Spaces:
Sleeping
Sleeping
Upload 3 files
Browse files
app.py
ADDED
|
@@ -0,0 +1,67 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
import gradio as gr
|
| 2 |
+
import pandas as pd
|
| 3 |
+
from models import Reciept_Analyzer
|
| 4 |
+
from utils import find_product, get_info
|
| 5 |
+
import os
|
| 6 |
+
model = Reciept_Analyzer()
|
| 7 |
+
|
| 8 |
+
sample_images = []
|
| 9 |
+
for img_file in os.listdir("samples/"):
|
| 10 |
+
sample_images.append(os.path.join("samples", img_file))
|
| 11 |
+
|
| 12 |
+
def predict(image):
|
| 13 |
+
results = model.forward(image)
|
| 14 |
+
return results
|
| 15 |
+
|
| 16 |
+
|
| 17 |
+
|
| 18 |
+
# Thiết kế giao diện với Gradio
|
| 19 |
+
def create_interface():
|
| 20 |
+
with gr.Blocks() as app:
|
| 21 |
+
gr.Markdown("# Ứng dụng phân tích hóa đơn siêu thị")
|
| 22 |
+
|
| 23 |
+
with gr.Row():
|
| 24 |
+
# Cột bên trái
|
| 25 |
+
with gr.Column():
|
| 26 |
+
gr.Markdown("### Tải lên hóa đơn hoặc chọn ảnh mẫu")
|
| 27 |
+
image_input = gr.Image(label="Ảnh hóa đơn", type="filepath")
|
| 28 |
+
|
| 29 |
+
|
| 30 |
+
|
| 31 |
+
res = None
|
| 32 |
+
def on_image_selected(image_path):
|
| 33 |
+
global res
|
| 34 |
+
res = predict(image_path)
|
| 35 |
+
final = get_info(res)
|
| 36 |
+
print(res)
|
| 37 |
+
return final
|
| 38 |
+
|
| 39 |
+
def handle_input(item_name):
|
| 40 |
+
global res
|
| 41 |
+
result = find_product(item_name, res)
|
| 42 |
+
return result
|
| 43 |
+
|
| 44 |
+
|
| 45 |
+
gr.Markdown("### Ảnh mẫu")
|
| 46 |
+
example = gr.Examples(
|
| 47 |
+
inputs=image_input,
|
| 48 |
+
examples=sample_images
|
| 49 |
+
)
|
| 50 |
+
|
| 51 |
+
# Cột bên phải
|
| 52 |
+
with gr.Column():
|
| 53 |
+
result_output = gr.Textbox(label="Kết quả phân tích")
|
| 54 |
+
image_input.change(fn=on_image_selected, inputs=image_input, outputs=result_output)
|
| 55 |
+
gr.Markdown("### Tìm kiếm thông tin item")
|
| 56 |
+
item_input = gr.Textbox(label="Tên item cần tìm")
|
| 57 |
+
output = gr.Textbox(label="Kết quả")
|
| 58 |
+
|
| 59 |
+
search_button = gr.Button("Tìm kiếm")
|
| 60 |
+
search_button.click(fn=handle_input, inputs=item_input, outputs=output)
|
| 61 |
+
|
| 62 |
+
return app
|
| 63 |
+
|
| 64 |
+
|
| 65 |
+
# Chạy ứng dụng
|
| 66 |
+
app = create_interface()
|
| 67 |
+
app.launch()
|
models.py
ADDED
|
@@ -0,0 +1,169 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
import os
|
| 2 |
+
import torch
|
| 3 |
+
import numpy as np
|
| 4 |
+
from ultralytics import YOLO
|
| 5 |
+
from transformers import AutoProcessor
|
| 6 |
+
from transformers import AutoModelForTokenClassification
|
| 7 |
+
from utils import normalize_box, unnormalize_box, draw_output, create_df
|
| 8 |
+
from PIL import Image, ImageDraw
|
| 9 |
+
from vietocr.tool.predictor import Predictor
