修正docker镜像构建

services/paddle_services/Dockerfile

@@ -12,7 +12,7 @@ ENV PYTHONUNBUFFERED=1 \
     PIP_INDEX_URL=https://pypi.tuna.tsinghua.edu.cn/simple
 
 # 安装依赖
-COPY ./services/paddle_services/requestments.txt /app/requirements.txt
+COPY requestments.txt /app/requirements.txt
 RUN ln -snf /usr/share/zoneinfo/$TZ /etc/localtime && echo '$TZ' > /etc/timezone \
     && pip install --no-cache-dir -r requirements.txt \
     && pip uninstall -y onnxruntime onnxruntime-gpu \

services/paddle_services/__init__.py

@@ -0,0 +1,34 @@
+"""
+信息抽取关键词配置
+"""
+
+# 患者姓名
+PATIENT_NAME = ['患者姓名']
+# 入院日期
+ADMISSION_DATE = ['入院日期']
+# 出院日期
+DISCHARGE_DATE = ['出院日期']
+# 发生医疗费
+MEDICAL_EXPENSES = ['费用总额']
+# 个人现金支付
+PERSONAL_CASH_PAYMENT = ['个人现金支付']
+# 个人账户支付
+PERSONAL_ACCOUNT_PAYMENT = ['个人账户支付']
+# 个人自费金额
+PERSONAL_FUNDED_AMOUNT = ['自费金额', '个人自费']
+# 医保类别
+MEDICAL_INSURANCE_TYPE = ['医保类型']
+# 就诊医院
+HOSPITAL = ['医院']
+# 就诊科室
+DEPARTMENT = ['科室']
+# 主治医生
+DOCTOR = ['主治医生']
+# 住院号
+ADMISSION_ID = ['住院号']
+# 医保结算单号码
+SETTLEMENT_ID = ['医保结算单号码']
+# 年龄
+AGE = ['年龄']
+# 大写总额
+UPPERCASE_MEDICAL_EXPENSES = ['大写总额']

services/paddle_services/clas_api.py

@@ -0,0 +1,26 @@
+from flask import Flask, request
+from paddleclas import PaddleClas
+
+from utils import process_request
+
+app = Flask(__name__)
+CLAS = PaddleClas(model_name='text_image_orientation')
+
+
+@app.route('/clas/orientation', methods=['POST'])
+@process_request
+def orientation():
+    """
+    判断图片旋转角度，逆时针旋转该角度后为正。可能值['0', '90', '180', '270']
+    :return: 最有可能的两个角度
+    """
+    img_path = request.form.get('img_path')
+    clas_result = CLAS.predict(input_data=img_path)
+    clas_result = next(clas_result)[0]
+    if clas_result['scores'][0] < 0.5:
+        return ['0', '90']
+    return clas_result['label_names']
+
+
+if __name__ == '__main__':
+    app.run('0.0.0.0', 5005)

services/paddle_services/cost_api.py

@@ -0,0 +1,24 @@
+import json
+
+from flask import Flask, request
+from paddlenlp import Taskflow
+
+from utils import process_request
+from . import PATIENT_NAME, ADMISSION_DATE, DISCHARGE_DATE, MEDICAL_EXPENSES
+
+app = Flask(__name__)
+COST_LIST_SCHEMA = PATIENT_NAME + ADMISSION_DATE + DISCHARGE_DATE + MEDICAL_EXPENSES
+COST = Taskflow('information_extraction', schema=COST_LIST_SCHEMA, model='uie-x-base',
+                task_path='../../model/cost_list_model', layout_analysis=False, precision='fp16')
+
+
+@app.route('/nlp/cost', methods=['POST'])
+@process_request
+def cost():
+    img_path = request.form.get('img_path')
+    layout = request.form.get('layout')
+    return COST({'doc': img_path, 'layout': json.loads(layout)})
+
+
+if __name__ == '__main__':
+    app.run('0.0.0.0', 5004)

services/paddle_services/det_api.py

@@ -0,0 +1,28 @@
+import os.path
+
+import cv2
+from flask import Flask, request
+
+from paddle_detection import detector
+from utils import process_request, parse_img_path
+
+app = Flask(__name__)
+
+
+@app.route('/det/books', methods=['POST'])
+@process_request
+def books():
+    img_path = request.form.get('img_path')
+    result = detector.get_book_areas(img_path)
+
+    dirname, img_name, ext = parse_img_path(img_path)
+    books_path = []
+    for i in range(len(result)):
+        save_path = os.path.join(dirname, img_name + '_book_' + str(i) + '.' + ext)
+        cv2.imwrite(save_path, result[i])
+        books_path.append(save_path)
+    return books_path
+
+
+if __name__ == '__main__':
+    app.run('0.0.0.0', 5006)

