fix(接口4): worker 移除脸型 Mock + face_shape 改本机 MediaPipe 计算

- worker /api/v1/face/features 不再返回假成功数据,直接告知仅网关实现,
  避免本机误打 :8187 被 Mock 结果误导。
- 网关 ark_api_key 加载优先级改为 gateway/config.json 优先(原先误读
  worker_config.json 里的失效 key)。
- 接口4 face_shape 不再采信豆包结果,改用本机 face/face_shape_classifier.py
  (MediaPipe 7 类)计算覆盖;其余 5 项特征仍走豆包。
- 修复 face_shape_classifier 共享 FaceMesh 实例的线程安全问题(加锁),
  避免网关侧接口4 并发请求时崩溃/结果错乱。
- 新增 /api/v1/debug/face-shape 调试接口 + static/test_face_shape.html
  单图调试页(worker 侧)。
- 更新文档:网关机现在也需要 mediapipe/opencv-python/numpy<2。

⚠️ 部署前提醒:网关机需先安装 mediapipe==0.10.14 / opencv-python==4.10.0.84 /
numpy==1.26.4,否则接口4 会返回 1007「分析服务异常」。

Co-authored-by: Cursor <cursoragent@cursor.com>
This commit is contained in:
xsl
2026-07-29 14:41:49 +08:00
co-authored by Cursor
parent a748dfd1e5
commit 9fb5b486c0
8 changed files with 446 additions and 72 deletions
+2 -1
View File
@@ -82,7 +82,8 @@ model.safetensors https://huggingface.co/jonathandinu/face-parsing/resol
模型已就位,但**内网机还需要 Python 依赖的离线 wheel 包**,否则 `pip install` 在内网无法联网安装。这部分**与目标机的操作系统、Python 版本、CUDA 版本强相关**,需确认后单独打包: 模型已就位,但**内网机还需要 Python 依赖的离线 wheel 包**,否则 `pip install` 在内网无法联网安装。这部分**与目标机的操作系统、Python 版本、CUDA 版本强相关**,需确认后单独打包:
- **workerGPU 机)**`mediapipe` / `opencv-python` / `numpy<2` / `Pillow` / **`torch`+`torchvision` 的 CUDA 版**(按 GPU 的 CUDA 版本选 cu118/cu121 等)/ `transformers`(接口2 SegFormer+ FastAPI/uvicorn 全家桶。 - **workerGPU 机)**`mediapipe` / `opencv-python` / `numpy<2` / `Pillow` / **`torch`+`torchvision` 的 CUDA 版**(按 GPU 的 CUDA 版本选 cu118/cu121 等)/ `transformers`(接口2 SegFormer+ FastAPI/uvicorn 全家桶。
- **网关机**很轻,只需 FastAPI/uvicorn/httpx 等代理依赖**不需要 torch/mediapipe**。 - **网关机**FastAPI/uvicorn/httpx 等代理依赖 + **接口4 现需 `mediapipe`/`opencv-python`/`numpy<2`**(脸型本机计算),
仍**不需要 torch**(无 GPU 推理需求)。
> 架构已拆分(见 `docs/实现说明.md`):算法依赖只装在 worker,网关保持轻量。 > 架构已拆分(见 `docs/实现说明.md`):算法依赖只装在 worker,网关保持轻量。
+93 -47
View File
@@ -1,6 +1,6 @@
"""旷视五接口 — worker 侧(高性能 GPU 后端)。 """旷视五接口 — worker 侧(高性能 GPU 后端)。
接口 1(四庭七眼测量)已为**真实算法实现**(见 face_analysis 包);接口 2~5 仍为 Mock。 接口 1 等算法在 worker;接口4(用户特征)已迁到网关本机(调豆包),worker 同路径只返回明确错误、无 Mock。
拆分架构:worker 跑算法、返回 `annotated_image_base64`(不落盘、不拼 URL,由网关完成)。 拆分架构:worker 跑算法、返回 `annotated_image_base64`(不落盘、不拼 URL,由网关完成)。
worker 对 `/api/*` 校验内网鉴权头 `X-Internal-Token``/health` 供网关探测不校验。 worker 对 `/api/*` 校验内网鉴权头 `X-Internal-Token``/health` 供网关探测不校验。
""" """
@@ -885,56 +885,28 @@ async def hair_grow_b(
@app.post( @app.post(
"/api/v1/face/features", "/api/v1/face/features",
summary="接口4 用户特征分析", summary="接口4 用户特征分析(仅网关)",
tags=["人脸分析"], tags=["人脸分析"],
description=f""" description="""
输入用户照片,返回 N 个用户面部特征字段。 **本接口不在 worker 实现。** 请调用网关(本机默认 `http://127.0.0.1:8080`)的同路径;
网关本机调火山方舟豆包视觉模型,不转发到 worker。
{_image_fields_desc} 直接打 worker(如 `:8187`)会返回错误,避免误用假数据。
图片同时支持 `multipart/form-data` 文件上传(字段名 `image_file`)。
---
由**火山方舟 豆包视觉模型**分析,返回**固定 6 个英文字段**。
**返回格式**`data.features` 为一个 **JSON 字符串**(不是对象),需要在客户端 `JSON.parse()` 后使用。
| 字段 | 说明 |
|------|------|
| face_shape | 脸形(如"鹅蛋脸" |
| eyebrow_shape | 眉形(如"平眉" |
| facial_age | 面部年龄(区间,如"18-25岁" |
| dynamic_static_type | 动静类型("静态型"/"动态型" |
| gender | 性别(""/"" |
| gene_style | 基因风格(如"自然型" |
> 无人脸返回 `1001`。
""", """,
responses={ responses={
200: { 200: {
"description": "成功", "description": "worker 不提供本接口",
"content": { "content": {
"application/json": { "application/json": {
"example": { "example": {
"code": 0, "code": 1007,
"message": "success", "message": "接口4 仅在网关实现,请访问网关(本机默认 :8080),worker 不提供本接口",
"request_id": "mock-request-id", "request_id": "mock-request-id",
"data": { "data": None,
"features": '{"face_shape":"鹅蛋脸","eyebrow_shape":"平眉","facial_age":"18-25岁","dynamic_static_type":"静态型","gender":"","gene_style":"少年型"}',