|
| 10 |
+
from vietocr.tool.config import Cfg
|
| 11 |
+
|
| 12 |
+
class Reciept_Analyzer:
|
| 13 |
+
def __init__(self,
|
| 14 |
+
processor_pretrained='microsoft/layoutlmv3-base',
|
| 15 |
+
layoutlm_pretrained=os.path.join(
|
| 16 |
+
'models', 'checkpoint'),
|
| 17 |
+
yolo_pretrained=os.path.join(
|
| 18 |
+
'models', 'best.pt'),
|
| 19 |
+
vietocr_pretrained=os.path.join(
|
| 20 |
+
'models', 'vietocr', 'vgg_seq2seq.pth')
|
| 21 |
+
):
|
| 22 |
+
|
| 23 |
+
print("Initializing processor")
|
| 24 |
+
if torch.cuda.is_available():
|
| 25 |
+
print("Using GPU")
|
| 26 |
+
else:
|
| 27 |
+
print("No GPU detected, using CPU")
|
| 28 |
+
|
| 29 |
+
self.processor = AutoProcessor.from_pretrained(
|
| 30 |
+
processor_pretrained, apply_ocr=False)
|
| 31 |
+
print("Finished initializing processor")
|
| 32 |
+
|
| 33 |
+
print("Initializing LayoutLM model")
|
| 34 |
+
self.lalm_model = AutoModelForTokenClassification.from_pretrained(
|
| 35 |
+
layoutlm_pretrained)
|
| 36 |
+
print("Finished initializing LayoutLM model")
|
| 37 |
+
|
| 38 |
+
if yolo_pretrained is not None:
|
| 39 |
+
print("Initializing YOLO model")
|
| 40 |
+
self.yolo_model = YOLO(yolo_pretrained)
|
| 41 |
+
print("Finished initializing YOLO model")
|
| 42 |
+
|
| 43 |
+
print("Initializing VietOCR model")
|
| 44 |
+
config = Cfg.load_config_from_name('vgg_seq2seq')
|
| 45 |
+
config['weights'] = vietocr_pretrained
|
| 46 |
+
config['cnn']['pretrained']= False
|
| 47 |
+
config['device'] = 'cuda:0' if torch.cuda.is_available() else 'cpu'
|
| 48 |
+
self.vietocr = Predictor(config)
|
| 49 |
+
print("Finished initializing VietOCR model")
|
| 50 |
+
|
| 51 |
+
def forward(self, img, output_path="output", is_save_cropped_img=False):
|
| 52 |
+
input_image = Image.open(img)
|
| 53 |
+
|
| 54 |
+
# detection with YOLOv8
|
| 55 |
+
bboxes = self.yolov8_det(input_image)
|
| 56 |
+
|
| 57 |
+
# sort
|
| 58 |
+
sorted_bboxes = self.sort_bboxes(bboxes)
|
| 59 |
+
|
| 60 |
+
# draw bbox
|
| 61 |
+
image_draw = input_image.copy()
|
| 62 |
+
self.draw_bbox(image_draw, sorted_bboxes, output_path)
|
| 63 |
+
|
| 64 |
+
# crop images
|
| 65 |
+
cropped_images, normalized_boxes = self.get_cropped_images(input_image, sorted_bboxes, is_save_cropped_img, output_path)
|
| 66 |
+
|
| 67 |
+
# recognition with VietOCR
|
| 68 |
+
texts, mapping_bbox_texts = self.ocr(cropped_images, normalized_boxes)
|
| 69 |
+
|
| 70 |
+
# KIE with LayoutLMv3
|
| 71 |
+
pred_texts, pred_label, boxes = self.kie(input_image, texts, normalized_boxes, mapping_bbox_texts, output_path)
|
| 72 |
+
|
| 73 |
+
# create dataframe
|
| 74 |
+
return create_df(pred_texts, pred_label)
|
| 75 |
+
|
| 76 |
+
|
| 77 |
+
def yolov8_det(self, img):
|
| 78 |
+
return self.yolo_model.predict(source=img, conf=0.3, iou=0.1)[0].boxes.xyxy.int()
