()
    conn.execute("INSERT OR REPLACE INTO settings(key,value,value_enc) VALUES ('inference_base_url',NULL,?)", (encrypt("https://infer.example.com"),))
    conn.execute("INSERT OR REPLACE INTO settings(key,value,value_enc) VALUES ('inference_api_key',NULL,?)", (encrypt("sk-test"),))
    conn.execute("INSERT OR REPLACE INTO settings(key,value,value_enc) VALUES ('inference_model','gpt-test',NULL)")
    conn.commit()
    from backend.app import app

    yield TestClient(app), conn
    db_mod._reset_for_tests()
    shutil.rmtree(tmp, ignore_errors=True)


def test_prompt_seed_and_revision(env):
    client, conn = env
    r = client.get("/api/v1/prompts/call1")
    assert r.status_code == 200
    body = r.json()
    assert body["active_content"] is not None
    assert len(body["revisions"]) == 1  # seed only

    r = client.post("/api/v1/prompts/call1/revisions", json={"content": "revised prompt v2"})
    assert r.status_code == 201
    new_rev = r.json()["id"]
    # new revision becomes active
    r = client.get("/api/v1/prompts/call1")
    assert r.json()["active_revision_id"] == new_rev
    assert r.json()["active_content"] == "revised prompt v2"
    assert len(r.json()["revisions"]) == 2

    # activate back to seed
    seed_rev = 1
    r = client.post("/api/v1/prompts/call1/activate", json={"revision_id": seed_rev})
    assert r.status_code == 200
    assert client.get("/api/v1/prompts/call1").json()["active_content"] is not None


def test_call1_seed_is_locked_expert_prompt(env):
    """#86: shipped Call 1 seed is the locked expert prompt (no free-text bags, no grade)."""
    client, _conn = env
    from backend.inference import CALL1_SEED

    body = client.get("/api/v1/prompts/call1").json()
    assert body["active_content"] == CALL1_SEED
    assert "wood identification expert" in CALL1_SEED
    assert "feature_tag_ids" in CALL1_SEED
    assert "Decide wood categories before edge type" in CALL1_SEED
    assert "highest-confidence" in CALL1_SEED
    assert "Wild edge" in CALL1_SEED
    assert "cookies and burl" in CALL1_SEED.lower() or "cookies and burl" in CALL1_SEED
    assert "grade_term_id" not in CALL1_SEED
    assert '"character"' not in CALL1_SEED
    assert '"inclusions"' not in CALL1_SEED
    assert "character\":" not in CALL1_SEED.replace(" ", "")


def test_call2_seed_has_short_long_structure(env):
    """Call 2 seed carries short+long structure and JSON dual output (not brand voice)."""
    client, _conn = env
    from backend.inference import CALL2_SEED

    body = client.get("/api/v1/prompts/call2").json()
    assert body["active_content"] == CALL2_SEED
    assert "SHORT DESCRIPTION" in CALL2_SEED
    assert "LONG DESCRIPTION" in CALL2_SEED
    assert "Lead with the outcome" in CALL2_SEED
    assert "Hook" in CALL2_SEED
    assert "Brand voice and Geography sections when provided" in CALL2_SEED
    assert "Whidbey" not in CALL2_SEED
    assert "Pacific Northwest" not in CALL2_SEED
    assert '"short_description"' in CALL2_SEED
    assert '"description"' in CALL2_SEED