},
} }
} }
}, },
}, },
400: {
"description": "参数错误 / 图片识别失败",
"content": {
"application/json": {
"example": {"code": 1001, "message": "无法识别人像", "request_id": "x", "data": None}
}
},
},
}, },
) )
async def face_features( async def face_features(
@@ -942,16 +914,12 @@ async def face_features(
image_url: Optional[str] = Form(default=None, description="图片 URL"), image_url: Optional[str] = Form(default=None, description="图片 URL"),
image_base64: Optional[str] = Form(default=None, description="图片 base64(需带 data:image/...;base64, 前缀)"), image_base64: Optional[str] = Form(default=None, description="图片 base64(需带 data:image/...;base64, 前缀)"),
): ):
# ⚠️ 接口4 已迁到**网关本机**实现(直接调豆包视觉模型,见 gateway/app.py)。 # 接口4 只在网关实现(gateway/app.py → face_features.analyze_features)。
# 网关不会把本接口转发到 worker,故此处仅留 Mock 占位、保持 worker 无外网依赖 # 不再返回 Mock 成功数据,避免本机打 :8187 时被假结果误导
features = json.dumps( return err(
{ 1007,
"face_shape": "鹅蛋脸", "eyebrow_shape": "平眉", "facial_age": "18-25岁", "接口4 仅在网关实现,请访问网关(本机默认 :8080),worker 不提供本接口",
"dynamic_static_type": "静态型", "gender": "", "gene_style": "少年型",
},
ensure_ascii=False,
) )
return ok({"features": features})
# --------------------------------------------------------------------------- # ---------------------------------------------------------------------------
@@ -1662,6 +1630,84 @@ async def download_hairline_log(rid: Optional[str] = None, tail: int = 500):
return PlainTextResponse("".join(lines), media_type="text/plain; charset=utf-8") return PlainTextResponse("".join(lines), media_type="text/plain; charset=utf-8")
# ---------------------------------------------------------------------------
# 调试:MediaPipe 脸型分类(face/face_shape_classifier.py,非接口4 豆包)
# ---------------------------------------------------------------------------
@app.post(
"/api/v1/debug/face-shape",
summary="调试 单张脸型分类(MediaPipe)",
tags=["调试"],
description="""
离线脸型分类调试接口(`face/face_shape_classifier.py`),**不是**接口4 的豆包视觉分析。
上传正面照 → MediaPipe 468 点 → 7 类脸型评分 + 特征标注图。
""",
include_in_schema=True,
)
async def debug_face_shape(
image_file: Optional[UploadFile] = File(default=None, description="上传图片文件(JPG/PNG"),
image_url: Optional[str] = Form(default=None, description="图片 URL"),
image_base64: Optional[str] = Form(default=None, description="图片 base64"),
):
raw, e = await resolve_image_bytes(image_file, image_url, image_base64)
if e:
return e
try:
nparr = np.frombuffer(raw, np.uint8)
bgr = cv2.imdecode(nparr, cv2.IMREAD_COLOR)
if bgr is None:
return err(1008, "图片格式不支持(仅 JPG / PNG)")
except Exception: # noqa: BLE001
return err(1008, "图片格式不支持(仅 JPG / PNG)")
def _jsonable(obj):
if isinstance(obj, dict):
return {k: _jsonable(v) for k, v in obj.items()}
if isinstance(obj, (list, tuple)):
return [_jsonable(v) for v in obj]
if hasattr(obj, "item"):
return obj.item()
if isinstance(obj, (float, int, str, bool)) or obj is None:
return obj
return obj
from fastapi.concurrency import run_in_threadpool
try:
from face.face_shape_classifier import classify_from_image
result = await run_in_threadpool(
classify_from_image, bgr, True, True,
)
except ValueError as ex:
return err(1001, str(ex) or "无法识别人像")
except Exception as ex: # noqa: BLE001
return err(1007, f"脸型分类失败:{ex}")
details = result.get("details") or {}
ranked = [
{"shape": name, "score": round(float(score), 2)}
for name, score in (details.get("ranked") or [])
]
annotated = result.pop("annotated", None)
h, w = bgr.shape[:2]
data = {
"face_shape": result["face_shape"],
"display": result["display"],
"confidence": round(float(result["confidence"]), 4),
"is_mixed": bool(details.get("is_mixed")),
"second_shape": details.get("second_shape"),
