v88 功能增强：产品落盘持久化 + 生图网关适配 + 模板导出优化

- 产品持久化：每完成一个产品立即追加写入 products_pending.jsonl，崩溃不丢已完成产品，finish 读盘合并后统一写模板 - 503 致命错误提前终止：compose/product/seed_shot 端到端识别，提前终止搜索分析，丢弃未完成简报，保留已完成落盘产品直接合成模板 - 模特分配：material_library 合格模特图按任务序号独立随机，同 SPU 多款不再共用同一模特 - 图像网关适配：execution_mode/background 默认不再传入 yunfei 等标准网关，base_url 需带 /v1；429/5xx/空响应退避重试 - Pinterest 分析：删除 term 注入与纯文本降级，失败直接放弃；图片上传前 PIL 完整性校验；suitable_for_print=False 过滤丢弃 - 模板导出：不再产生空白 xlsx，文件名=模板原文件名_已填写；写入前按货号末 3 位升序排序 - 删除对接文档.md，更新 README，gitignore 排除测试产物
2026-08-28 10:28:35 +08:00
parent 685b7b0862
commit 2a96ec0870
28 changed files with 1187 additions and 729 deletions
@@ -180,14 +180,11 @@ class MockBackend:
        paths = list(image_paths or [])
        return [{
            "topic": term,
-            "concept": f"（启发式兜底）围绕「{term}」做原创{art_style}风格印花",
-            "motif": motif,
-            "art_style": art_style,
-            "color_palette": palette,
-            "composition": composition,
+            "suitable_for_print": True,
            "negative_prompt": negative,
            "image_prompt": (f"{motif}, {art_style}, {palette}, {composition}, "
-                             f"original {art_style} t-shirt print design"),
+                             f"original {art_style} t-shirt print design, "
+                             f"no brand logo, no trademark, no character, no watermark"),
            # 生图参考：每条简报对应其来源爬取图（mock 按图逐张产出简报，顺序一一对应）
            "ref_images": [str(paths[i])] if i < len(paths) else [],
            "source": "pinterest",
@@ -318,31 +318,56 @@ def build_pinterest_term_user_prompt(context: Dict[str, Any]) -> str:
 # —— Pinterest 参考模式：图片分析 → 原创设计简报（多模态）——
 # image_prompt 由 LLM 直接输出完整的英文生图提示词（多模态对图片的描述拼接），
 # 不再走「四要素 + 固定模板」装配；尺寸/白底等统一约束段由 prompt_node 自动追加。
-PINTEREST_ANALYZE_SYSTEM_PROMPT = """You are a POD (print-on-demand) T-shirt design analyst.
-You receive Pinterest reference images for one search term. For each image, extract the VISUAL CONCEPT
-(style, mood, motif, color palette, composition) that makes it appealing, then produce an ORIGINAL
-T-shirt print design brief that captures that VIBE WITHOUT copying the image.
+PINTEREST_ANALYZE_SYSTEM_PROMPT = """You are a POD T-shirt design analyst. Given ONE Pinterest reference image,
+judge whether it can inspire a T-shirt print, then write an ORIGINAL design brief
+capturing its vibe WITHOUT copying.
+Your image_prompt will be sent TOGETHER WITH this reference image to an image
+generator, so it must actively override visual imitation.

 RULES:
- NEVER copy the image, never reproduce the exact artwork, characters, logos, or any text from it.
- Extract only the abstract style/mood/motif concept as inspiration.
- Produce an original, flat, print-ready design brief (no garment, no model, no background scene).
- COPYRIGHT-SAFE: no brands, no logos, no characters, no celebrities, no real persons, no franchises.
- AVOID: politics, religion, hate, violence, sexual content, alcohol, national flags.
- motif: English, concrete central subject of the print (e.g. "a smiling cat with a fish", "geometric mountain layers").
- art_style: English visual technique (e.g. "clean flat vector", "retro screen print").
- color_palette: English colors (e.g. "sunset orange, cream, dusty blue").
- composition: English layout (e.g. "centered emblem with balanced negative space").
- concept: Chinese, one sentence describing the design idea.
- negative_prompt: what to avoid (real people, likeness, characters, logos, text).
- image_prompt: a COMPLETE, fluent English text-to-image prompt for generating the ORIGINAL flat print
-  design artwork (the print itself, NOT a garment photo). Describe the motif, art style, colors, layout
-  and mood in natural English, as a standalone print. Do NOT include garment / shirt / model / mannequin /
-  background-scene / watermark words. Do NOT mention any size or white-background suffix — a fixed
-  "small centered print on pure white" suffix will be appended automatically by the system.
+1. NO COPYING — never reproduce or closely imitate the reference's artwork,
+ characters, layout or text. Deliberately change motif, arrangement and/or
+ palette so the two read as clearly different works sharing only a general
+ style. Distill inspiration into generic style words (retro, y2k, minimal,
+ grunge, boho, kawaii...); never imitate an identifiable artist/studio/IP
+ style.
+2. FORBIDDEN — brand logos, trademarks, slogans, mascots, copyrighted
+ characters, real people/celebrities, movie/game/anime/band IP, lyrics,
+ even stylized or silhouette versions. Avoid politics, religion, violence,
+ sexual content, alcohol, drugs, gambling, flags, death/occult themes.
+3. FORM — ONE clear central subject with strong graphic composition;
+ print-ready standalone artwork. ANY colors are fine — rich palettes,
+ gradients and detailed shading are all acceptable. Photographic references
+ may be rendered as detailed full-color illustrations, retro badges or
+ vintage stickers.