def test_taxonomy_lists_call1_closed_shape():
    """#86: Call 1 lists are closed Woo picks only — no grade/thickness/moisture; feat names stripped."""
    import tempfile, os, shutil
    tmp = tempfile.mkdtemp(prefix="tax_")
    os.environ["SLAB_DB_PATH"] = os.path.join(tmp, "t.db")
    os.environ["SLAB_IMAGES_DIR"] = tmp
    os.environ.setdefault("WOO_BASE_URL", "https://x.com")
    os.environ.setdefault(
        "SLAB_AES_KEY",
        base64.b64encode(b"0123456789abcdef0123456789abcdef").decode(),
    )
    db_mod._reset_for_tests()
    conn = db_mod.init_db()
    conn.execute("INSERT INTO woo_categories(id,name,slug,parent_id,woo_id) VALUES (10,'Species','species',NULL,10)")
    conn.execute("INSERT INTO woo_categories(id,name,slug,parent_id,woo_id) VALUES (11,'Walnut','walnut',10,11)")
    conn.execute("INSERT INTO woo_categories(id,name,slug,parent_id,woo_id) VALUES (20,'Wcat','wcat',NULL,20)")
    conn.execute("INSERT INTO woo_categories(id,name,slug,parent_id,woo_id) VALUES (21,'Live edge wood','liveedge',20,21)")
    conn.execute("INSERT INTO woo_attributes(id,name,slug,woo_id) VALUES (6,'Edge type','pa_etype',6)")
    conn.execute("INSERT INTO woo_attributes(id,name,slug,woo_id) VALUES (8,'Figure','pa_fig',8)")
    conn.execute("INSERT INTO woo_attributes(id,name,slug,woo_id) VALUES (9,'Grade','pa_gr',9)")
    conn.execute("INSERT INTO woo_attributes(id,name,slug,woo_id) VALUES (10,'Thickness','pa_thk',10)")
    conn.execute("INSERT INTO woo_attributes(id,name,slug,woo_id) VALUES (7,'Moisture','pa_mst',7)")
    conn.execute("INSERT INTO woo_attribute_terms(id,attribute_id,name,slug,woo_id) VALUES (321,6,'Square edge','sqe',321)")
    conn.execute("INSERT INTO woo_attribute_terms(id,attribute_id,name,slug,woo_id) VALUES (328,8,'Plain','pl',328)")
    conn.execute("INSERT INTO woo_attribute_terms(id,attribute_id,name,slug,woo_id) VALUES (342,9,'Fancy','fnc',342)")
    conn.execute("INSERT INTO woo_attribute_terms(id,attribute_id,name,slug,woo_id) VALUES (350,10,'1\"- 1 1/2\"','4-6qtr',350)")
    conn.execute("INSERT INTO woo_attribute_terms(id,attribute_id,name,slug,woo_id) VALUES (326,7,'Kiln dried','kd',326)")
    conn.execute("INSERT INTO woo_tags(id,name,slug,woo_id) VALUES (1,'feat-knot','feat-knot',1)")
    conn.execute("INSERT INTO woo_tags(id,name,slug,woo_id) VALUES (2,'fig-curl','fig-curl',2)")
    conn.commit()
    from backend.inference import _taxonomy_lists

    tax = _taxonomy_lists(conn)
    assert set(tax.keys()) == {"species", "wood_categories", "edge_terms", "figure_terms", "feature_tags"}
    assert tax["species"] == [{"id": 11, "name": "Walnut"}]
    assert tax["wood_categories"] == [{"id": 21, "name": "Live edge wood"}]
    assert tax["edge_terms"] == [{"id": 321, "name": "Square edge"}]
    assert tax["figure_terms"] == [{"id": 328, "name": "Plain"}]
    # feat- prefix stripped for the model; fig- tags are not feature_tags
    assert tax["feature_tags"] == [{"id": 1, "name": "knot"}]
    assert "grade_terms" not in tax
    assert "thickness_terms" not in tax
    assert "moisture_terms" not in tax
    db_mod._reset_for_tests()
    shutil.rmtree(tmp, ignore_errors=True)


def test_gate_feature_tag_ids_not_free_text():
    """#86: gate keeps feature_tag_ids at threshold; drops free-text bags and grade."""
    from backend.inference import _gate