"score_gap": round(float(details["score_gap"]), 2) if details.get("score_gap") is not None else None,
"ranked": ranked,
"features": _jsonable(result.get("features") or {}),
"zscores": _jsonable(details.get("zscores") or {}),
"image_size": {"width": w, "height": h},
"annotated_image_base64": _jpg_b64(annotated) if annotated is not None else None,
}
return ok(data)
# --------------------------------------------------------------------------- # ---------------------------------------------------------------------------
# 健康检查 # 健康检查
# --------------------------------------------------------------------------- # ---------------------------------------------------------------------------
+8 -4
View File
@@ -85,8 +85,10 @@
### 接口4 用户特征 `/api/v1/face/features`**网关本机** ### 接口4 用户特征 `/api/v1/face/features`**网关本机**
- **做什么**:照片 → 几十项面部特征(脸型/眉形/肤色/三庭五眼/四季色彩季型/量感/基因风格/性别…)。`data.features` 是 JSON 字符串。 - **做什么**:照片 → 几十项面部特征(脸型/眉形/肤色/三庭五眼/四季色彩季型/量感/基因风格/性别…)。`data.features` 是 JSON 字符串。
- **怎么实现**`gateway/`,逻辑参考 worker `face_features.py` / `/home/xsl/fuyan`):调**火山方舟 豆包视觉模型** - **怎么实现**`gateway/`,逻辑参考 worker `face_features.py` / `/home/xsl/fuyan`):调**火山方舟 豆包视觉模型**
`doubao-seed-1-6-vision`(OpenAI 兼容,base64 data URI 喂图),解析 JSON + 映射 6 个英文优先字段并保留全部中文 `doubao-seed-1-6-vision`(OpenAI 兼容,base64 data URI 喂图),解析眉形/年龄/动静/性别/基因风格 5 项
无人脸→1001。**唯一调外网的接口**:网关需可达 `ark.cn-beijing.volces.com`API Key 走网关配置(不入 git) **`face_shape`(脸型)改为本机 `face/face_shape_classifier.py`(MediaPipe)计算并覆盖豆包结果**
无人脸→1001。网关需可达 `ark.cn-beijing.volces.com`API Key 走网关配置(不入 git)。
⚠️ 因此**网关机不再是纯轻量代理**,需额外安装 `mediapipe`/`opencv-python`/`numpy<2`(见 `requirements.txt`)。
### 接口5 发际线PNG生成 `/api/v1/hairline/generate`worker ### 接口5 发际线PNG生成 `/api/v1/hairline/generate`worker
- **做什么**:入参同接口2`gender` + 多选 `hair_style` 必填)。对每个选中发型 → `middle`/`high`/`low` 三档发际线叠图 + 生发图 + 首个选中发型的面部中间点坐标。 - **做什么**:入参同接口2`gender` + 多选 `hair_style` 必填)。对每个选中发型 → `middle`/`high`/`low` 三档发际线叠图 + 生发图 + 首个选中发型的面部中间点坐标。
@@ -108,8 +110,10 @@
- `worker_config.json`(不入 git)`accept_passwords`(鉴权) + 鉴权头 `X-Internal-Token` - `worker_config.json`(不入 git)`accept_passwords`(鉴权) + 鉴权头 `X-Internal-Token`
**网关机** **网关机**
- 很轻:FastAPI/uvicorn/httpx + **接口4 的 `volcengine-python-sdk[ark]`**(或直接 httpx 调,OpenAI 兼容)。 - FastAPI/uvicorn/httpx + **接口4 的 `volcengine-python-sdk[ark]`**(或直接 httpx 调,OpenAI 兼容)。
- 不装 torch/mediapipe/opencv。配置 `gateway/config.json`(不入 git)`workers` 列表、`shared_password` - ⚠️ 接口4 `face_shape` 改本机 MediaPipe 计算后,网关机**也需要装** `mediapipe`/`opencv-python`/`numpy<2`
(不再是"网关不装 torch/mediapipe/opencv",只是仍不需要 torch/transformers/scikit-image 等重依赖)。
- 配置 `gateway/config.json`(不入 git)`workers` 列表、`shared_password`
`ark` 的 api_key/base_url/model、`public_base_url`、超时(**生发接口慢,`request_timeout_seconds` 调大 ≥120s**)。 `ark` 的 api_key/base_url/model、`public_base_url`、超时(**生发接口慢,`request_timeout_seconds` 调大 ≥120s**)。
- 托管 `/static/annotations/`(落盘的图),定期清理。 - 托管 `/static/annotations/`(落盘的图),定期清理。
+8 -2
View File
@@ -14,6 +14,7 @@ face_shape_classifier.py
from __future__ import annotations from __future__ import annotations
import math import math
import threading
from pathlib import Path from pathlib import Path
from typing import Dict, List, Optional, Tuple, Union from typing import Dict, List, Optional, Tuple, Union
@@ -343,6 +344,9 @@ def get_mixed_description(details: Dict) -> str:
_face_mesh = None _face_mesh = None
# mediapipe Solutions API 的单个 FaceMesh 实例不是线程安全的;被 web 服务用
# run_in_threadpool 并发调用时(如网关接口4、worker 调试接口)必须加锁串行化。
_face_mesh_lock = threading.Lock()
def _get_face_mesh(): def _get_face_mesh():
@@ -502,7 +506,8 @@ def annotate_face_features(
if landmarks is None: if landmarks is None:
