#!/bin/bash set -e # ============================================================ # Passport 生成环境部署脚本 - 最终版 # 目标机器：支持 opencode，联网环境 # 包含：m03 模板坐标 + 8 可编辑字段 + 4 固定字段 + MRZ # ============================================================ echo "🚀 Deploying passport generation environment (final version)..." # 1. 安装系统依赖 echo "📦 Installing system packages..." apt-get update && apt-get install -y --no-install-recommends \ python3 python3-pip python3-venv \ fonts-dejavu-core \ wget curl \ && rm -rf /var/lib/apt/lists/* # 2. 安装 Python 依赖 echo "🐍 Installing Python packages..." pip3 install --break-system-packages --no-cache-dir \ pillow==10.4.0 \ opencv-python-headless==4.10.0.84 \ numpy==1.26.4 \ flask # 3. 创建目录结构 echo "📁 Creating directories..." mkdir -p /data/passport/{fonts,models,input,output,scripts,data_json} mkdir -p /data/passport/fonts mkdir -p /data/passport/models # 4. 安装字体 echo "🔤 Installing fonts..." cp /usr/share/fonts/truetype/dejavu/DejaVuSans*.ttf /data/passport/fonts/ 2>/dev/null || true # 5. 下载 LaMa 模型 echo "📥 Downloading LaMa model..." cd /data/passport/models if [ ! -f big-lama.pt ]; then wget -q "https://github.com/Sanster/models/releases/download/add_big_lama/big-lama.pt" -O big-lama.pt fi # 6. 部署核心脚本 echo "📝 Deploying core scripts..." # 6.1 主写入脚本 cat > /data/passport/scripts/passport_m03.py << 'PYEOF' #!/usr/bin/env python3 """ Passport Generation API - m03 Template Version 支持：8 个可编辑字段 + 4 个固定字段 + MRZ 自动生成 基于 m03 OCR 坐标校准的字段位置常量 """ # ============================================================ # m03 模板字段位置常量 (OCR 校准) # ============================================================ # 模板尺寸 M03_WIDTH = 1672 M03_HEIGHT = 1840 # Scale 因子 M03_SCALE_F = min(M03_WIDTH / 900, M03_HEIGHT / 630) # 1.858 # 字段位置映射 (x, y) - 基于 m03 OCR 结果 # 格式: (x_label, y_label, x_value, y_value) M03_FIELD_POS = { # 固定字段 (4个) "type": (626, 1065, 626, 1065), # PV - 左上 "country_code": (784, 1066, 784, 1066), # MMR - 右上 "nationality": (600, 1200, 600, 1200), # MYANMAR - 左列 "authority": (1091, 1397, 1091, 1397), # MOHA, KYAINGTONG - 右列 # 可编辑字段 (8个) - Web 页面变量 "surname": (802, 1039, 802, 1039), # Surname / Nom - 右列第1行 "given_name": (1083, 1039, 1083, 1039), # Given Name / Prenoms - 右列第1行 "sex": (632, 1105, 632, 1105), # Sex / Sexe - 左列第2行 "dob": (649, 1170, 649, 1170), # Date of birth - 左列第3行 "birth_place": (655, 1240, 655, 1240), # Place of birth - 左列第4行 "issue_date": (618, 1305, 618, 1305), # Date of issue - 左列第5行 "expiry_date": (661, 1371, 661, 1371), # Date of expiry - 左列第6行 "passport_no": (1642, 46, 1642, 46), # Passport No - 右上角 } # MRZ 位置 (来自 m02 OCR，m03 无 MRZ 残留) M03_MRZ_POS = { "prefix": (45, 1110), # "P<" "line1": (836, 1126), # MRZ Line 1 "line2": (836, 1170), # MRZ Line 2 } # 字体大小 M03_FONT_SIZES = { "title": max(12, int(24 * M03_SCALE_F)), # 44 "pass": max(10, int(17 * M03_SCALE_F)), # 31 "label": max(8, int(11 * M03_SCALE_F)), # 20 "val": max(9, int(13 * M03_SCALE_F)), # 24 "mrz": max(10, int(18 * M03_SCALE_F)), # 33 "tiny": max(6, int(8 * M03_SCALE_F)), # 14 } # 颜色 M03_COLORS = { "title": "#1a1a1a", "pass": "#6B0000", "gold": "#C4A35A", "label": "#666", "val": "#111", "mrz": "#111", } # ============================================================ # MRZ 生成 (ICAO 