backend/modules/text_checks.py

import re
import emoji
from PIL import Image
from backend import config
from backend.utils import get_roi, clean_text, are_strings_similar, blur_image, is_blank, is_english, is_valid_english, destroy_text_roi
from backend.model_handler import model_handler

def is_unreadable_tagline(htag, tag):
    clean_htag = clean_text(htag)
    clean_tag = clean_text(tag)
    return not are_strings_similar(clean_htag, clean_tag)

def is_hyperlink_tagline(tag):
    substrings = ['www', '.com', 'http']
    return any(sub in tag for sub in substrings)

def is_price_tagline(tag):
    exclude_keywords = ["crore", "thousand", "million", "billion", "trillion"]
    exclude_pattern = r'(₹\.?\s?\d+\s*(lac|lacs|lakh|lakhs|cr|k))|(\brs\.?\s?\d+\s*(lac|lacs|lakh|lakhs|cr|k))|(\$\.?\s?\d+\s*(lac|lacs|lakh|lakhs|cr|k))'
    price_pattern = r'(₹\s?\d+)|(\brs\.?\s?\d+)|(\$\s?\d+)|(र\d+)'
    
    if any(keyword in tag for keyword in exclude_keywords):
        return False
    if re.search(exclude_pattern, tag):
        return False
    return bool(re.search(price_pattern, tag))

def is_multiple_emoji(emoji_text):
    words = emoji_text.split()
    last_word = words[-1]
    return last_word not in ['0', '1']

def is_incomplete_tagline(tag, is_eng):
    tag = emoji.replace_emoji(tag, '')
    tag = tag.strip()
    if tag.endswith(('...', '..')):
        return True
    if not is_eng and tag.endswith(('.')):
        return True
    return False

def tagline(image_path):
    results = {
        "Empty/Illegible/Black Tagline": 0,
        "Multiple Taglines": 0,
        "Incomplete Tagline": 0,
        "Hyperlink": 0,
        "Price Tag": 0,
        "Excessive Emojis": 0
    }

    image = get_roi(image_path, *config.TAG)
    himage = blur_image(image, 0.3)
    easytag = model_handler.easyocr_ocr(image).lower().strip()
    unr = model_handler.easyocr_ocr(himage).lower().strip()

    if is_blank(easytag) or is_blank(unr):
        results["Empty/Illegible/Black Tagline"] = 1
        return results

    is_eng = is_english(easytag)
    if not is_eng:
        results["Empty/Illegible/Black Tagline"] = 0
        tag = easytag
    else:
        Tag = model_handler.intern(image, config.PTAG, 25).strip()
        tag = Tag.lower()

        htag = model_handler.intern(himage, config.PTAG, 25).lower().strip()
        if is_unreadable_tagline(htag, tag):
            results["Empty/Illegible/Black Tagline"] = 1
            
    results["Incomplete Tagline"] = 1 if is_incomplete_tagline(tag, is_eng) else 0
    results["Hyperlink"] = 1 if is_hyperlink_tagline(tag) else 0
    results["Price Tag"] = 1 if is_price_tagline(tag) else 0

    imagedt = get_roi(image_path, *config.DTAG)
    dtag = model_handler.easyocr_ocr(imagedt).strip()
    results["Multiple Taglines"] = 0 if is_blank(dtag) else 1

    emoji_resp = model_handler.intern(image, config.PEMO, 100)
    results["Excessive Emojis"] = 1 if is_multiple_emoji(emoji_resp) else 0

    return results

def cta(image_path):
    image = get_roi(image_path, *config.CTA)
    cta_text = model_handler.intern(image, config.PTAG, 5).strip()
    veng = is_valid_english(cta_text)
    eng = is_english(cta_text)

    if '.' in cta_text or '..' in cta_text or '...' in cta_text:
        return {"Bad CTA": 1}

    if any(emoji.is_emoji(c) for c in cta_text):
        return {"Bad CTA": 1}

    clean_cta_text = clean_text(cta_text)
    # print(len(clean_cta_text)) # Removed print

    if eng and len(clean_cta_text) <= 2:
        return {"Bad CTA": 1}

    if len(clean_cta_text) > 15:
        return {"Bad CTA": 1}

    return {"Bad CTA": 0}

def tnc(image_path):
    image = get_roi(image_path, *config.TNC)
    tnc_text = model_handler.easyocr_ocr(image)
    clean_tnc = clean_text(tnc_text)

    return {"Terms & Conditions": 0 if is_blank(clean_tnc) else 1}

def tooMuchText(image_path):
    DRIB = (0.04, 0.625, 1.0, 0.677)
    DUP = (0, 0, 1.0, 0.25)
    DBEL = (0, 0.85, 1.0, 1)
    image = Image.open(image_path).convert('RGB')
    image = destroy_text_roi(image, *DRIB)
    image = destroy_text_roi(image, *DUP)
    image = destroy_text_roi(image, *DBEL)
    bd = model_handler.easyocr_ocr(image).lower().strip()
    return {"Too Much Text": 1 if len(bd) > 55 else 0}