rgb = cv2.cvtColor(bgr, cv2.COLOR_BGR2RGB) rgb = cv2.cvtColor(bgr, cv2.COLOR_BGR2RGB)
results = _get_face_mesh().process(rgb) with _face_mesh_lock:
results = _get_face_mesh().process(rgb)
if not results.multi_face_landmarks: if not results.multi_face_landmarks:
raise ValueError("未检测到人脸关键点") raise ValueError("未检测到人脸关键点")
landmarks = results.multi_face_landmarks[0].landmark landmarks = results.multi_face_landmarks[0].landmark
@@ -741,7 +746,8 @@ def classify_from_image(
bgr = _load_image(image) bgr = _load_image(image)
h, w = bgr.shape[:2] h, w = bgr.shape[:2]
rgb = cv2.cvtColor(bgr, cv2.COLOR_BGR2RGB) rgb = cv2.cvtColor(bgr, cv2.COLOR_BGR2RGB)
results = _get_face_mesh().process(rgb) with _face_mesh_lock:
results = _get_face_mesh().process(rgb)
if not results.multi_face_landmarks: if not results.multi_face_landmarks:
raise ValueError("未检测到人脸关键点") raise ValueError("未检测到人脸关键点")
+63 -16
View File
@@ -1,10 +1,11 @@
"""接口4:用户面部特征分析(调用火山方舟 豆包视觉模型 doubao-seed-1-6-vision """接口4:用户面部特征分析。
算法来源:/home/xsl/fuyanFaceArk.py)。worker 把图片以 base64 data URI 传给方舟 - 眉形 / 面部年龄 / 动静类型 / 性别 / 基因风格:火山方舟豆包视觉模型
多模态模型,模型返回一大堆人脸特征 JSON;本模块解析后映射出接口4 的英文优先字段 - face_shape(脸型):本机 MediaPipe 分类(face/face_shape_classifier.py)覆盖,
face_shape 等),并保留 doubao 返回的全部中文字段。 不用豆包结果
⚠️ 这是**唯一调外网云模型**的接口(其余接口全本地)。API Key 走配置/环境变量,不入 git。 ⚠️ 仍依赖外网豆包(其余 5 项)。API Key 走配置/环境变量,不入 git。
网关本机需可 import face 包(opencv + mediapipe)。
""" """
from __future__ import annotations from __future__ import annotations
@@ -18,9 +19,8 @@ logger = logging.getLogger("hair.worker")
ARK_BASE_URL = os.getenv("ARK_BASE_URL", "https://ark.cn-beijing.volces.com/api/v3") ARK_BASE_URL = os.getenv("ARK_BASE_URL", "https://ark.cn-beijing.volces.com/api/v3")
ARK_MODEL = os.getenv("ARK_MODEL", "doubao-seed-1-6-vision-250815") ARK_MODEL = os.getenv("ARK_MODEL", "doubao-seed-1-6-vision-250815")
# doubao 中文键 → 接口4 英文优先字段(仅保留这 6 项 # doubao 中文键 → 接口4 英文字段(脸型不走豆包,见 _local_face_shape
_KEY_MAP = { _KEY_MAP = {
"脸型": "face_shape",
"眉形": "eyebrow_shape", "眉形": "eyebrow_shape",
"面部年龄": "facial_age", "面部年龄": "facial_age",
"动静类型": "dynamic_static_type", "动静类型": "dynamic_static_type",
@@ -28,11 +28,11 @@ _KEY_MAP = {
"基因风格": "gene_style", "基因风格": "gene_style",
} }
# 仅请求接口4 需要的 6 个字段(+「图片是否有人脸」用于 1001 判定,不进最终输出) # 豆包只问 5 项 + 是否有人脸;脸型由本地分类器给出
_PROMPT = ( _PROMPT = (
"分析一下图片告诉我以下特征,只要答案,格式为json字符串," "分析一下图片告诉我以下特征,只要答案,格式为json字符串,"
"图片是否有人脸(有人/没人) " "图片是否有人脸(有人/没人) "
"脸型(圆形脸/心形脸/菱形脸/鹅蛋脸/方形脸/长形脸/瓜子脸) 眉形 " "眉形 "
"面部年龄(给出区间年龄) 动静类型(静态型/动态型) 性别(男/女) " "面部年龄(给出区间年龄) 动静类型(静态型/动态型) 性别(男/女) "
"基因风格(戏剧型/睿智型/自然型/古典型/优雅型/浪漫型/前卫型/少女型/少年型)" "基因风格(戏剧型/睿智型/自然型/古典型/优雅型/浪漫型/前卫型/少女型/少年型)"
) )
@@ -42,12 +42,13 @@ _client_key: str | None = None # _client 构建时使用的 api_key,用
def _load_api_key() -> str | None: def _load_api_key() -> str | None:
"""ARK_API_KEY 环境变量优先否则读 worker_config.json / gateway/config.json 的 ark_api_key。""" """ARK_API_KEY 环境变量优先否则优先 gateway/config.json(接口4 已迁网关),
再回退 worker_config.json(兼容旧配置)。"""
key = os.getenv("ARK_API_KEY") key = os.getenv("ARK_API_KEY")
if key: if key:
return key return key
base = os.path.dirname(__file__) base = os.path.dirname(__file__)
for cfg_name in ("worker_config.json", "gateway/config.json"): for cfg_name in ("gateway/config.json", "worker_config.json"):
cfg = os.path.join(base, cfg_name) cfg = os.path.join(base, cfg_name)
if os.path.isfile(cfg): if os.path.isfile(cfg):
try: try:
@@ -97,10 +98,47 @@ def _image_to_url(image_bytes: bytes = None, image_url: str = None) -> str:
return f"data:image/{fmt};base64," + base64.b64encode(image_bytes).decode() return f"data:image/{fmt};base64," + base64.b64encode(image_bytes).decode()