9303) # ============================================================ WEIGHTS = (7, 3, 1) def to_val(ch: str) -> int: ch = ch.upper() if ch == "<": return 0 if ch.isdigit(): return int(ch) return ord(ch) - ord("A") + 10 def check_digit(data: str) -> int: return sum(to_val(c) * WEIGHTS[i % 3] for i, c in enumerate(data)) % 10 def _padded(raw: str, n: int) -> str: return raw.upper().ljust(n, "<")[:n] def build_line1(surname: str, given: str = "", type_code: str = "PV", country: str = "MMR") -> str: s = surname.upper().replace(" ", "<") g = given.upper().replace(" ", "<") if given else "" t = type_code[1] if len(type_code) > 1 and type_code[1] != "<" else "<" c = country.upper()[:3] id_part = s + "<<" + g if g else s return ("P" + t + c + id_part).ljust(44, "<")[:44] def build_line2(passport_no: str, birth_date: str, expiry_date: str, sex: str = "M", issuing_state: str = "MMR", personal_no: str = "") -> str: p = _padded(passport_no, 9) b = _padded(birth_date, 6) e = _padded(expiry_date, 6) n = _padded(personal_no, 14) ps = f"{p}{check_digit(p)}" bs = f"{b}{check_digit(b)}" es = f"{e}{check_digit(e)}" ns = n + str(check_digit(n)) if personal_no.strip("<") else n + "<" l43 = ps + issuing_state + bs + sex + es + ns return l43 + str(check_digit(ps + bs + es + ns)) def gen_mrz(data: dict) -> tuple[str, str]: from datetime import datetime dob = datetime.strptime(data["dob"], "%d %b %Y") expiry = datetime.strptime(data["expiry_date"], "%d %b %Y") return build_line1(data["surname"], data["given_name"], data.get("type", "PV"), data.get("country_code", "MMR")), \ build_line2(data["passport_no"], dob.strftime("%y%m%d"), expiry.strftime("%y%m%d"), data.get("sex", "M"), "MMR") # ============================================================ # 主写入函数 (Mask-based rendering) # ============================================================ from PIL import Image, ImageDraw, ImageFont from pathlib import Path def load_font(name: str, size: int) -> ImageFont.FreeTypeFont: font_dir = Path(__file__).parent.parent / "fonts" font_files = { "title": "DejaVuSans-Bold.ttf", "pass": "DejaVuSans-Bold.ttf", "label": "DejaVuSans.ttf", "val": "DejaVuSans-Bold.ttf", "mrz": "OCR-B.otf", "tiny": "DejaVuSans.ttf", } if name not in font_files: raise ValueError(f"Unknown font name: {name}") path = font_dir / font_files[name] if not path.exists(): raise FileNotFoundError(f"Font not found: {path}") return ImageFont.truetype(str(path), size) def hex_to_rgb(hex_color: str) -> tuple: h = hex_color.lstrip('#') if len(h) == 3: h = ''.join(c*2 for c in h) return tuple(int(h[i:i+2], 16) for i in (0, 2, 4)) def draw_text_mask(img: Image.Image, xy: tuple, text: str, font: ImageFont.FreeTypeFont, fill: str, anchor: str = "lt", debug: bool = False, label: str = "") -> None: """使用 font.getmask() 绘制文本，绕过 Pillow 9.5.0 draw.text() bug""" mask = font.getmask(text, mode='L') mask_img = Image.frombytes('L', mask.size, bytes(mask)) mw, mh = mask_img.size x, y = xy # 计算左上角 anchors = { "lt": (x, y), "mt": (x - mw//2, y), "rt": (x - mw, y), "mm": (x - mw//2, y - mh//2), "lb": (x, y - mh), "mb": (x - mw//2, y - mh), "rb": (x - mw, y - mh) } tl_x, tl_y = anchors.get(anchor, (x, y)) # 创建文本层并贴图 text_layer = Image.new('RGB', (mw, mh), hex_to_rgb(fill) if