FORMAT_CONVERSION_MAXCOMPUTE/ocr/tools/infer/predict_rec_pytorch.py

# encoding=utf8
# Copyright (c) 2020 PaddlePaddle Authors. All Rights Reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
#     http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
import io
import logging
import os
import sys
# __dir__ = os.path.dirname(os.path.abspath(__file__))
import zlib

import requests
# sys.path.append(__dir__)
# sys.path.append(os.path.abspath(os.path.join(__dir__, '../..')))

sys.path.append(os.path.dirname(os.path.abspath(__file__)) + "/../../../")
os.environ["FLAGS_allocator_strategy"] = 'auto_growth'
import cv2
import numpy as np
import math
import time
import traceback
os.environ['FLAGS_eager_delete_tensor_gb'] = '0'
import paddle

import ocr.tools.infer.utility as utility
from ocr.ppocr.postprocess import build_post_process
from ocr.ppocr.utils.logging import get_logger
from ocr.ppocr.utils.utility import get_image_file_list, check_and_read_gif

from format_convert.utils import judge_error_code, log, namespace_to_dict,get_platform
from format_convert import _global

import torch
from torch import nn
from ocr.tools.infer.torch_rec_model import Rec_ResNet_34
import gc

logger = get_logger()


class TextRecognizer(object):
    shrink_memory_count = 0

    def __init__(self, args):
        self.rec_image_shape = [int(v) for v in args.rec_image_shape.split(",")]
        self.character_type = args.rec_char_type
        self.rec_batch_num = args.rec_batch_num
        self.rec_algorithm = args.rec_algorithm
        postprocess_params = {
            'name': 'CTCLabelDecode',
            "character_type": args.rec_char_type,
            "character_dict_path": args.rec_char_dict_path,
            # "use_space_char": args.use_space_char
            "use_space_char": False
        }
        # if self.rec_algorithm == "SRN":
        #     postprocess_params = {
        #         'name': 'SRNLabelDecode',
        #         "character_type": args.rec_char_type,
        #         "character_dict_path": args.rec_char_dict_path,
        #         "use_space_char": args.use_space_char
        #     }
        # elif self.rec_algorithm == "RARE":
        #     postprocess_params = {
        #         'name': 'AttnLabelDecode',
        #         "character_type": args.rec_char_type,
        #         "character_dict_path": args.rec_char_dict_path,
        #         "use_space_char": args.use_space_char
        #     }
        self.postprocess_op = build_post_process(postprocess_params)
        # self.predictor, self.input_tensor, self.output_tensors = \
        #     utility.create_predictor(args, 'rec', logger)

        rec_model_path = args.rec_model_dir
        self.device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
        model = Rec_ResNet_34()
        mode_state_dict = torch.load(rec_model_path, self.device)['state_dict']
        if str(self.device)=='cpu': # cpu处理时精度调整，加速推理
            for name, value in mode_state_dict.items():
                if get_platform() != "Windows":
                    value = value.double()
                value = torch.where((value < 1.0e-23) & (value > 0.0), 1.0e-23, value)
                value = torch.where((value > -1.0e-23) & (value < 0.0), -1.0e-23, value)
                mode_state_dict[name] = value

        model.load_state_dict(mode_state_dict)

        self.predictor = model
        self.predictor.to(self.device)
        self.predictor.eval()

    def resize_norm_img(self, img, max_wh_ratio):
        h, w = img.shape[:2]
        imgC, imgH, imgW = self.rec_image_shape
        assert imgC == img.shape[2]
        # print('max_wh_ratio', max_wh_ratio)
        if max_wh_ratio < 0.1:
            # if h > imgW:
            #     resized_image = cv2.resize(img, (w, imgW))
            # else:
            #     resized_image = img

            # max_wh_ratio h是w的10倍，直接跳过
            resized_w = None
        else:
            if self.character_type == "ch":
                imgW = int((32 * max_wh_ratio))

            ratio = w / float(h)
            if math.ceil(imgH * ratio) > imgW:
                resized_w = imgW
            else:
                resized_w = int(math.ceil(imgH * ratio))

            try:
                resized_image = cv2.resize(img, (resized_w, imgH))
            except:
                log("predict_rec.py resize_norm_img resize shape " + str((resized_w, imgH, imgW, h, w, ratio, max_wh_ratio)) + ' ' + str(self.rec_image_shape))
                raise

        resized_image = resized_image.astype('float32')
        resized_image = resized_image.transpose((2, 0, 1)) / 255
        resized_image -= 0.5
        resized_image /= 0.5
        padding_im = np.zeros((imgC, imgH, imgW), dtype=np.float32)
        if resized_w is not None:
            padding_im[:, :, 0:resized_w] = resized_image
        return padding_im