    parsed = {
        "species_id": 11, "species_confidence": 0.9,
        "wood_category_ids": [21], "wood_category_confidence": 0.8,
        "edge_type_term_id": 321, "edge_confidence": 0.85,
        "figure_term_ids": [328, 335], "figure_confidence": 0.75,
        "feature_tag_ids": [1], "feature_tag_confidence": 0.8,
        "grade_term_id": 342, "grade_confidence": 0.99,
        "character": ["rustic"], "inclusions": ["bark"], "voids": [], "checks": ["end"],
    }
    out = _gate(parsed)
    assert out["species_id"] == 11
    assert out["wood_category_ids"] == [21]
    assert out["edge_type_term_id"] == 321
    assert out["figure_term_ids"] == [328, 335]
    assert out["feature_tag_ids"] == [1]
    assert "grade_term_id" not in out
    assert "character" not in out
    assert "inclusions" not in out
    assert "voids" not in out
    assert "checks" not in out
    # below threshold feature tags dropped
    low = {**parsed, "feature_tag_confidence": 0.2}
    assert "feature_tag_ids" not in _gate(low)


def test_call1_hard_disabled_410(env, monkeypatch):
    """overnight-whole-app: Call 1 OUT — POST always 410, never chats the model."""
    client, conn = env
    conn.execute("INSERT OR REPLACE INTO settings(key,value,value_enc) VALUES ('inference_enabled','true',NULL)")
    conn.execute("INSERT INTO slabs(id,sku,length_in,thickness_in,status,server_rev,client_rev,created_at,updated_at) VALUES ('s1','SLAB-01',100,2,'calibrated',1,1,'t','t')")
    conn.execute("INSERT INTO slab_photos(id,slab_id,seq,processed_blob,created_at) VALUES ('p1','s1',1,X'89504E47','t')")
    conn.commit()
    from backend.inference import InferenceClient

    called = {"n": 0}

    def fake_chat(self, messages):
        called["n"] += 1
        return "{}"

    monkeypatch.setattr(InferenceClient, "chat", fake_chat)
    r = client.post("/api/v1/slabs/s1/inference/call1")
    assert r.status_code == 410, r.text
    body = r.json()
    assert body["error"]["code"] == "call1_disabled"
    assert called["n"] == 0
    assert conn.execute("SELECT COUNT(*) AS n FROM inference_log").fetchone()["n"] == 0


def test_call2_stores_short_and_description(env, monkeypatch):
    """Call 2 persists both short_description and description from JSON output."""
    client, conn = env
    conn.execute("INSERT OR REPLACE INTO settings(key,value,value_enc) VALUES ('content_llm_enabled','true',NULL)")
    conn.execute(
        "INSERT INTO slabs(id,sku,length_in,thickness_in,width_avg_in,bdft,title,short_description,"
        "status,server_rev,client_rev,created_at,updated_at) "
        "VALUES ('s1','SLAB-01',100,2,20,18,'Walnut Slab','old short','ready',1,1,'t','t')"
    )
    conn.execute("INSERT INTO slab_photos(id,slab_id,seq,processed_blob,created_at) VALUES ('p1','s1',1,X'89504E47','t')")
    conn.commit()
    from backend.inference import InferenceClient
    import json as _json

    payload = _json.dumps({
        "short_description": "Turn a blank into a showpiece table — kiln-dried walnut with live edge character. Ready to finish.",
        "description": "<p><strong>Showpiece ready.</strong></p><ul><li><strong>Stable:</strong> kiln dried</li></ul>",
    })
    monkeypatch.setattr(InferenceClient, "chat", lambda self, messages: payload)
    r = client.post("/api/v1/slabs/s1/inference/call2")
    assert r.status_code == 200, r.text
    body = r.json()
    assert "Showpiece ready" in body["description"] or "kiln" in body["description"]
    assert body["short_description"].startswith("Turn a blank")
    row = conn.execute(
        "SELECT short_description, description, content_source FROM slabs WHERE id='s1'"
    ).fetchone()
    assert row["short_description"].startswith("Turn a blank")
    assert "<p>" in row["description"]
    assert row["content_source"] == "llm"


