File size: 5,975 Bytes
e7bb669
 
 
f5065e2
 
e7bb669
f5065e2
 
 
 
 
 
 
 
 
 
 
 
e7bb669
f5065e2
e7bb669
7e87f71
e7bb669
f5065e2
 
 
e7bb669
f5065e2
 
e7bb669
 
f5065e2
e7bb669
f5065e2
 
 
e7bb669
f5065e2
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
e7bb669
f5065e2
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
e7bb669
 
 
f5065e2
e7bb669
f5065e2
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
import sys
import json
import re
from typing import List, Dict, Any, Optional

from PIL import Image
import pytesseract
from transformers import AutoTokenizer, AutoModelForSequenceClassification, pipeline

CLS_MODEL_ID = "MoritzLaurer/mDeBERTa-v3-base-mnli-xnli"

classifier = pipeline(
    "zero-shot-classification",
    model=CLS_MODEL_ID,
    tokenizer=CLS_MODEL_ID,
    device=0 if hasattr(sys, "gettrace") is False else -1,
)


def ocr_image_to_text(image: Image.Image) -> str:

    image = image.convert("L")

    config = r"--psm 6"
    text = pytesseract.image_to_string(image, lang="rus", config=config)
    return text


def pick_total_from_text(text: str) -> Optional[float]:
    if not text:
        return None

    text = text.replace("\xa0", " ")
    text = re.sub(r"[₽€$]", " ", text)

    def _to_float(s: str) -> Optional[float]:
        s = s.replace(" ", "").replace(",", ".")
        try:
            return float(s)
        except Exception:
            return None

    pattern_num = r"(-?\d{1,3}(?:[ .,\u00A0]?\d{3})*(?:[.,]\d{2}))"

    strong_candidates = []
    medium_candidates = []
    all_candidates = []
    line_totals = []

    for line in text.splitlines():
        line_clean = line.strip()
        if not line_clean:
            continue
        lower = line_clean.lower()

        if "сдач" in lower:
            continue

        m_eq = re.search(r"=\s*([0-9][0-9 .,\u00A0]*[.,]\d{2})", line_clean)
        if m_eq:
            v_eq = _to_float(m_eq.group(1))
            if v_eq and 0 < v_eq < 1e7:
                line_totals.append(v_eq)

        nums = re.findall(pattern_num, line_clean)
        if not nums:
            continue

        for m in nums:
            v = _to_float(m)
            if not v or v <= 0 or v > 1e7:
                continue

            if any(k in lower for k in ["итог", "итого", "к оплате", "всего к оплате"]):
                strong_candidates.append(v)
            elif any(k in lower for k in ["наличн", "карта", "безнал", "оплачено"]):
                medium_candidates.append(v)

            all_candidates.append(v)

    if strong_candidates:
        return max(strong_candidates)

    if medium_candidates:
        return max(medium_candidates)

    if len(line_totals) >= 3:
        s = sum(line_totals)
        if 0 < s < 1e7:
            return s

    if all_candidates:
        return max(all_candidates)

    return None

def classify_category_zeroshot(
    receipt_text: str,
    categories: List[Dict[str, Any]],
) -> Dict[str, Any]:
    if not categories:
        return {"id": None, "name": None}

    receipt_short = receipt_text[:1500]
    labels = [cat["name"] for cat in categories]

    result = classifier(
        receipt_short,
        candidate_labels=labels,
        multi_label=False,
        hypothesis_template="Это покупка по категории {}.",
    )

    best_label = result["labels"][0]

    best_cat = next((c for c in categories if c["name"] == best_label), None)

    if best_cat is None:
        best_label_low = best_label.lower()
        for c in categories:
            if c["name"].lower() == best_label_low:
                best_cat = c
                break

    if best_cat is None:
        best_cat = categories[-1]

    return best_cat

def guess_shop_name(receipt_text: str) -> Optional[str]:
    lines = [ln.strip() for ln in receipt_text.splitlines() if ln.strip()]
    top = lines[:5]

    candidates = []
    for ln in top:
        lower = ln.lower()
        if any(x in lower for x in ["инн", "ккт", "касса", "рн кк", "россия", "г. ", "ул."]):
            continue
        if 2 <= len(ln) <= 40:
            candidates.append(ln)

    if candidates:
        return candidates[0]
    return None


def build_description(
    receipt_text: str,
    category_name: Optional[str],
    total: Optional[float],
) -> str:
    cat = category_name or "покупка"
    shop = guess_shop_name(receipt_text)

    if shop:
        return f"Покупка по категории {cat} в {shop}"
    else:
        if total is not None:
            return f"Покупка по категории {cat} на {total:.2f}"
        else:
            return f"Покупка по категории {cat}"

DEFAULT_CATEGORIES: List[Dict[str, Any]] = [
    {"id": 1, "name": "Еда"},
    {"id": 2, "name": "Спорт"},
    {"id": 3, "name": "Обучение"},
    {"id": 4, "name": "Транспорт"},
    {"id": 5, "name": "Развлечения"},
    {"id": 6, "name": "Медицина"},
    {"id": 7, "name": "Бытовые товары"},
    {"id": 8, "name": "Прочее"},
]


def extract_info(
    image_path: str,
    categories: Optional[List[Dict[str, Any]]] = None,
) -> Dict[str, Any]:

    if categories is None:
        categories = DEFAULT_CATEGORIES

    image = Image.open(image_path).convert("RGB")
    text = ocr_image_to_text(image)

    total = pick_total_from_text(text)
    best_cat = classify_category_zeroshot(text, categories)
    description = build_description(text, best_cat["name"], total)

    return {
        "total": total,
        "category_id": best_cat.get("id"),
        "category_name": best_cat.get("name"),
        "description": description,
        "raw_text": text,
    }


if __name__ == "__main__":
    if len(sys.argv) < 2:
        print("Usage: receipt_info_api.py path/to/receipt.jpg [categories.json]", file=sys.stderr)
        sys.exit(1)

    image_path = sys.argv[1]

    cats = None
    if len(sys.argv) >= 3:
        with open(sys.argv[2], "r", encoding="utf-8") as f:
            cats = json.load(f)

    info = extract_info(image_path, categories=cats)

    out = {
        "total": info["total"],
        "category_id": info["category_id"],
        "category_name": info["category_name"],
        "description": info["description"],
    }
    print(json.dumps(out, ensure_ascii=False))