-Return JSON with the field "designs" (array of objects with keys:
-motif, art_style, color_palette, composition, concept, negative_prompt, image_prompt)."""
+TEXT: short ORIGINAL English wording (1-6 words) allowed; wrap exact words in
+double quotes and demand exact spelling; integrate into composition. Never
+reuse/translate reference text; no brand/band/movie names or famous slogans.
+When unsure, omit.
+
+suitable_for_print: DEFAULT TRUE for graphics, illustrations, badges, vector
+art, typography posters, or prints on mockups (judge only the printed artwork).
+FALSE only for: subjectless photo scenery, memes/screenshots/collages,
+watermarked or very low-quality images, decor/food/candid photos with no
+usable motif. Even when FALSE, still fill all fields so downstream never breaks.
+
+image_prompt = two parts:
+ 1) mandatory opener, e.g.: "Use the attached reference image only as loose
+ inspiration for overall mood, theme and era — do NOT reproduce, trace,
+ rearrange, recolor or closely imitate any element, character, layout or
+ text shown in it."
+ 2) the new design: [central motif] + [style] + [color treatment] +
+ [composition] + [mood], plus quoted original text if used.
+ NEVER mention shirts, apparel, models, scenes, sizes, backgrounds or
+ watermarks — placement is handled externally.
+
+negative_prompt: copy of reference artwork, likenesses, characters, logos,
+trademarks, watermark, photorealistic shirt/product mockups, busy background;
+add garbled-lettering terms only if your design includes text.
+
+OUTPUT — ONLY valid JSON, no fences:
+{"designs":[{"suitable_for_print":<bool>,"negative_prompt":"<str>","image_prompt":"<str>"}]}"""

 PINTEREST_ANALYZE_SCHEMA = {
    "name": "pinterest_design_briefs",
@@ -354,17 +379,11 @@ PINTEREST_ANALYZE_SCHEMA = {
                "items": {
                    "type": "object",
                    "properties": {
-                        "image_index": {"type": "integer"},
-                        "motif": {"type": "string"},
-                        "art_style": {"type": "string"},
-                        "color_palette": {"type": "string"},
-                        "composition": {"type": "string"},
-                        "concept": {"type": "string"},
+                        "suitable_for_print": {"type": "boolean"},
                        "negative_prompt": {"type": "string"},
                        "image_prompt": {"type": "string"},
                    },
-                    "required": ["image_index", "motif", "art_style", "color_palette",
-                                 "composition", "concept", "negative_prompt", "image_prompt"],
+                    "required": ["suitable_for_print", "negative_prompt", "image_prompt"],
                    "additionalProperties": False,
                },
            }
@@ -375,18 +394,10 @@ PINTEREST_ANALYZE_SCHEMA = {
 }


-def build_pinterest_analyze_user_prompt(term: str, country: str, image_count: int) -> str:
+def build_pinterest_analyze_user_prompt() -> str:
    return (
-        f"Country: {country}\n"
-        f"Pinterest search term: {term}\n"
-        f"Reference images attached: {image_count} images.\n\n"
-        f"Analyze the attached images and produce {image_count} ORIGINAL design briefs "
-        f"(one per image), each capturing the visual vibe as an original T-shirt print design. "
-        f"Do NOT copy the images.\n"
-        f"For EACH brief you MUST set image_index to the 0-based position of the input image "
-        f"it was derived from (first image = 0, second = 1, ...). Every image_index from 0 to "
-        f"{max(image_count - 1, 0)} must appear exactly once — this links each brief to its "
-        f"source image so the design is generated from the SAME image that was analyzed."
+        "Analyze the attached image and produce one ORIGINAL T-shirt print design brief "
+        "that captures its visual vibe without copying it."
    )


@@ -587,8 +598,7 @@ class OpenAICompatBackend(LLMBackend):
                                 on_400=None) -> List[Dict[str, Any]]:
        """多模态分析 Pinterest 图片 → 原创设计简报列表。

