import gzip
import requests
from lxml import etree as ET
from deep_translator import GoogleTranslator
import time
from pathlib import Path

# Translation setup
TRANSLATOR = GoogleTranslator(source='auto', target='en')
BATCH_SIZE = 10  # Process titles in batches for better performance

def translate_batch(texts):
    """Translate a batch of texts more efficiently"""
    try:
        # Join texts with special separator (to split back later)
        combined = " ||| ".join(texts)
        translated = TRANSLATOR.translate(combined)
        return translated.split(" ||| ")
    except Exception as e:
        print(f"Batch translation failed: {e}")
        return texts  # Return originals if batch fails

def translate_titles(titles):
    """Process titles in batches with priority"""
    translated_titles = []
    
    # First pass: try all as Hebrew (most Bollywood titles in this EPG are Hebrew)
    hebrew_batch = []
    for title in titles:
        hebrew_batch.append(title)
        if len(hebrew_batch) >= BATCH_SIZE:
            translated_titles.extend(translate_batch(hebrew_batch))
            hebrew_batch = []
            time.sleep(0.3)  # Short delay between batches
    
    # Process remaining
    if hebrew_batch:
        translated_titles.extend(translate_batch(hebrew_batch))
    
    # Second pass: retry any that weren't translated (fallback to auto)
    for i, title in enumerate(titles):
        if translated_titles[i] == title:
            try:
                translated_titles[i] = TRANSLATOR.translate(title)
                time.sleep(0.1)
            except:
                pass
    
    return translated_titles

# Download and parse EPG
print("Downloading EPG data...")
url = "https://epgshare01.online/epgshare01/epg_ripper_IL1.xml.gz"
response = requests.get(url, stream=True)
response.raise_for_status()

with gzip.open(response.raw) as f:
    xml_bytes = f.read().replace(b'xmlns="http://xmltv.org/xmltv.dtd"', b'')

root = ET.fromstring(xml_bytes)
target_id = "Bollywood.Movies.il"
channel = root.xpath(f"//channel[@id='{target_id}']")[0]
programs = root.xpath(f"//programme[@channel='{target_id}']")

# Extract all titles first for batch processing
print("Extracting titles for batch translation...")
all_titles = [program.find("title").text for program in programs if program.find("title") is not None]

# Translate all titles in optimized batches
print(f"Translating {len(all_titles)} titles...")
translated_titles = translate_titles(all_titles)

# Build new EPG
print("Building translated EPG...")
new_root = ET.Element("tv")
new_root.append(channel)

for i, program in enumerate(programs):
    new_program = ET.SubElement(new_root, "programme")
    # Copy attributes
    for attr, value in program.items():
        new_program.set(attr, value)
    
    # Add translated title
    title_elem = program.find("title")
    if title_elem is not None:
        new_title = ET.SubElement(new_program, "title", lang="en")
        new_title.text = translated_titles[i] if i < len(translated_titles) else title_elem.text
    
    # Add other elements
    for elem in program:
        if elem.tag == "desc":
            new_desc = ET.SubElement(new_program, "desc", lang="en")
            new_desc.text = TRANSLATOR.translate(elem.text) if elem.text else ""
            time.sleep(0.1)
        elif elem.tag != "title":
            new_elem = ET.SubElement(new_program, elem.tag)
            new_elem.text = elem.text if elem.text else ""
            for attr, value in elem.items():
                new_elem.set(attr, value)

# Save compact XML

output_dir = Path(__file__).parent / "tempest_config" / "epg"
output_file = output_dir / "yes_bollywood.xml"

print("Saving EPG file...")
with open(output_file, "wb") as f:
    f.write(b'<?xml version="1.0" encoding="utf-8"?>\n<tv>\n')
    f.write(ET.tostring(new_root[0], encoding='utf-8') + b'\n')
    for program in new_root[1:]:
        f.write(ET.tostring(program, encoding='utf-8') + b'\n')
    f.write(b'</tv>')

print("✅ Translation complete! Saved to 'yes_bollywood.xml'")