    def resize_norm_img_srn(self, img, image_shape):
        imgC, imgH, imgW = image_shape

        img_black = np.zeros((imgH, imgW))
        im_hei = img.shape[0]
        im_wid = img.shape[1]

        if im_wid <= im_hei * 1:
            img_new = cv2.resize(img, (imgH * 1, imgH))
        elif im_wid <= im_hei * 2:
            img_new = cv2.resize(img, (imgH * 2, imgH))
        elif im_wid <= im_hei * 3:
            img_new = cv2.resize(img, (imgH * 3, imgH))
        else:
            img_new = cv2.resize(img, (imgW, imgH))

        img_np = np.asarray(img_new)
        img_np = cv2.cvtColor(img_np, cv2.COLOR_BGR2GRAY)
        img_black[:, 0:img_np.shape[1]] = img_np
        img_black = img_black[:, :, np.newaxis]

        row, col, c = img_black.shape
        c = 1

        return np.reshape(img_black, (c, row, col)).astype(np.float32)

    def srn_other_inputs(self, image_shape, num_heads, max_text_length):

        imgC, imgH, imgW = image_shape
        feature_dim = int((imgH / 8) * (imgW / 8))

        encoder_word_pos = np.array(range(0, feature_dim)).reshape(
            (feature_dim, 1)).astype('int64')
        gsrm_word_pos = np.array(range(0, max_text_length)).reshape(
            (max_text_length, 1)).astype('int64')

        gsrm_attn_bias_data = np.ones((1, max_text_length, max_text_length))
        gsrm_slf_attn_bias1 = np.triu(gsrm_attn_bias_data, 1).reshape(
            [-1, 1, max_text_length, max_text_length])
        gsrm_slf_attn_bias1 = np.tile(
            gsrm_slf_attn_bias1,
            [1, num_heads, 1, 1]).astype('float32') * [-1e9]

        gsrm_slf_attn_bias2 = np.tril(gsrm_attn_bias_data, -1).reshape(
            [-1, 1, max_text_length, max_text_length])
        gsrm_slf_attn_bias2 = np.tile(
            gsrm_slf_attn_bias2,
            [1, num_heads, 1, 1]).astype('float32') * [-1e9]

        encoder_word_pos = encoder_word_pos[np.newaxis, :]
        gsrm_word_pos = gsrm_word_pos[np.newaxis, :]

        return [
            encoder_word_pos, gsrm_word_pos, gsrm_slf_attn_bias1,
            gsrm_slf_attn_bias2
        ]

    def process_image_srn(self, img, image_shape, num_heads, max_text_length):
        norm_img = self.resize_norm_img_srn(img, image_shape)
        norm_img = norm_img[np.newaxis, :]

        [encoder_word_pos, gsrm_word_pos, gsrm_slf_attn_bias1, gsrm_slf_attn_bias2] = \
            self.srn_other_inputs(image_shape, num_heads, max_text_length)

        gsrm_slf_attn_bias1 = gsrm_slf_attn_bias1.astype(np.float32)
        gsrm_slf_attn_bias2 = gsrm_slf_attn_bias2.astype(np.float32)
        encoder_word_pos = encoder_word_pos.astype(np.int64)
        gsrm_word_pos = gsrm_word_pos.astype(np.int64)

        return (norm_img, encoder_word_pos, gsrm_word_pos, gsrm_slf_attn_bias1,
                gsrm_slf_attn_bias2)

    def __call__(self, img_list):
        img_num = len(img_list)
        # Calculate the aspect ratio of all text bars
        width_list = []
        i = 0
        for img in img_list:
            # cv2.imwrite('D:/myProject/format_conversion_maxcompute/ocr/test/'+str(i)+'.jpg',img)
            # i+=1
            # cv2.imshow('img', img)
            # cv2.waitKey(1000)
            width_list.append(img.shape[1] / float(img.shape[0]))
        # Sorting can speed up the recognition process
        indices = np.argsort(np.array(width_list))