def analyze_features(image_bytes: bytes = None, image_url: str = None): def _resolve_image_bytes(image_bytes: bytes = None, image_url: str = None) -> bytes:
"""调 doubao 视觉模型分析人脸特征。 """本地分类器用:优先已有字节;仅有 URL 时下载。"""
if image_bytes:
return image_bytes
if not image_url:
raise ValueError("缺少图片数据")
if image_url.startswith("data:"):
# data URI
b64 = image_url.split(",", 1)[1] if "," in image_url else image_url
return base64.b64decode(b64)
import httpx
with httpx.Client(timeout=15.0, follow_redirects=True) as client:
r = client.get(image_url)
r.raise_for_status()
return r.content
Returns: dict —— 仅含接口4 的 6 个英文字段(face_shape/eyebrow_shape/facial_age/
def _local_face_shape(image_bytes: bytes = None, image_url: str = None) -> str:
"""MediaPipe 脸型分类,返回 display(含混合脸型描述)或主脸型。"""
import cv2
import numpy as np
from face.face_shape_classifier import classify_from_image
raw = _resolve_image_bytes(image_bytes, image_url)
bgr = cv2.imdecode(np.frombuffer(raw, np.uint8), cv2.IMREAD_COLOR)
if bgr is None:
raise ValueError("图片格式不支持,无法解码")
result = classify_from_image(bgr, return_details=True, return_annotated=False)
shape = result.get("display") or result["face_shape"]
logger.info(
"local face_shape=%s conf=%.3f",
shape,
float(result.get("confidence") or 0),
)
return shape
def analyze_features(image_bytes: bytes = None, image_url: str = None):
"""豆包分析 5 项特征 + 本机 MediaPipe 覆盖 face_shape。
Returns: dict —— 6 个英文字段(face_shape/eyebrow_shape/facial_age/
dynamic_static_type/gender/gene_style)**无人脸返回 None**(调用方据此判 1001)。 dynamic_static_type/gender/gene_style)**无人脸返回 None**(调用方据此判 1001)。
""" """
url = _image_to_url(image_bytes, image_url) url = _image_to_url(image_bytes, image_url)
@@ -131,8 +169,17 @@ def analyze_features(image_bytes: bytes = None, image_url: str = None):
raise RuntimeError(f"豆包模型返回格式异常,无法解析为 JSON:{text[:200]}") from e raise RuntimeError(f"豆包模型返回格式异常,无法解析为 JSON:{text[:200]}") from e
if not has_face(raw): if not has_face(raw):
return None return None
# 只保留 6 个英文字段(doubao 缺某字段则跳过) # 豆包 5 项 + 本地脸型覆盖
return {en: raw[zh] for zh, en in _KEY_MAP.items() if zh in raw} feats = {en: raw[zh] for zh, en in _KEY_MAP.items() if zh in raw}
try:
feats["face_shape"] = _local_face_shape(image_bytes, image_url)
except ValueError as e:
logger.warning("本地脸型分类未检测到人脸: %s", e)
return None
except Exception as e: # noqa: BLE001
logger.exception("本地脸型分类失败")
raise RuntimeError(f"本地脸型分类失败:{e}") from e
return feats
def has_face(features: dict) -> bool: def has_face(features: dict) -> bool:
+1 -1
View File
@@ -552,7 +552,7 @@ async def face_features(
image_url: Optional[str] = Form(default=None, description="图片 URL"), image_url: Optional[str] = Form(default=None, description="图片 URL"),
image_base64: Optional[str] = Form(default=None, description="图片 base64(需带前缀)"), image_base64: Optional[str] = Form(default=None, description="图片 base64(需带前缀)"),
): ):
"""接口4:用户特征分析 — 本机直接调豆包视觉模型,不经过 worker。""" """接口4:用户特征分析 — 豆包 5 项 + 本机 MediaPipe 脸型(face/,不经过 worker。"""
import uuid as _uuid import uuid as _uuid
# 三选一校验 # 三选一校验
+4 -1
View File
@@ -26,8 +26,11 @@ transformers==4.45.2 # SegFormer 人脸分割(jonathandinu/face-parsing
# ⚠️ 必须 0.24.x —— 0.25+ 强依赖 numpy>=2,会顶掉 mediapipe 需要的 numpy<2 # ⚠️ 必须 0.24.x —— 0.25+ 强依赖 numpy>=2,会顶掉 mediapipe 需要的 numpy<2
scikit-image==0.24.0 # route_through_array(黑帽响应图上的 Dijkstra 最小路径) scikit-image==0.24.0 # route_through_array(黑帽响应图上的 Dijkstra 最小路径)
# 接口4:用户特征(火山方舟 豆包视觉模型)—— 已迁到**网关**实现,worker 不需要。 # 接口4:用户特征(火山方舟 豆包视觉模型 + 本机 MediaPipe 脸型覆盖)—— 已迁到**网关**实现,worker 不需要。
# 网关机装:volcengine-python-sdk[ark]from volcenginesdkarkruntime import Ark);API Key 走配置不入 git # 网关机装:volcengine-python-sdk[ark]from volcenginesdkarkruntime import Ark);API Key 走配置不入 git
# ⚠️ face_shape 字段改为本机 face/face_shape_classifier.py 计算(覆盖豆包结果),
# 因此网关机现在也需要 mediapipe==0.10.14 + opencv-python==4.10.0.84 + numpy==1.26.4
# (不再是"网关不需要 mediapipe",见 docs/实现说明.md 需同步更新)
# 测试 # 测试
pytest==8.3.3 pytest==8.3.3