isinstance(fill, str) else fill) img.paste(text_layer, (tl_x, tl_y), mask_img) if debug: draw = ImageDraw.Draw(img) draw.rectangle([tl_x, tl_y, tl_x + mw, tl_y + mh], outline="#FF0000", width=2) draw.ellipse([x-3, y-3, x+3, y+3], fill="#00FF00") debug_font = load_font("sans", 12) draw.text((tl_x, tl_y - 14), f"{label} {mw}x{mh}", fill="#FF0000", font=debug_font) def write_passport(img_path: str, out_path: str, data: dict, debug: bool = False, skip_header: bool = False) -> None: img = Image.open(img_path).convert("RGB") W, H = img.size # 加载字体 F = {k: load_font(k, M03_FONT_SIZES[k]) for k in ["title", "pass", "label", "val", "mrz", "tiny"]} C = M03_COLORS draw = ImageDraw.Draw(img) if not skip_header: # 1. 页眉 draw_text_mask(img, (W//2, 55), "REPUBLIC OF THE UNION OF MYANMAR", F["title"], C["title"], "mt", debug, "title1") draw_text_mask(img, (W//2, 55 + 66), "P A S S P O R T", F["pass"], C["pass"], "mt", debug, "title2") # 2. 右上区域 draw_text_mask(img, (W - 30, 46), f"Passport No {data['passport_no']}", F["label"], "#1a1a1a", "rt", debug, "passport_no") draw_text_mask(img, (W - 30, 83), f"Type {data.get('type','PV')} Code {data.get('country_code','MMR')}", F["tiny"], "#555", "rt", debug, "type_code") # 3. 固定字段 (4个) - 直接写入值 # Type (PV) draw_text_mask(img, (626, 1065), data.get("type", "PV"), F["val"], C["val"], "lt", debug, "type_val") # Country Code (MMR) draw_text_mask(img, (784, 1066), data.get("country_code", "MMR"), F["val"], C["val"], "lt", debug, "country_val") # Nationality (MYANMAR) draw_text_mask(img, (600, 1200), data.get("nationality", "MYANMAR"), F["val"], C["val"], "lt", debug, "nationality_val") # Authority draw_text_mask(img, (1091, 1397), data.get("authority", "MOHA, KYAINGTONG"), F["val"], C["val"], "lt", debug, "authority_val") # 4. 可编辑字段 (8个) - Web 页面传入的变量 editable = [ ("surname", 802, 1039, "Surname"), ("given_name", 1083, 1039, "GivenName"), ("sex", 632, 1105, "Sex"), ("dob", 649, 1170, "DOB"), ("birth_place", 655, 1240, "BirthPlace"), ("issue_date", 618, 1305, "IssueDate"), ("expiry_date", 661, 1371, "ExpiryDate"), ("passport_no", 1642, 46, "PassportNo"), ] for key, x, y, label in editable: val = data.get(key, "") # passport_no 在右上角用 rt anchor anchor = "rt" if key == "passport_no" else "lt" draw_text_mask(img, (x, y), str(val), F["val"], C["val"], anchor, debug, f"field_{label}") # 5. MRZ 生成 l1, l2 = gen_mrz(data) draw_text_mask(img, (45, 1110), "P<", F["mrz"], C["mrz"], "lt", debug, "mrz_prefix") draw_text_mask(img, (836, 1126), l1, F["mrz"], C["mrz"], "mt", debug, "mrz_l1") draw_text_mask(img, (836, 1170), l2, F["mrz"], C["mrz"], "mt", debug, "mrz_l2") img.save(out_path, dpi=(300, 300)) print(f"✓ Generated: {out_path} ({W}x{H})") # ============================================================ # CLI 入口 # ============================================================ if __name__ == "__main__": import argparse, json, sys p = argparse.ArgumentParser(description="Passport generator - m03 template") p.add_argument("--img", required=True) p.add_argument("--out", required=True) p.add_argument("--data", required=True, help="JSON or @file.json") p.add_argument("--debug", action="store_true") p.add_argument("--skip-header", action="store_true") args = p.parse_args() if