def test_call2_html_fallback_keeps_prior_short(env, monkeypatch):
    """Non-JSON Call 2 body becomes description only; prior short is left alone."""
    client, conn = env
    conn.execute("INSERT OR REPLACE INTO settings(key,value,value_enc) VALUES ('content_llm_enabled','true',NULL)")
    conn.execute(
        "INSERT INTO slabs(id,sku,length_in,thickness_in,width_avg_in,bdft,title,short_description,"
        "status,server_rev,client_rev,created_at,updated_at) "
        "VALUES ('s1','SLAB-01',100,2,20,18,'Walnut Slab','keep me','ready',1,1,'t','t')"
    )
    conn.execute("INSERT INTO slab_photos(id,slab_id,seq,processed_blob,created_at) VALUES ('p1','s1',1,X'89504E47','t')")
    conn.commit()
    from backend.inference import InferenceClient

    monkeypatch.setattr(InferenceClient, "chat", lambda self, messages: "<p>A fine walnut slab.</p>")
    r = client.post("/api/v1/slabs/s1/inference/call2")
    assert r.status_code == 200, r.text
    assert r.json()["description"] == "<p>A fine walnut slab.</p>"
    assert "short_description" not in r.json()
    row = conn.execute(
        "SELECT short_description, description, content_source FROM slabs WHERE id='s1'"
    ).fetchone()
    assert row["short_description"] == "keep me"
    assert row["description"] == "<p>A fine walnut slab.</p>"
    assert row["content_source"] == "llm"


def test_parse_call2_response_json_and_fenced():
    from backend.inference import parse_call2_response
    import json as _json

    short, desc = parse_call2_response(
        _json.dumps({"short_description": " Benefit first. ", "description": " <p>Long</p> "})
    )
    assert short == "Benefit first."
    assert desc == "<p>Long</p>"
    fenced = "```json\n" + _json.dumps({"short_description": "S", "description": "<p>D</p>"}) + "\n```"
    short2, desc2 = parse_call2_response(fenced)
    assert short2 == "S"
    assert desc2 == "<p>D</p>"
    short3, desc3 = parse_call2_response("<p>only long</p>")
    assert short3 is None
    assert desc3 == "<p>only long</p>"


def test_call1_disabled_even_when_inference_off(env):
    """Call 1 hard-disable wins over inference_enabled=false (still 410, not 409)."""
    client, conn = env
    conn.execute("INSERT INTO slabs(id,sku,status,server_rev,client_rev,created_at,updated_at) VALUES ('s1','SLAB-01','calibrated',1,1,'t','t')")
    conn.commit()
    r = client.post("/api/v1/slabs/s1/inference/call1")
    assert r.status_code == 410
    assert r.json()["error"]["code"] == "call1_disabled"


def test_server_description_inference_off(env):
    from backend.inference import server_description

    d = server_description({"title": "Figured Walnut Slab", "length_in": 114.0, "width_avg_in": 24.0, "thickness_in": 1.75, "bdft": 18.4})
    assert "114.0in" in d and "18.4" in d


def test_test_inference_unconfigured(env):
    client, conn = env
    conn.execute("UPDATE settings SET value_enc=NULL WHERE key='inference_api_key'")
    conn.commit()
    r = client.post("/api/v1/admin/test-inference")
    assert r.status_code == 200
    assert r.json()["ok"] is False and r.json()["mode"] == "fail"


def _rgba_png(w: int, h: int) -> bytes:
    """RGBA PNG: opaque body + transparent top-left block (capture-like alpha)."""
    from PIL import Image
    import io
    im = Image.new("RGBA", (w, h), (10, 20, 30, 255))
    for y in range(min(80, h)):
        for x in range(min(80, w)):
            im.putpixel((x, y), (0, 0, 0, 0))
    buf = io.BytesIO()
    im.save(buf, format="PNG")
    return buf.getvalue()