|
| 79 |
+
|
| 80 |
+
def sort_bboxes(self, bboxes):
|
| 81 |
+
bbox_list = []
|
| 82 |
+
for box in bboxes:
|
| 83 |
+
tlx, tly, brx, bry = map(int, box)
|
| 84 |
+
bbox_list.append([tlx, tly, brx, bry])
|
| 85 |
+
bbox_list.sort(key=lambda x: (x[1], x[2]))
|
| 86 |
+
return bbox_list
|
| 87 |
+
|
| 88 |
+
def draw_bbox(self, image_draw, bboxes, output_path):
|
| 89 |
+
# draw bbox
|
| 90 |
+
draw = ImageDraw.Draw(image_draw)
|
| 91 |
+
for box in bboxes:
|
| 92 |
+
draw.rectangle(box, outline='red', width=2)
|
| 93 |
+
image_draw.save(os.path.join(output_path, 'bbox.jpg'))
|
| 94 |
+
print(f"Exported image with bounding boxes to {os.path.join(output_path, 'bbox.jpg')}")
|
| 95 |
+
|
| 96 |
+
def get_cropped_images(self, input_image, bboxes, is_save_cropped=False, output_path="output"):
|
| 97 |
+
normalized_boxes = []
|
| 98 |
+
cropped_images = []
|
| 99 |
+
|
| 100 |
+
# OCR
|
| 101 |
+
if is_save_cropped:
|
| 102 |
+
cropped_folder = os.path.join(output_path, "cropped")
|
| 103 |
+
if not os.path.exists(cropped_folder):
|
| 104 |
+
os.makedirs(cropped_folder)
|
| 105 |
+
i = 0
|
| 106 |
+
for box in bboxes:
|
| 107 |
+
tlx, tly, brx, bry = map(int, box)
|
| 108 |
+
normalized_box = normalize_box(box, input_image.width, input_image.height)
|
| 109 |
+
normalized_boxes.append(normalized_box)
|
| 110 |
+
cropped_ = input_image.crop((tlx, tly, brx, bry))
|
| 111 |
+
if is_save_cropped:
|
| 112 |
+
cropped_.save(os.path.join(cropped_folder, f'cropped_{i}.jpg'))
|
| 113 |
+
i += 1
|
| 114 |
+
cropped_images.append(cropped_)
|
| 115 |
+
|
| 116 |
+
return cropped_images, normalized_boxes
|
| 117 |
+
|
| 118 |
+
def ocr(self, cropped_images, normalized_boxes):
|
| 119 |
+
mapping_bbox_texts = {}
|
| 120 |
+
texts = []
|
| 121 |
+
for img, normalized_box in zip(cropped_images, normalized_boxes):
|
| 122 |
+
result = self.vietocr.predict(img)
|
| 123 |
+
text = result.strip().replace('\n', ' ')
|
| 124 |
+
texts.append(text)
|
| 125 |
+
mapping_bbox_texts[','.join(map(str, normalized_box))] = text
|
| 126 |
+
|
| 127 |
+
return texts, mapping_bbox_texts
|
| 128 |
+
|
| 129 |
+
def kie(self, img, texts, boxes, mapping_bbox_texts, output_path):
|
| 130 |
+
encoding = self.processor(img, texts,
|
| 131 |
+
boxes=boxes,
|
| 132 |
+
return_offsets_mapping=True,
|
| 133 |
+
return_tensors='pt',
|
| 134 |
+
max_length=512,
|
| 135 |
+
padding='max_length')
|
| 136 |
+
offset_mapping = encoding.pop('offset_mapping')
|
| 137 |
+
|
| 138 |
+
with torch.no_grad():
|
| 139 |
+
outputs = self.lalm_model(**encoding)
|
| 140 |
+
|
| 141 |
+
id2label = self.lalm_model.config.id2label
|
| 142 |
+
logits = outputs.logits
|
| 143 |
+
token_boxes = encoding.bbox.squeeze().tolist()
|
| 144 |
+