+267
View File
@@ -0,0 +1,267 @@
<!DOCTYPE html>
<html lang="zh-CN">
<head>
<meta charset="UTF-8">
<meta name="viewport" content="width=device-width, initial-scale=1.0">
<title>脸型分类调试 — MediaPipe</title>
<style>
* { box-sizing: border-box; margin: 0; padding: 0; }
body { font-family: -apple-system, BlinkMacSystemFont, "Segoe UI", Roboto, sans-serif; background: #f5f5f5; color: #333; }
.container { max-width: 1200px; margin: 0 auto; padding: 24px; }
h1 { font-size: 22px; margin-bottom: 6px; }
.subtitle { color: #888; font-size: 13px; margin-bottom: 24px; }
.subtitle code { background: #eef2ff; color: #3730a3; padding: 1px 6px; border-radius: 4px; font-size: 12px; }
.card { background: #fff; border-radius: 12px; box-shadow: 0 1px 4px rgba(0,0,0,.06); margin-bottom: 20px; }
.card-header { font-weight: 700; font-size: 14px; padding: 14px 18px; border-bottom: 1px solid #f0f0f0; background: #fafafa; display: flex; justify-content: space-between; align-items: center; }
.card-body { padding: 18px; }
.upload-row { display: flex; gap: 12px; align-items: center; flex-wrap: wrap; }
.file-input { flex: 1; min-width: 200px; }
.file-input input[type=file] { width: 100%; padding: 8px; border: 2px dashed #ddd; border-radius: 8px; cursor: pointer; }
.btn { padding: 10px 28px; border: none; border-radius: 8px; font-size: 15px; cursor: pointer; font-weight: 600; transition: .2s; }
.btn-primary { background: #2563eb; color: #fff; }
.btn-primary:hover { background: #1d4ed8; }
.btn-primary:disabled { background: #93c5fd; cursor: not-allowed; }
.btn-sm { padding: 6px 14px; font-size: 13px; }
.btn-outline { background: #fff; border: 1px solid #d1d5db; color: #374151; }
.btn-outline:hover { background: #f9fafb; }
.hint { font-size: 12px; color: #9ca3af; margin-top: 8px; }
.status { padding: 10px 16px; border-radius: 8px; font-size: 14px; margin-bottom: 16px; display: none; }
.status.info { background: #dbeafe; color: #1e40af; display: block; }
.status.error { background: #fee2e2; color: #991b1b; display: block; }
.status.success { background: #d1fae5; color: #065f46; display: block; }
.results-layout { display: flex; gap: 24px; }
.col-main { flex: 1.4; min-width: 0; }
.col-side { flex: 1; min-width: 0; }
.verdict { display: flex; gap: 16px; flex-wrap: wrap; align-items: stretch; }
.verdict-item { flex: 1; min-width: 140px; background: #f8fafc; border: 1px solid #e2e8f0; border-radius: 10px; padding: 14px 16px; }
.verdict-item .label { font-size: 11px; color: #64748b; text-transform: uppercase; letter-spacing: .4px; margin-bottom: 6px; }
.verdict-item .value { font-size: 20px; font-weight: 700; color: #0f172a; }
.verdict-item .value.accent { color: #2563eb; }
.badge-mixed { display: inline-block; margin-left: 8px; background: #fef3c7; color: #92400e; font-size: 11px; padding: 2px 8px; border-radius: 10px; font-weight: 700; vertical-align: middle; }
.img-preview { text-align: center; background: #222; border-radius: 8px; overflow: hidden; min-height: 200px; display: flex; align-items: center; justify-content: center; }
.img-preview img { max-width: 100%; max-height: 560px; object-fit: contain; display: block; }
.img-preview .placeholder { color: #9ca3af; padding: 40px; font-size: 14px; }
.bar-list { display: flex; flex-direction: column; gap: 10px; }
.bar-row { display: grid; grid-template-columns: 72px 1fr 52px; gap: 10px; align-items: center; font-size: 13px; }
.bar-row .name { font-weight: 600; color: #334155; }