-        图片输入不被模型支持（纯文本模型 400）时自动降级为纯文本分析（仅用搜索词）。
-        失败返回 []，由节点兜底（回退 mock 规则简报）。
+        图片输入失败/无有效图片时直接放弃（返回 []，不降级纯文本），由节点跳过该产品。
        on_400: 每次 HTTP 400（且含「内容/图片」）时回调（供调用方累计放弃计数）。
        """
        cfg = self._cfg
@@ -605,11 +615,26 @@ class OpenAICompatBackend(LLMBackend):
        data_uris: List[str] = []
        for p in image_paths:
            try:
+                pk = Path(p)
+                raw = pk.read_bytes()
+                # 校验图片完整性：损坏/截断的图片会被火山方舟等多模态接口直接 400 拒绝，
+                # 必须滤掉后才能编码 base64（PIL 打开失败即视为损坏）。
+                if not raw or len(raw) < 100:
+                    print(f"[pinterest_analyze] 图片文件过小/为空，跳过: {p} ({len(raw)}B)")
+                    continue
+                try:
+                    from PIL import Image
+                    _im = Image.open(pk)
+                    _im.verify()   # 校验文件头/结构，不完整则抛异常
+                    _im.close()
+                except Exception as _ve:  # noqa: BLE001
+                    print(f"[pinterest_analyze] 图片损坏/不完整，跳过: {p} ({_ve})")
+                    continue
                import base64 as b64
                mime = "image/png"
-                if Path(p).suffix.lower() in (".jpg", ".jpeg"):
+                if pk.suffix.lower() in (".jpg", ".jpeg"):
                    mime = "image/jpeg"
-                data_uris.append(f"data:{mime};base64,{b64.b64encode(Path(p).read_bytes()).decode()}")
+                data_uris.append(f"data:{mime};base64,{b64.b64encode(raw).decode()}")
            except Exception as e:  # noqa: BLE001
                print(f"[pinterest_analyze] 图片读取失败 {p}: {e}")

@@ -623,12 +648,11 @@ class OpenAICompatBackend(LLMBackend):
            except Exception:  # noqa: BLE001
                pass

-        def _call(use_images: bool) -> str:
+        def _call() -> str:
            user_content: List[Any] = [
-                {"type": "text", "text": build_pinterest_analyze_user_prompt(term, country, len(data_uris))},
+                {"type": "text", "text": build_pinterest_analyze_user_prompt()},
            ]
-            if use_images:
-                user_content += [{"type": "image_url", "image_url": {"url": u}} for u in data_uris]
+            user_content += [{"type": "image_url", "image_url": {"url": u}} for u in data_uris]
            payload = {
                "model": model,
                "messages": [
@@ -656,19 +680,15 @@ class OpenAICompatBackend(LLMBackend):
                resp.raise_for_status()
                return str(resp.json()["choices"][0]["message"].get("content") or "")

-        raw = ""
-        if data_uris:
-            try:
-                raw = _call(use_images=True)
-            except Exception as e:  # noqa: BLE001  纯文本模型不支持图片 → 降级纯文本
-                print(f"[pinterest_analyze] 图片输入失败，降级纯文本分析: {e}")
-                raw = ""
-        if not raw:
-            try:
-                raw = _call(use_images=False)
-            except Exception as e:  # noqa: BLE001
-                print(f"[pinterest_analyze] 分析失败: {e}")
-                return []
+        # 图片输入失败/无有效图片 → 直接放弃该产品（不降级纯文本），由节点跳过后续流程
+        if not data_uris:
+            print("[pinterest_analyze] 无有效图片输入，放弃该产品（不降级纯文本）")
+            return []
+        try:
+            raw = _call()
+        except Exception as e:  # noqa: BLE001
+            print(f"[pinterest_analyze] 图片分析失败，放弃该产品（不降级纯文本）: {e}")
+            return []
        try:
            parsed = _extract_json(raw)
        except Exception as e:  # noqa: BLE001
@@ -687,11 +707,7 @@ class OpenAICompatBackend(LLMBackend):
                continue
            designs.append({
                "topic": term,
-                "concept": str(d.get("concept", "")).strip(),
-                "motif": str(d.get("motif", "")).strip(),
-                "art_style": str(d.get("art_style", "")).strip(),
-                "color_palette": str(d.get("color_palette", "")).strip(),
-                "composition": str(d.get("composition", "")).strip(),
+                "suitable_for_print": bool(d.get("suitable_for_print", True)),
                "negative_prompt": str(d.get("negative_prompt", "")).strip(),
                "image_prompt": str(d.get("image_prompt", "")).strip(),
                # 生图参考：每条简报对应其来源爬取图（LLM 按图逐张产出简报，顺序一一对应）