        # rec_res = []
        rec_res = [['', 0.0]] * img_num
        batch_num = self.rec_batch_num
        elapse = 0
        for beg_img_no in range(0, img_num, batch_num):
            end_img_no = min(img_num, beg_img_no + batch_num)
            norm_img_batch = []
            max_wh_ratio = 0
            for ino in range(beg_img_no, end_img_no):
                # h, w = img_list[ino].shape[0:2]
                h, w = img_list[indices[ino]].shape[0:2]
                wh_ratio = w * 1.0 / h
                max_wh_ratio = max(max_wh_ratio, wh_ratio)
            # print('max_wh_ratio',max_wh_ratio)
            for ino in range(beg_img_no, end_img_no):
                if self.rec_algorithm != "SRN":
                    # print('max_wh_ratio', max_wh_ratio)
                    norm_img = self.resize_norm_img(img_list[indices[ino]],
                                                    max_wh_ratio)
                    # cv2.imshow('img', norm_img.transpose(1,2,0))
                    # cv2.waitKey(1000)
                    norm_img = norm_img[np.newaxis, :]
                    norm_img_batch.append(norm_img)
                else:
                    # norm_img = self.process_image_srn(
                    #     img_list[indices[ino]], self.rec_image_shape, 8, 25)
                    # encoder_word_pos_list = []
                    # gsrm_word_pos_list = []
                    # gsrm_slf_attn_bias1_list = []
                    # gsrm_slf_attn_bias2_list = []
                    # encoder_word_pos_list.append(norm_img[1])
                    # gsrm_word_pos_list.append(norm_img[2])
                    # gsrm_slf_attn_bias1_list.append(norm_img[3])
                    # gsrm_slf_attn_bias2_list.append(norm_img[4])
                    # norm_img_batch.append(norm_img[0])
                    pass
            norm_img_batch = np.concatenate(norm_img_batch)
            norm_img_batch = norm_img_batch.copy()

            if self.rec_algorithm == "SRN":
                # starttime = time.time()
                # encoder_word_pos_list = np.concatenate(encoder_word_pos_list)
                # gsrm_word_pos_list = np.concatenate(gsrm_word_pos_list)
                # gsrm_slf_attn_bias1_list = np.concatenate(
                #     gsrm_slf_attn_bias1_list)
                # gsrm_slf_attn_bias2_list = np.concatenate(
                #     gsrm_slf_attn_bias2_list)
                #
                # inputs = [
                #     norm_img_batch,
                #     encoder_word_pos_list,
                #     gsrm_word_pos_list,
                #     gsrm_slf_attn_bias1_list,
                #     gsrm_slf_attn_bias2_list,
                # ]
                # input_names = self.predictor.get_input_names()
                # for i in range(len(input_names)):
                #     input_tensor = self.predictor.get_input_handle(input_names[
                #         i])
                #     input_tensor.copy_from_cpu(inputs[i])
                # self.predictor.run()
                # outputs = []
                # for output_tensor in self.output_tensors:
                #     output = output_tensor.copy_to_cpu()
                #     outputs.append(output)
                # preds = {"predict": outputs[2]}
                pass
            else:
                starttime = time.time()

                tensor = torch.from_numpy(norm_img_batch).float()
                start_time = time.time()
                tensor = tensor.to(self.device)
                with torch.no_grad():
                    out = self.predictor(tensor)
                logging.info("ocr model predict time - rec" + str(time.time()-start_time))
                out = out.cpu().numpy()
                preds = out

            # print("tools/infer/predict_rec preds", preds)
            rec_result = self.postprocess_op(preds)
            for rno in range(len(rec_result)):
                # print("predict_rec", img_num, batch_num, beg_img_no,
                #       indices[beg_img_no + rno], len(rec_res))
                rec_res[indices[beg_img_no + rno]] = rec_result[rno]
            elapse += time.time() - starttime
            # 释放内存
            # self.predictor.clear_intermediate_tensor()
            # self.predictor.try_shrink_memory()

            # gc.collect()
            if str(self.device)!='cpu':
                torch.cuda.empty_cache()
            #     gc.collect()
        return rec_res, elapse


class TextRecognizer2(object):
    shrink_memory_count = 0

    def __init__(self, args):
        self.rec_image_shape = [int(v) for v in args.rec_image_shape.split(",")]
        self.character_type = args.rec_char_type
        self.rec_batch_num = args.rec_batch_num
        self.rec_algorithm = args.rec_algorithm
        postprocess_params = {
            'name': 'CTCLabelDecode',
            "character_type": args.rec_char_type,
            "character_dict_path": args.rec_char_dict_path,
            "use_space_char": args.use_space_char
        }
        self.postprocess_op = build_post_process(postprocess_params)
        self.args = args
        # self.predictor, self.input_tensor, self.output_tensors = \
        #     utility.create_predictor(args, 'rec', logger)