.bar-row .track { height: 10px; background: #e2e8f0; border-radius: 999px; overflow: hidden; }
.bar-row .fill { height: 100%; background: #94a3b8; border-radius: 999px; }
.bar-row.top .fill { background: #2563eb; }
.bar-row.second .fill { background: #60a5fa; }
.bar-row .score { text-align: right; font-variant-numeric: tabular-nums; color: #475569; }
.feat-table { width: 100%; border-collapse: collapse; font-size: 13px; }
.feat-table th, .feat-table td { text-align: left; padding: 8px 12px; border-bottom: 1px solid #f1f5f9; }
.feat-table th { background: #f8fafc; font-weight: 700; color: #475569; font-size: 11px; text-transform: uppercase; letter-spacing: .3px; }
.feat-table td:first-child { font-weight: 600; color: #1e293b; width: 180px; }
.feat-table tr:hover td { background: #f8fafc; }
.json-panel { max-height: 520px; overflow: auto; }
.json-content { padding: 14px 16px; font-family: "SF Mono", "Fira Code", monospace; font-size: 12px; line-height: 1.6; white-space: pre-wrap; word-break: break-all; }
.hidden { display: none !important; }
@media (max-width: 800px) { .results-layout { flex-direction: column; } }
</style>
<script src="/static/img_downscale.js"></script>
</head>
<body>
<div class="container">
<h1>脸型分类调试(MediaPipe</h1>
<p class="subtitle">
POST <code>/api/v1/debug/face-shape</code>
&nbsp;|&nbsp; 本地 <code>face/face_shape_classifier.py</code>7 类)
&nbsp;|&nbsp; 非接口4 豆包分析
</p>
<div class="card">
<div class="card-body">
<div class="upload-row">
<div class="file-input"><input type="file" id="imageFile" accept="image/jpeg,image/png,.jpg,.jpeg,.png"></div>
<button class="btn btn-primary" id="submitBtn" onclick="submitTest()">分析脸型</button>
<button class="btn btn-outline btn-sm" onclick="clearResults()">清除</button>
</div>
<div class="hint">JPG/PNG 正面照 &nbsp;|&nbsp; 走本机 workerMediaPipe),约 1s 内</div>
<div id="statusBar" class="status hidden"></div>
</div>
</div>
<div class="results-layout hidden" id="resultsArea">
<div class="col-main">
<div class="card">
<div class="card-header"><span>判定结果</span></div>
<div class="card-body">
<div class="verdict" id="verdictBox"></div>
</div>
</div>
<div class="card">
<div class="card-header"><span>特征标注图</span></div>
<div class="card-body">
<div class="img-preview" id="imgPreview"><span class="placeholder"></span></div>
</div>
</div>
<div class="card">
<div class="card-header"><span>各脸型得分</span></div>
<div class="card-body">
<div class="bar-list" id="scoreBars"></div>
</div>
</div>
</div>
<div class="col-side">
<div class="card">
<div class="card-header"><span>几何特征</span></div>
<div class="card-body" style="padding:0;max-height:360px;overflow:auto">
<table class="feat-table" id="featTable"></table>
</div>
</div>
<div class="card">
<div class="card-header"><span>原始 JSON</span><button class="btn btn-outline btn-sm" onclick="copyJson()">复制</button></div>
<div class="json-panel"><pre class="json-content" id="jsonContent"></pre></div>
</div>
</div>
</div>
</div>
<script>
const API_BASE = window.location.origin;
const TOKEN = 'dev-shared-secret-2026';
const FEAT_LABELS = {
face_height: '脸高 (px)',
face_width: '脸宽 (px)',
forehead_width: '额宽 (px)',
cheekbone_width: '颧宽 (px)',
jaw_width: '下颌宽 (px)',
chin_width: '下巴宽 (px)',
aspect_ratio: '长宽比 (宽/高)',
forehead_ratio: '额宽/面宽',
cheekbone_ratio: '颧宽/面宽',
jaw_ratio: '下颌宽/面宽',
chin_ratio: '下巴宽/面宽',
chin_sharpness: '下巴尖锐度',
taper_ratio: '额头→下巴收窄',
jaw_angle: '下颌角 (°)',
width_uniformity: '宽度均匀度',
face_curve_score: '面部曲线分',
};
function $(id) { return document.getElementById(id); }
function setStatus(t, type) {
const b = $('statusBar');
b.textContent = t;
b.className = 'status ' + type;
}
function clearResults() {
$('resultsArea').classList.add('hidden');