services/paddle_services/dewarp_api.py

@@ -0,0 +1,25 @@
+import os
+
+import cv2
+from flask import Flask, request
+
+from doc_dewarp import dewarper
+from utils import process_request, parse_img_path
+
+app = Flask(__name__)
+
+
+@app.route('/dewarp', methods=['POST'])
+@process_request
+def dewarp():
+    img_path = request.form.get('img_path')
+    img = cv2.imread(img_path)
+    dewarped_img = dewarper.dewarp_image(img)
+    dirname, img_name, ext = parse_img_path(img_path)
+    save_path = os.path.join(dirname, img_name + '_dewarped.' + ext)
+    cv2.imwrite(save_path, dewarped_img)
+    return save_path
+
+
+if __name__ == '__main__':
+    app.run('0.0.0.0', 5007)

services/paddle_services/discharge_api.py

@@ -0,0 +1,26 @@
+import json
+
+from flask import Flask, request
+from paddlenlp import Taskflow
+
+from utils import process_request
+from . import HOSPITAL, DEPARTMENT, PATIENT_NAME, ADMISSION_DATE, DISCHARGE_DATE, DOCTOR, ADMISSION_ID, AGE
+
+app = Flask(__name__)
+DISCHARGE_RECORD_SCHEMA = (
+        HOSPITAL + DEPARTMENT + PATIENT_NAME + ADMISSION_DATE + DISCHARGE_DATE + DOCTOR + ADMISSION_ID + AGE
+)
+DISCHARGE = Taskflow('information_extraction', schema=DISCHARGE_RECORD_SCHEMA, model='uie-x-base',
+                     task_path='../../model/discharge_record_model', layout_analysis=False, precision='fp16')
+
+
+@app.route('/nlp/discharge', methods=['POST'])
+@process_request
+def discharge():
+    img_path = request.form.get('img_path')
+    layout = request.form.get('layout')
+    return DISCHARGE({'doc': img_path, 'layout': json.loads(layout)})
+
+
+if __name__ == '__main__':
+    app.run('0.0.0.0', 5003)

services/paddle_services/ocr_api.py

@@ -0,0 +1,18 @@
+from flask import Flask, request
+from paddleocr import PaddleOCR
+
+from utils import process_request
+
+app = Flask(__name__)
+OCR = PaddleOCR(use_angle_cls=False, show_log=False, gpu_id=0, det_db_box_thresh=0.3)
+
+
+@app.route('/ocr', methods=['POST'])
+@process_request
+def ocr():
+    img_path = request.form.get('img_path')
+    return OCR.ocr(img_path, cls=False)
+
+
+if __name__ == '__main__':
+    app.run('0.0.0.0', 5001)

services/paddle_services/settlement_api.py

@@ -0,0 +1,30 @@
+import json
+
+from flask import Flask, request
+from paddlenlp import Taskflow
+
+from utils import process_request
+from . import PATIENT_NAME, ADMISSION_DATE, DISCHARGE_DATE, MEDICAL_EXPENSES, PERSONAL_CASH_PAYMENT, \
+    PERSONAL_ACCOUNT_PAYMENT, PERSONAL_FUNDED_AMOUNT, MEDICAL_INSURANCE_TYPE, ADMISSION_ID, SETTLEMENT_ID, \
+    UPPERCASE_MEDICAL_EXPENSES
+
+app = Flask(__name__)
+SETTLEMENT_LIST_SCHEMA = (
+        PATIENT_NAME + ADMISSION_DATE + DISCHARGE_DATE + MEDICAL_EXPENSES + PERSONAL_CASH_PAYMENT
+        + PERSONAL_ACCOUNT_PAYMENT + PERSONAL_FUNDED_AMOUNT + MEDICAL_INSURANCE_TYPE + ADMISSION_ID + SETTLEMENT_ID
+        + UPPERCASE_MEDICAL_EXPENSES
+)
+SETTLEMENT_IE = Taskflow('information_extraction', schema=SETTLEMENT_LIST_SCHEMA, model='uie-x-base',
+                         task_path='../../model/settlement_list_model', layout_analysis=False, precision='fp16')
+
+
+@app.route('/nlp/settlement', methods=['POST'])
+@process_request
+def settlement():
+    img_path = request.form.get('img_path')
+    layout = request.form.get('layout')
+    return SETTLEMENT_IE({'doc': img_path, 'layout': json.loads(layout)})
+
+
+if __name__ == '__main__':
+    app.run('0.0.0.0', 5002)

services/paddle_services/utils.py

@@ -0,0 +1,24 @@
+import os
+
+from flask import jsonify
+
+
+def process_request(func):
+    """
+    api通用处理函数
+    """
+
+    def wrapper(*args, **kwargs):
+        try:
+            result = func(*args, **kwargs)
+            return jsonify(result), 200
+        except Exception as e:
+            return jsonify({'error': str(e)}), 500
+
+    return wrapper
+
+
+def parse_img_path(img_path):
+    dirname = os.path.dirname(img_path)
+    img_name, ext = os.path.basename(img_path).rsplit('.', 1)
+    return dirname, img_name, ext