def test_fit_long_edge_png_downscales_over_1024():
    """METADATA-35: Call 1 images long edge <= 1024; never upscale; keep alpha."""
    from backend.inference import fit_long_edge_png
    from PIL import Image
    import io

    src = _rgba_png(800, 1067)  # inventory 3:4, long edge 1067
    out = fit_long_edge_png(src, max_long_edge=1024)
    im = Image.open(io.BytesIO(out)).convert("RGBA")
    assert max(im.size) == 1024
    assert min(im.size) < 1024
    w, h = im.size
    assert abs((w / h) - (800 / 1067)) < 0.02
    # Transparent block still transparent near origin after downscale
    assert im.getpixel((0, 0))[3] == 0
    assert im.getpixel((5, 5))[3] == 0


def test_fit_long_edge_png_leaves_small_unchanged_dims():
    from backend.inference import fit_long_edge_png
    from PIL import Image
    import io

    src = _rgba_png(400, 300)
    out = fit_long_edge_png(src, max_long_edge=1024)
    im = Image.open(io.BytesIO(out))
    assert im.size == (400, 300)


def test_redact_request_for_log_strips_data_urls_keeps_text():
    """#87: log sanitizer replaces data-URL images with photo refs; keeps text."""
    from backend.inference import _redact_request_for_log

    b64 = "iVBORw0KGgoAAAANSUhEUgAAAAEAAAABCAYAAAAfFcSJAAAADUlEQVR42mP8z8BQDwAEhQGAhKmMIQAAAABJRU5ErkJggg=="
    messages = [{
        "role": "user",
        "content": [
            {"type": "image_url", "image_url": {"url": f"data:image/png;base64,{b64}"}},
            {"type": "image_url", "image_url": {"url": f"data:image/png;base64,{b64 * 20}"}},
            {"type": "text", "text": "LISTS:\n{\"species\":[]}\n\nSKU=X length_in=1 thickness_in=2"},
        ],
    }]
    meta = [
        {"photo_id": "p1", "seq": 1, "bytes": 100, "infer_long_edge_px": 64},
        {"photo_id": "p2", "seq": 2, "bytes": 200, "infer_long_edge_px": 128},
    ]
    out = _redact_request_for_log(messages, meta)
    blob = json.dumps(out)
    assert "data:image" not in blob
    assert "base64" not in blob.lower()
    assert "LISTS:" in blob
    assert "SKU=X" in blob
    parts = out[0]["content"]
    assert parts[0]["type"] == "image_ref"
    assert parts[0]["photo_id"] == "p1"
    assert parts[0]["seq"] == 1
    assert parts[1]["photo_id"] == "p2"
    assert parts[2]["type"] == "text"
    assert "LISTS:" in parts[2]["text"]


def test_call1_log_skipped_while_disabled(env):
    """#87 Call 1 log path dormant while Call 1 is hard-disabled (overnight-whole-app)."""
    client, conn = env
    conn.execute("INSERT OR REPLACE INTO settings(key,value,value_enc) VALUES ('inference_enabled','true',NULL)")
    conn.execute(
        "INSERT INTO slabs(id,sku,length_in,thickness_in,status,server_rev,client_rev,created_at,updated_at) "
        "VALUES ('s1','SLAB-01',100,2,'calibrated',1,1,'t','t')"
    )
    conn.commit()
    r = client.post("/api/v1/slabs/s1/inference/call1")
    assert r.status_code == 410
    assert conn.execute("SELECT COUNT(*) AS n FROM inference_log").fetchone()["n"] == 0