offset_mapping = offset_mapping.squeeze().tolist()
|
| 145 |
+
|
| 146 |
+
predictions = logits.argmax(-1).squeeze().tolist()
|
| 147 |
+
is_subword = np.array(offset_mapping)[:, 0] != 0
|
| 148 |
+
|
| 149 |
+
true_predictions = []
|
| 150 |
+
true_boxes = []
|
| 151 |
+
true_texts = []
|
| 152 |
+
for idx in range(len(predictions)):
|
| 153 |
+
if not is_subword[idx] and token_boxes[idx] != [0, 0, 0, 0]:
|
| 154 |
+
true_predictions.append(id2label[predictions[idx]])
|
| 155 |
+
true_boxes.append(unnormalize_box(
|
| 156 |
+
token_boxes[idx], img.width, img.height))
|
| 157 |
+
true_texts.append(mapping_bbox_texts.get(
|
| 158 |
+
','.join(map(str, token_boxes[idx])), ''))
|
| 159 |
+
|
| 160 |
+
if isinstance(output_path, str):
|
| 161 |
+
os.makedirs(output_path, exist_ok=True)
|
| 162 |
+
img_output = draw_output(
|
| 163 |
+
image=img,
|
| 164 |
+
true_predictions=true_predictions,
|
| 165 |
+
true_boxes=true_boxes
|
| 166 |
+
)
|
| 167 |
+
img_output.save(os.path.join(output_path, 'result.jpg'))
|
| 168 |
+
print(f"Exported result to {os.path.join(output_path, 'result.jpg')}")
|
| 169 |
+
return true_texts, true_predictions, true_boxes
|
utils.py
ADDED
|
@@ -0,0 +1,179 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
import numpy as np
|
| 2 |
+
from datasets import load_metric
|
| 3 |
+
from PIL import ImageDraw, ImageFont
|
| 4 |
+
import pandas as pd
|
| 5 |
+
|
| 6 |
+
|
| 7 |
+
metric = load_metric("seqeval")
|
| 8 |
+
|
| 9 |
+
|
| 10 |
+
def unnormalize_box(bbox, width, height):
|
| 11 |
+
return [
|
| 12 |
+
width * (bbox[0] / 1000),
|
| 13 |
+
height * (bbox[1] / 1000),
|
| 14 |
+
width * (bbox[2] / 1000),
|
| 15 |
+
height * (bbox[3] / 1000)
|
| 16 |
+
]
|
| 17 |
+
|
| 18 |
+
|
| 19 |
+
def normalize_box(bbox, width, height):
|
| 20 |
+
return [
|
| 21 |
+
int((bbox[0] / width) * 1000),
|
| 22 |
+
int((bbox[1] / height) * 1000),
|
| 23 |
+
int((bbox[2] / width) * 1000),
|
| 24 |
+
int((bbox[3] / height) * 1000)
|
| 25 |
+
]
|
| 26 |
+
|
| 27 |
+
|
| 28 |
+
def draw_output(image, true_predictions, true_boxes):
|
| 29 |
+
def iob_to_label(label):
|
| 30 |
+
label = label
|
| 31 |
+
if not label:
|
| 32 |
+
return 'other'
|
| 33 |
+
return label
|
| 34 |
+
|
| 35 |
+
# width, height = image.size
|
| 36 |
+
|
| 37 |
+
# predictions = logits.argmax(-1).squeeze().tolist()
|
| 38 |
+
# is_subword = np.array(offset_mapping)[:,0] != 0
|
| 39 |
+
# true_predictions = [id2label[pred] for idx, pred in enumerate(predictions) if not is_subword[idx]]
|
| 40 |
+
# true_boxes = [unnormalize_box(box, width, height) for idx, box in enumerate(token_boxes) if not is_subword[idx]]
|
| 41 |
+
|
| 42 |
+
# draw
|
| 43 |
+
draw = ImageDraw.Draw(image)
|
| 44 |
+
font = ImageFont.load_default()
|
| 45 |
+
|
| 46 |
+