$('statusBar').className = 'status hidden';
$('imageFile').value = '';
$('jsonContent').textContent = '';
$('imgPreview').innerHTML = '<span class="placeholder">—</span>';
}
async function submitTest() {
let f = $('imageFile').files[0];
if (!f) { setStatus('请选择图片', 'error'); return; }
if (window.downscaleImageFile) f = await window.downscaleImageFile(f);
$('submitBtn').disabled = true;
$('submitBtn').textContent = '分析中...';
setStatus('调用 MediaPipe 脸型分类...', 'info');
$('resultsArea').classList.add('hidden');
const fd = new FormData();
fd.append('image_file', f);
const t0 = performance.now();
try {
const r = await fetch(API_BASE + '/api/v1/debug/face-shape', {
method: 'POST',
headers: { 'X-Internal-Token': TOKEN },
body: fd,
});
const json = await r.json();
const elapsed = ((performance.now() - t0) / 1000).toFixed(2);
$('jsonContent').textContent = JSON.stringify(json, null, 2);
$('resultsArea').classList.remove('hidden');
if (json.code === 0) {
setStatus('完成 (' + elapsed + 's)', 'success');
renderResult(json.data);
} else {
setStatus('(' + elapsed + 's) code=' + json.code + ' ' + json.message, 'error');
}
} catch (e) {
setStatus('请求失败: ' + e.message, 'error');
} finally {
$('submitBtn').disabled = false;
$('submitBtn').textContent = '分析脸型';
}
}
function renderResult(data) {
const mixed = data.is_mixed
? '<span class="badge-mixed">混合 · 次选 ' + (data.second_shape || '—') + '</span>'
: '';
$('verdictBox').innerHTML =
'<div class="verdict-item"><div class="label">脸型</div><div class="value accent">' +
esc(data.display || data.face_shape) + mixed + '</div></div>' +
'<div class="verdict-item"><div class="label">置信度</div><div class="value">' +
(data.confidence * 100).toFixed(1) + '%</div></div>' +
'<div class="verdict-item"><div class="label">分差 score_gap</div><div class="value">' +
(data.score_gap == null ? '—' : data.score_gap) + '</div></div>' +
'<div class="verdict-item"><div class="label">尺寸</div><div class="value" style="font-size:16px">' +
(data.image_size ? data.image_size.width + '×' + data.image_size.height : '—') + '</div></div>';
if (data.annotated_image_base64) {
$('imgPreview').innerHTML =
'<img src="data:image/jpeg;base64,' + data.annotated_image_base64 + '" alt="annotated">';
} else {
$('imgPreview').innerHTML = '<span class="placeholder">无标注图</span>';
}
const ranked = data.ranked || [];
const maxScore = ranked.length ? Math.max.apply(null, ranked.map(function (x) { return x.score; })) : 100;
$('scoreBars').innerHTML = ranked.map(function (row, i) {
const cls = i === 0 ? ' top' : (i === 1 ? ' second' : '');
const pct = maxScore > 0 ? (100 * row.score / maxScore) : 0;
return '<div class="bar-row' + cls + '">' +
'<div class="name">' + esc(row.shape) + '</div>' +
'<div class="track"><div class="fill" style="width:' + pct.toFixed(1) + '%"></div></div>' +
'<div class="score">' + Number(row.score).toFixed(1) + '</div>' +
'</div>';
}).join('');
const feats = data.features || {};
const keys = Object.keys(feats);
let html = '<tr><th>特征</th><th>值</th></tr>';
keys.forEach(function (k) {
const label = FEAT_LABELS[k] || k;
const v = feats[k];
const text = typeof v === 'number' ? (Number.isInteger(v) ? v : v.toFixed(4)) : String(v);
html += '<tr><td>' + esc(label) + '</td><td>' + esc(String(text)) + '</td></tr>';
});
$('featTable').innerHTML = html;
}
function esc(s) {
return String(s).replace(/[&<>"']/g, function (c) {
return ({ '&': '&amp;', '<': '&lt;', '>': '&gt;', '"': '&quot;', "'": '&#39;' })[c];
});
}
function copyJson() {
const t = $('jsonContent').textContent;
if (!t) return;
navigator.clipboard.writeText(t).then(function () {
setStatus('已复制 JSON', 'success');
});
}
</script>
</body>
</html>