args.data.startswith("@"): with open(args.data[1:]) as f: data = json.load(f) else: data = json.loads(args.data) # 必需字段：8 可编辑 + 4 固定 required = ["surname", "given_name", "nationality", "dob", "sex", "birth_place", "issue_date", "expiry_date", "authority", "passport_no", "type", "country_code"] missing = [k for k in required if k not in data] if missing: sys.exit(f"Missing fields: {missing}") write_passport(args.img, args.out, data, args.debug, args.skip_header) PYEOF chmod +x /data/passport/scripts/passport_m03.py # 6.2 MRZ 辅助脚本 cat > /data/passport/scripts/mrz.py << 'PYEOF' WEIGHTS=(7,3,1) def to_val(ch): ch=ch.upper() if ch=="<": return 0 if ch.isdigit(): return int(ch) return ord(ch)-ord("A")+10 def check_digit(data): return sum(to_val(c)*WEIGHTS[i%3] for i,c in enumerate(data))%10 def _padded(raw,n): return raw.upper().ljust(n,"<")[:n] def build_line1(surname, given="", type_code="PV", country="MMR"): s=surname.upper().replace(" ","<"); g=given.upper().replace(" ","<") if given else "" t=type_code[1] if len(type_code)>1 and type_code[1]!="<" else "<" c=country.upper()[:3]; idp=s+"<<"+g if g else s return ("P"+t+c+idp).ljust(44,"<")[:44] def build_line2(passport_no, birth_date, expiry_date, sex="M", issuing_state="MMR", personal_no=""): p=_padded(passport_no,9); b=_padded(birth_date,6); e=_padded(expiry_date,6); n=_padded(personal_no,14) ps=f"{p}{check_digit(p)}"; bs=f"{b}{check_digit(b)}"; es=f"{e}{check_digit(e)}" ns=n+check_digit(n) if personal_no.strip("<") else n+"<" l43=ps+issuing_state+bs+sex+es+ns return l43+str(check_digit(ps+bs+es+ns)) PYEOF # 6.3 生成掩码脚本 cat > /data/passport/scripts/gen_mask.py << 'PYEOF' #!/usr/bin/env python3 import argparse, cv2, numpy as np def main(): p = argparse.ArgumentParser() p.add_argument("--img", required=True); p.add_argument("--mask", required=True); p.add_argument("--template", default="m03") args = p.parse_args() img = cv2.imread(args.img); h,w = img.shape[:2] mask = np.zeros((h,w), np.uint8) if args.template=="m03": boxes = [[536,939,1109,968],[302,1023,557,1062],[606,1031,648,1054],[741,1029,863,1050],[1026,1029,1140,1050],[604,1050,648,1081],[741,1052,828,1081],[600,1092,664,1119],[602,1160,697,1181],[598,1184,784,1217],[597,1224,714,1256],[601,1289,641,1322],[911,1292,1033,1313],[598,1361,724,1382],[911,1359,995,1380],[908,1380,1275,1415],[598,1432,728,1453],[910,1428,1086,1453],[1006,1457,1041,1489]] else: raise ValueError(args.template) sf = lambda v: max(1, int(v*min(w/900,h/630))) for x1,y1,x2,y2 in boxes: cv2.rectangle(mask,(max(0,x1-8),max(0,y1-8)),(min(w,x2+8),min(h,y2+8)),255,-1) protect = [(sf(280),sf(118),sf(445),sf(318)),(0,0,w,sf(100)),(0,sf(1080),w,sf(1200)),(0,0,sf(20),h),(w-sf(20),0,w,h),(0,0,w,sf(20)),(0,h-sf(20),w,h)] for x1,y1,x2,y2 in protect: cv2.rectangle(mask,(x1,y1),(x2,y2),0,-1) mask = cv2.morphologyEx(mask, cv2.MORPH_CLOSE, np.ones((5,5),np.uint8)) cv2.imwrite(args.mask, mask) print(f"✓ Mask: {args.mask}") if __name__=="__main__": main() PYEOF chmod +x /data/passport/scripts/gen_mask.py # 6.4 LaMa inpaint 脚本 cat > /data/passport/scripts/lama_inpaint.py << 'PYEOF' #!