for prediction, box in zip(true_predictions, true_boxes):
|
| 47 |
+
predicted_label = iob_to_label(prediction).lower()
|
| 48 |
+
draw.rectangle(box, outline='red')
|
| 49 |
+
draw.text((box[0] + 10, box[1] - 10),
|
| 50 |
+
text=predicted_label, fill='red', font=font)
|
| 51 |
+
|
| 52 |
+
return image
|
| 53 |
+
|
| 54 |
+
|
| 55 |
+
def create_df(true_texts,
|
| 56 |
+
true_predictions,
|
| 57 |
+
chosen_labels=['SHOP_NAME', 'ADDR', 'TITLE', 'PHONE',
|
| 58 |
+
'PRODUCT_NAME', 'AMOUNT', 'UNIT', 'UPRICE', 'SUB_TPRICE', 'UDISCOUNT',
|
| 59 |
+
'TAMOUNT', 'TPRICE', 'FPRICE', 'TDISCOUNT',
|
| 60 |
+
'RECEMONEY', 'REMAMONEY',
|
| 61 |
+
'BILLID', 'DATETIME', 'CASHIER']
|
| 62 |
+
):
|
| 63 |
+
|
| 64 |
+
data = {'text': [], 'class_label': [], 'product_id': []}
|
| 65 |
+
product_id = -1
|
| 66 |
+
for text, prediction in zip(true_texts, true_predictions):
|
| 67 |
+
if prediction not in chosen_labels:
|
| 68 |
+
continue
|
| 69 |
+
|
| 70 |
+
if prediction == 'PRODUCT_NAME':
|
| 71 |
+
product_id += 1
|
| 72 |
+
|
| 73 |
+
|
| 74 |
+
if prediction in ['AMOUNT', 'UNIT', 'UDISCOUNT', 'UPRICE', 'SUB_TPRICE',
|
| 75 |
+
'UDISCOUNT', 'TAMOUNT', 'TPRICE', 'FPRICE', 'TDISCOUNT',
|
| 76 |
+
'RECEMONEY', 'REMAMONEY']:
|
| 77 |
+
text = reformat(text)
|
| 78 |
+
|
| 79 |
+
|
| 80 |
+
if prediction in ['AMOUNT', 'SUB_TPRICE', 'UPRICE', 'PRODUCT_NAME']:
|
| 81 |
+
data['product_id'].append(product_id)
|
| 82 |
+
else:
|
| 83 |
+
data['product_id'].append('')
|
| 84 |
+
|
| 85 |
+
|
| 86 |
+
data['class_label'].append(prediction)
|
| 87 |
+
data['text'].append(text)
|
| 88 |
+
|
| 89 |
+
|
| 90 |
+
df = pd.DataFrame(data)
|
| 91 |
+
|
| 92 |
+
return df
|
| 93 |
+
|
| 94 |
+
|
| 95 |
+
def reformat(text: str):
|
| 96 |
+
try:
|
| 97 |
+
text = text.replace('.', '').replace(',', '').replace(':', '').replace('/', '').replace('|', '').replace(
|
| 98 |
+
'\\', '').replace(')', '').replace('(', '').replace('-', '').replace(';', '').replace('_', '')
|
| 99 |
+
return int(text)
|
| 100 |
+
except:
|
| 101 |
+
return text
|
| 102 |
+
|
| 103 |
+
def find_product(product_name, df):
|
| 104 |
+
product_name = product_name.lower()
|
| 105 |
+
product_df = df[df['class_label'] == 'PRODUCT_NAME']
|
| 106 |
+
mask = product_df['text'].str.lower().str.contains(product_name, case=False, na=False)
|
| 107 |
+
if mask.any():
|
| 108 |
+
product_id = product_df.loc[mask, 'product_id'].iloc[0]
|
| 109 |
+
product_info = df[df['product_id'] == product_id]
|
| 110 |
+
|
| 111 |
+
prod_name = product_info.loc[product_info['class_label'] == 'PRODUCT_NAME', 'text'].iloc[0]
|
| 112 |
+
|
| 113 |
+
try:
|
| 114 |
+
amount = product_info.loc[product_info['class_label'] == 'AMOUNT', 'text'].iloc[0]
|
| 115 |
+
except:
|
| 116 |
+
print("Error: cannot find amount")
|
| 117 |
+