def test_call2_log_redacts_images(env, monkeypatch):
    """#87: Call 2 log also redacts image data-URLs."""
    client, conn = env
    conn.execute("INSERT OR REPLACE INTO settings(key,value,value_enc) VALUES ('content_llm_enabled','true',NULL)")
    conn.execute(
        "INSERT INTO slabs(id,sku,length_in,thickness_in,width_avg_in,bdft,title,status,server_rev,client_rev,created_at,updated_at) "
        "VALUES ('s1','SLAB-01',100,2,20,18,'Walnut','ready',1,1,'t','t')"
    )
    png = _rgba_png(200, 200)
    conn.execute(
        "INSERT INTO slab_photos(id,slab_id,seq,processed_blob,created_at) VALUES ('p2','s1',1,?, 't')",
        (png,),
    )
    conn.commit()
    from backend.inference import InferenceClient

    monkeypatch.setattr(InferenceClient, "chat", lambda self, messages: "<p>ok</p>")
    r = client.post("/api/v1/slabs/s1/inference/call2")
    assert r.status_code == 200, r.text
    payload = conn.execute(
        "SELECT request_payload FROM inference_log WHERE call=2 ORDER BY id DESC LIMIT 1"
    ).fetchone()["request_payload"]
    assert "data:image" not in payload
    assert "image_ref" in payload
    assert "p2" in payload
    assert "DIMENSION FACT BLOCK" in payload


def test_call1_sends_long_edge_dormant_while_disabled(env):
    """METADATA-35 Call 1 sizing path dormant while Call 1 is hard-disabled."""
    client, _conn = env
    r = client.post("/api/v1/slabs/missing/inference/call1")
    assert r.status_code == 410
    assert r.json()["error"]["code"] == "call1_disabled"


def test_assemble_call2_includes_declared_attrs_and_dims(env):
    """Call 2 text = seed + DECLARED ATTRIBUTES + dimension fact + brand/geo; shared assembler."""
    client, conn = env
    conn.execute(
        "INSERT INTO woo_categories(id,name,slug,parent_id,woo_id) VALUES "
        "(10,'Species','species',NULL,10),"
        "(11,'Black Walnut','black-walnut',10,11),"
        "(20,'Wcat','wcat',NULL,20),"
        "(21,'Slabs & Lumber','snl',20,21),"
        "(22,'Live edge wood','liveedge',21,22)"
    )
    conn.execute(
        "INSERT INTO woo_attributes(id,name,slug,woo_id) VALUES "
        "(6,'Edge type','pa_etype',6),(8,'Figure','pa_fig',8),"
        "(9,'Grade','pa_gr',9),(7,'Moisture','pa_mst',7)"
    )
    conn.execute(
        "INSERT INTO woo_attribute_terms(id,attribute_id,name,slug,woo_id) VALUES "
        "(321,6,'Live Edge 1 side','le1',321),"
        "(328,8,'Curl','curl',328),"
        "(342,9,'Fancy','fnc',342),"
        "(326,7,'Kiln dried','kd',326)"
    )
    conn.execute(
        "INSERT INTO woo_tags(id,name,slug,woo_id) VALUES (1,'feat-knot','feat-knot',1)"
    )
    conn.execute(
        "INSERT OR REPLACE INTO settings(key,value,value_enc) VALUES "
        "('brand_voice','Warm and precise.',NULL),"
        "('geo_context','Pacific Northwest mill.',NULL)"
    )
    conn.execute(
        "INSERT INTO slabs(id,sku,length_in,thickness_in,width_avg_in,width_min_in,width_max_in,bdft,"
        "species_id,wood_category_ids,edge_type_term_id,figure_term_ids,grade_term_id,moisture_term_id,"
        "feat_tag_ids,title,short_description,status,server_rev,client_rev,created_at,updated_at) "
        "VALUES ('s1','SLAB-01',114.0,1.75,24.3,18.0,24.3,18.4,11,?,321,?,342,326,?,"
        "'Figured Black Walnut Live Edge Slab','short line','ready',1,1,'t','t')",
        (json.dumps([22]), json.dumps([328]), json.dumps([1])),
    )
    png = _rgba_png(200, 200)
    conn.execute(
        "INSERT INTO slab_photos(id,slab_id,seq,processed_blob,created_at) VALUES ('p9','s1',1,?, 't')",
        (png,),
    )
    conn.commit()
    from backend.inference import assemble_call2, assemble_call2_text, CALL2_SEED