/usr/bin/env python3 import argparse, torch, cv2, numpy as np from pathlib import Path def pad(x, m=8): h,w = x.shape[:2] nh,nw = ((h+m-1)//m)*m, ((w+m-1)//m)*m ph,pw = nh-h, nw-w if x.ndim==3: return np.pad(x,((0,ph),(0,pw),(0,0)),mode="reflect"),(ph,pw) return np.pad(x,((0,ph),(0,pw)),mode="reflect"),(ph,pw) def main(): p = argparse.ArgumentParser() p.add_argument("--img", required=True); p.add_argument("--mask", required=True); p.add_argument("--out", required=True) p.add_argument("--model", default="/data/passport/models/big-lama.pt") args = p.parse_args() model = torch.jit.load(args.model, map_location="cpu"); model.eval() img = cv2.imread(args.img); mask = cv2.imread(args.mask, 0) img = cv2.cvtColor(img, cv2.COLOR_BGR2RGB); h,w = img.shape[:2] img_pad,_ = pad(img); mask_pad,_ = pad(mask) img_t = torch.from_numpy(img_pad).float()/255.0; img_t = img_t.permute(2,0,1).unsqueeze(0) mask_t = torch.from_numpy(mask_pad).float()/255.0; mask_t = mask_t.unsqueeze(0).unsqueeze(0) with torch.no_grad(): out = model(img_t, mask_t) out = out[0].permute(1,2,0).cpu().numpy() out = np.clip(out*255,0,255).astype(np.uint8) out = cv2.cvtColor(out, cv2.COLOR_RGB2BGR)[:h,:w] cv2.imwrite(args.out, out) print(f"✓ {args.out}") if __name__=="__main__": main() PYEOF chmod +x /data/passport/scripts/lama_inpaint.py # 7. 生成干净模板（如果尚不存在） echo "🧹 Generating clean template if needed..." cd /data/passport if [ ! -f input/m03_clean.png ]; then echo " Generating mask..." python3 scripts/gen_mask.py --img /passport/模板/m03.jpeg --mask input/m03_mask.png --template m03 echo " Running LaMa inpainting..." python3 scripts/lama_inpaint.py --img /passport/模板/m03.jpeg --mask input/m03_mask.png --out input/m03_clean.png fi # 8. 设置权限 echo "🔐 Setting permissions..." chown -R www-data:www-data /data/passport chmod -R 755 /data/passport/scripts echo "" echo "✅ Deployment complete!" echo "" echo "📁 目录结构:" echo " /data/passport/" echo " ├── fonts/ # DejaVu 字体" echo " ├── models/ # big-lama.pt" echo " ├── input/ # m03_clean.png (干净底图), m03_mask.png" echo " ├── output/ # 生成结果" echo " ├── scripts/ # 核心脚本:" echo " │ ├── passport_m03.py # 主生成脚本" echo " │ ├── mrz.py # MRZ 生成" echo " │ ├── gen_mask.py # 掩码生成" echo " │ └── lama_inpaint.py # LaMa 修复" echo " └── data_json/ # 批量数据目录" echo "" echo "🚀 使用示例:" echo "" echo " # 1. 单次生成 (命令行)" echo " python3 /data/passport/scripts/passport_m03.py \\" echo " --img /data/passport/input/m03_clean.png \\" echo " --out /data/passport/output/test.png \\" echo " --data '{\"surname\":\"LI\",\"given_name\":\"WEI\",\"nationality\":\"MYANMAR\",\"dob\":\"15 JAN 1995\",\"sex\":\"M\",\"birth_place\":\"YANGON\",\"issue_date\":\"10 MAR 2023\",\"expiry_date\":\"09 MAR 2028\",\"authority\":\"MOHA, KYAINGTONG\",\"passport_no\":\"MK888888\",\"type\":\"PV\",\"country_code\":\"MMR\"}' \\" echo " --skip-header" echo "" echo " # 2. 批量生成" echo " mkdir -p /data/passport/data_json" echo " echo '[{\"surname\":\"LI\",\"given_name\":\"WEI\",...}]' > /data/passport/data_json/batch.json" echo " python3 /data/passport/scripts/passport_m03.py \\" echo " --img /data/passport/input/m03_clean.png \\" echo " --out /data/passport/output/batch_001.png \\" echo " --data @/data/passport/data_json/batch.json" echo "" echo " # 3. 查看帮助" echo " python3 /data/passport/scripts/passport_m03.py --help" echo "" echo "💡 提示: 生成的护照将保存在 /data/passport/output/ 目录"