Detekcia jazyka z výstupu AIO z Marketing Mineru

import pandas as pd
from langdetect import detect, DetectorFactory

DetectorFactory.seed = 0

# Načítanie Excel súboru
input_file = "stiahnuty-subor-z-mm.xlsx"  # zmeň na názov svojho súboru
df = pd.read_excel(input_file)
df.columns = df.columns.str.strip()

# Prioritné URL fragmenty -> jazyk
url_lang_map = {
    "/en/": "en",
    "/cz/": "cs",
    "/cs/": "cs",
    "/sk/": "sk",
    "/pl/": "pl",
    "/hu/": "hu",
    "/de/": "de",
    "/fr/": "fr",
    "/it/": "it",
    "/es/": "es",
    "/ro/": "ro",
    "/bg/": "bg",
    "/gr/": "el",
    "/nl/": "nl",
    "/se/": "sv",
    "/no/": "no",
    "/fi/": "fi",
    "/pt/": "pt",
    "/ru/": "ru",
    "/cn/": "zh",
    "/jp/": "ja",
    "/kr/": "ko",
    "/ae/": "ar",
    "/ro/": "ro",
    "/sa/": "ar"
}

# TLD -> jazyk, použije sa len ak je URL fragment nenájdený
tld_lang_map = {
    ".cz": "cs",
    ".sk": "sk",
    ".pl": "pl",
    ".hu": "hu",
    ".de": "de",
    ".fr": "fr",
    ".it": "it",
    ".es": "es",
    ".ro": "ro",
    ".bg": "bg",
    ".gr": "el",
    ".nl": "nl",
    ".se": "sv",
    ".no": "no",
    ".fi": "fi",
    ".pt": "pt",
    ".ru": "ru",
    ".cn": "zh",
    ".jp": "ja",
    ".kr": "ko",
    ".ae": "ar",
    ".sa": "ar",
    ".com": "en",
    ".org": "en",
    ".net": "en"
}

def detect_language(title, url):
    if not isinstance(title, str) or len(title.strip()) == 0:
        return "unknown"

    url = str(url).lower()  # lowercase všetko

    # Priorita: URL fragmenty
    for key, lang in url_lang_map.items():
        if key in url:
            return lang

    # Fallback: TLD
    for key, lang in tld_lang_map.items():
        if key in url:
            return lang

    # Fallback: langdetect z titulku
    try:
        lang = detect(title)
        # krátke titulky môžu byť nepresné
        if len(title.split()) < 3 and lang not in tld_lang_map.values():
            return "unknown"
        return lang
    except:
        return "unknown"

# Vytvorenie stĺpca Language
df['Language'] = df.apply(lambda row: detect_language(row['Title'], row['URL']), axis=1)

# Uloženie do nového Excel súboru
output_file = "nazov-vystupneho-suboru.xlsx"
df.to_excel(output_file, index=False)
print(f"Hotovo! Súbor uložený: {output_file}")