    assembled = assemble_call2(conn, "s1")
    assert assembled is not None
    text = assembled["text"]
    assert "DECLARED ATTRIBUTES" in text
    assert "species: Black Walnut" in text
    assert "primary wood category: Live edge wood" in text
    assert "edge: Live Edge 1 side" in text
    assert "figure: Curl" in text
    assert "grade: Fancy" in text
    assert "moisture: Kiln dried" in text
    assert "features: knot" in text
    assert "DIMENSION FACT BLOCK" in text
    assert "length=114.0in" in text
    assert "Brand voice: Warm and precise." in text
    assert "Geo: Pacific Northwest mill." in text
    assert CALL2_SEED.split("\n")[0] in text
    assert "SHORT DESCRIPTION" in text
    assert "LONG DESCRIPTION" in text
    assert "short_description" in CALL2_SEED
    assert '"description"' in CALL2_SEED or "\"description\"" in CALL2_SEED
    # Prior short must not be fed back into the prompt (Call 2 writes it).
    assert "short_description: short line" not in text
    assert "Write BOTH" in CALL2_SEED or "BOTH the short" in CALL2_SEED
    # No raw base64 in View-prompt shape
    assert "photos" in assembled
    assert assembled["photos"][0]["id"] == "p9"
    assert assembled["photos"][0]["seq"] == 1
    assert assembled["photos"][0]["bytes"] > 0
    assert "base64" not in json.dumps(assembled["photos"])
    # View prompt GET mirrors assembler (no model)
    r = client.get("/api/v1/slabs/s1/inference/call2/prompt")
    assert r.status_code == 200, r.text
    body = r.json()
    assert body["text"] == text
    assert body["photos"][0]["id"] == "p9"
    assert "data:image" not in json.dumps(body)


def test_call2_post_uses_same_assembler_text(env, monkeypatch):
    """POST call2 sends the same assembled text View prompt returns."""
    client, conn = env
    conn.execute("INSERT OR REPLACE INTO settings(key,value,value_enc) VALUES ('content_llm_enabled','true',NULL)")
    conn.execute(
        "INSERT OR REPLACE INTO settings(key,value,value_enc) VALUES "
        "('brand_voice','voice',NULL),('geo_context','geo',NULL)"
    )
    conn.execute(
        "INSERT INTO slabs(id,sku,length_in,thickness_in,width_avg_in,width_min_in,width_max_in,bdft,"
        "title,status,server_rev,client_rev,created_at,updated_at) "
        "VALUES ('s1','SLAB-01',100,2,20,18,22,18,'Walnut Slab','ready',1,1,'t','t')"
    )
    conn.execute(
        "INSERT INTO slab_photos(id,slab_id,seq,processed_blob,created_at) VALUES ('p1','s1',1,?, 't')",
        (_rgba_png(64, 64),),
    )
    conn.commit()
    from backend.inference import InferenceClient, assemble_call2

    expected = assemble_call2(conn, "s1")["text"]
    captured = {}

    def fake_chat(self, messages):
        captured["text"] = messages[0]["content"][-1]["text"]
        return "<p>ok</p>"

    monkeypatch.setattr(InferenceClient, "chat", fake_chat)
    r = client.post("/api/v1/slabs/s1/inference/call2")
    assert r.status_code == 200, r.text
    assert captured["text"] == expected
    assert "DECLARED ATTRIBUTES" in captured["text"]
    assert "DIMENSION FACT BLOCK" in captured["text"]