    def resize_norm_img(self, img, max_wh_ratio):
        imgC, imgH, imgW = self.rec_image_shape
        assert imgC == img.shape[2]
        if self.character_type == "ch":
            imgW = int((32 * max_wh_ratio))
        h, w = img.shape[:2]
        ratio = w / float(h)
        if math.ceil(imgH * ratio) > imgW:
            resized_w = imgW
        else:
            resized_w = int(math.ceil(imgH * ratio))
        # print("predict_rec.py resize_norm_img resize shape", (resized_w, imgH))
        resized_image = cv2.resize(img, (resized_w, imgH))
        resized_image = resized_image.astype('float32')
        resized_image = resized_image.transpose((2, 0, 1)) / 255
        resized_image -= 0.5
        resized_image /= 0.5
        padding_im = np.zeros((imgC, imgH, imgW), dtype=np.float32)
        padding_im[:, :, 0:resized_w] = resized_image
        return padding_im

    def __call__(self, img_list):
        from format_convert.convert_need_interface import from_gpu_interface_redis
        img_num = len(img_list)
        # Calculate the aspect ratio of all text bars
        width_list = []
        for img in img_list:
            width_list.append(img.shape[1] / float(img.shape[0]))

        # Sorting can speed up the recognition process
        indices = np.argsort(np.array(width_list))

        rec_res = [['', 0.0]] * img_num
        batch_num = self.rec_batch_num
        elapse = 0
        all_gpu_time = 0
        for beg_img_no in range(0, img_num, batch_num):
            # 预处理
            end_img_no = min(img_num, beg_img_no + batch_num)
            norm_img_batch = []
            max_wh_ratio = 0
            for ino in range(beg_img_no, end_img_no):
                h, w = img_list[indices[ino]].shape[0:2]
                wh_ratio = w * 1.0 / h
                max_wh_ratio = max(max_wh_ratio, wh_ratio)
            for ino in range(beg_img_no, end_img_no):
                norm_img = self.resize_norm_img(img_list[indices[ino]],
                                                max_wh_ratio)
                norm_img = norm_img[np.newaxis, :]
                norm_img_batch.append(norm_img)
            norm_img_batch = np.concatenate(norm_img_batch)
            norm_img_batch = norm_img_batch.copy()
            starttime = time.time()

            # # 压缩numpy
            # compressed_array = io.BytesIO()
            # np.savez_compressed(compressed_array, norm_img_batch)
            # compressed_array.seek(0)
            # norm_img_batch = compressed_array.read()

            # 调用GPU接口
            _dict = {"inputs": norm_img_batch, "args": str(namespace_to_dict(self.args)), "md5": _global.get("md5")}
            result = from_gpu_interface_redis(_dict, model_type="ocr", predictor_type="rec")
            if judge_error_code(result):
                logging.error("from_gpu_interface failed! " + str(result))
                raise requests.exceptions.RequestException

            preds = result.get("preds")
            gpu_time = result.get("gpu_time")
            all_gpu_time += round(gpu_time, 2)

            # # 解压numpy
            # decompressed_array = io.BytesIO()
            # decompressed_array.write(preds)
            # decompressed_array.seek(0)
            # preds = np.load(decompressed_array, allow_pickle=True)['arr_0']
            # log("inputs.shape" + str(preds.shape))

            # 后处理
            rec_result = self.postprocess_op(preds)
            for rno in range(len(rec_result)):
                rec_res[indices[beg_img_no + rno]] = rec_result[rno]
            elapse += time.time() - starttime

        log("ocr model predict time - rec - time " + str(all_gpu_time) + " - num " + str(img_num))
        return rec_res, elapse


def main(args):
    image_file_list = get_image_file_list(args.image_dir)
    text_recognizer = TextRecognizer(args)
    valid_image_file_list = []
    img_list = []
    for image_file in image_file_list:
        img, flag = check_and_read_gif(image_file)
        if not flag:
            img = cv2.imread(image_file)
        if img is None:
            logger.info("error in loading image:{}".format(image_file))
            continue
        valid_image_file_list.append(image_file)
        img_list.append(img)
    try:
        rec_res, predict_time = text_recognizer(img_list)
    except:
        logger.info(traceback.format_exc())
        logger.info(
            "ERROR!!!! \n"
            "Please read the FAQ：https://github.com/PaddlePaddle/PaddleOCR#faq \n"
            "If your model has tps module:  "
            "TPS does not support variable shape.\n"
            "Please set --rec_image_shape='3,32,100' and --rec_char_type='en' ")
        exit()
    for ino in range(len(img_list)):
        logger.info("Predicts of {}:{}".format(valid_image_file_list[ino],
                                               rec_res[ino]))
    logger.info("Total predict time for {} images, cost: {:.3f}".format(
        len(img_list), predict_time))


if __name__ == "__main__":
    main(utility.parse_args())