amount = ''
|
| 118 |
+
|
| 119 |
+
try:
|
| 120 |
+
uprice = product_info.loc[product_info['class_label'] == 'UPRICE', 'text'].iloc[0]
|
| 121 |
+
except:
|
| 122 |
+
print("Error: cannot find unit price")
|
| 123 |
+
uprice = ''
|
| 124 |
+
|
| 125 |
+
try:
|
| 126 |
+
sub_tprice = product_info.loc[product_info['class_label'] == 'SUB_TPRICE', 'text'].iloc[0]
|
| 127 |
+
except:
|
| 128 |
+
print("Error: cannot find sub total price")
|
| 129 |
+
sub_tprice = ''
|
| 130 |
+
|
| 131 |
+
#print("Sản phẩm: ", product_info.loc[product_info['class_label'] == 'PRODUCT_NAME', 'text'].iloc[0])
|
| 132 |
+
#print("Số lượng: ", product_info.loc[product_info['class_label'] == 'AMOUNT', 'text'].iloc[0])
|
| 133 |
+
#print("Đơn giá: ", product_info.loc[product_info['class_label'] == 'UPRICE', 'text'].iloc[0])
|
| 134 |
+
#print("Thành tiền: ", product_info.loc[product_info['class_label'] == 'SUB_TPRICE', 'text'].iloc[0])
|
| 135 |
+
return f"Sản phẩm: {prod_name}\n Số lượng: {amount}\n Đơn giá: {uprice}\n Thành tiền: {sub_tprice}"
|
| 136 |
+
else:
|
| 137 |
+
#print("Không tìm thấy item nào phù hợp.")
|
| 138 |
+
return "Không tìm thấy item nào phù hợp."
|
| 139 |
+
#return result = product_df['text'].str.contains(product_name, case=False, na=False).any()
|
| 140 |
+
#return product_df[product_df['text'].str.contains(product_name, case=False, na=False)]
|
| 141 |
+
|
| 142 |
+
|
| 143 |
+
def get_info(df):
|
| 144 |
+
try:
|
| 145 |
+
shop_name = df.loc[df['class_label'] == 'SHOP_NAME', 'text'].iloc[0]
|
| 146 |
+
except:
|
| 147 |
+
print("Error: cannot find shop name")
|
| 148 |
+
shop_name = ''
|
| 149 |
+
print("Tên siêu thị: ", shop_name)
|
| 150 |
+
|
| 151 |
+
try:
|
| 152 |
+
addr = df.loc[df['class_label'] == 'ADDR', 'text'].iloc[0]
|
| 153 |
+
except:
|
| 154 |
+
print("Error: cannot find address")
|
| 155 |
+
addr = ''
|
| 156 |
+
print("Địa chỉ: ", addr)
|
| 157 |
+
|
| 158 |
+
try:
|
| 159 |
+
bill_id = df.loc[df['class_label'] == 'BILLID', 'text'].iloc[0]
|
| 160 |
+
except:
|
| 161 |
+
print("Error: cannot find bill id")
|
| 162 |
+
bill_id = ''
|
| 163 |
+
print("ID hóa đơn: ", bill_id)
|
| 164 |
+
|
| 165 |
+
try:
|
| 166 |
+
date_time = df.loc[df['class_label'] == 'DATETIME', 'text'].iloc[0]
|
| 167 |
+
except:
|
| 168 |
+
print("Error: cannot find date and time")
|
| 169 |
+
date_time = ''
|
| 170 |
+
print("Ngày: ", date_time)
|
| 171 |
+
|
| 172 |
+
try:
|
| 173 |
+
cashier = df.loc[df['class_label'] == 'CASHIER', 'text'].iloc[0]
|
| 174 |
+
except:
|
| 175 |
+
print("Error: cannot find cashier")
|
| 176 |
+
cashier = ''
|
| 177 |
+
print("Nhân viên: ", cashier)
|
| 178 |
+
|
| 179 |
+
return f"Tên siêu thị: {shop_name}\n Địa chỉ: {addr}\n ID hóa đơn: {bill_id}\n Ngày: {date_time}\n Nhân viên: {cashier}\n"
|