main.py

from transformers import pipeline

# Create translation pipeline
translator = pipeline("translation", model="facebook/nllb-200-3.3B")

# Translate English to Zulu (you prepend the target language code in input)
input_text = ">>zul_Latn<< Hello, how are you?"

result = translator(input_text)
print(result[0]["translation_text"])


# import requests
# import gradio as gr
# from dotenv import load_dotenv
# import os

# # Load environment variables from .env file
# load_dotenv()
# HF_TOKEN = os.getenv("HF_TOKEN")

# model_name = "Helsinki-NLP/opus-mt-en-nso"
# API_URL = f"https://api-inference.huggingface.co/models/{model_name}"
# headers = {"Authorization": f"Bearer {HF_TOKEN}"}


# def query(payload):
#     # HTTP POST Request
#     response = requests.post(API_URL, headers=headers, json=payload)
#     return response.json()


# def translate(input_text):
#     # API Request:
#     response = query({"inputs": input_text, "options": {"wait_for_model": True}})

#     translation = response[0]["translation_text"]

#     return translation


# translator = gr.Interface(
#     fn=translate,
#     inputs=[gr.Textbox(label="Input Text", placeholder="Input Text To Be Translated")],
#     outputs=gr.Textbox(label="Translation"),
#     title="Translademia",
# )

# translator.launch()


# # The one we are going with
# import requests
# import gradio as gr
# from dotenv import load_dotenv
# import os

# # Load environment variables
# load_dotenv()
# HF_TOKEN = os.getenv("HF_TOKEN")
# headers = {"Authorization": f"Bearer {HF_TOKEN}"}

# # Language to ISO 639-3 codes (used for NLLB-200)
# LANGUAGES = {
#     "English → Afrikaans": "afr",
#     "English → Xhosa": "xho",
#     "English → Zulu": "zul",
#     "English → Sesotho": "sot",
#     "English → Tswana": "tsn",
#     "English → Northern Sotho": "nso",
#     "English → Swati": "ssw",
#     "English → Tsonga": "tso",
#     "English → Venda": "ven",
# }

# MODEL_NAME = "facebook/nllb-200-distilled-600M"
# API_URL = f"https://api-inference.huggingface.co/models/{MODEL_NAME}"


# def query(payload):
#     response = requests.post(API_URL, headers=headers, json=payload)

#     if response.status_code != 200:
#         print(f"[ERROR] API failed: {response.status_code} - {response.text}")
#         return {"error": f"Request failed with {response.status_code}"}

#     try:
#         return response.json()
#     except requests.exceptions.JSONDecodeError:
#         print(f"[ERROR] Failed to parse JSON: {response.text}")
#         return {"error": "Invalid JSON from API"}


# def translate(input_text, language_label):
#     language_code = LANGUAGES[language_label]
#     formatted_input = f">>{language_code}<< {input_text}"

#     response = query({"inputs": formatted_input, "options": {"wait_for_model": True}})

#     if "error" in response:
#         return f"Error: {response['error']}"

#     return response[0]["translation_text"]


# translator = gr.Interface(
#     fn=translate,
#     inputs=[
#         gr.Textbox(label="Input Text", placeholder="Type text here..."),
#         gr.Dropdown(list(LANGUAGES.keys()), label="Select Language Target"),
#     ],
#     outputs=gr.Textbox(label="Translation"),
#     title="Translademia",
#     description="Translate English text to South African languages using Meta's NLLB-200 model.",
# )

# translator.launch()


# love

# import os
# from huggingface_hub import InferenceClient
# import gradio as gr
# from dotenv import load_dotenv

# # Load env
# load_dotenv()
# HF_TOKEN = os.getenv("HF_TOKEN")

# # Init client
# client = InferenceClient(token=HF_TOKEN)

# # Languages supported
# LANGUAGES = {
#     "English → Afrikaans": "afr",
#     "English → Xhosa": "xho",
#     "English → Zulu": "zul",
#     "English → Sesotho": "sot",
#     "English → Tswana": "tsn",
#     "English → Northern Sotho": "nso",
#     "English → Swati": "ssw",
#     "English → Tsonga": "tso",
#     "English → Venda": "ven",
# }

# MODEL_NAME = "facebook/nllb-200-distilled-600M"


# def translate(input_text: str, language_label: str) -> str:
#     if not input_text.strip():
#         return "Error: Please enter text to translate."

#     lang_code = LANGUAGES[language_label]
#     formatted_input = f">>{lang_code}<< {input_text}"

#     try:
#         response = client.text_generation(
#             prompt=formatted_input,
#             model=MODEL_NAME,
#             max_new_tokens=200,
#         )
#         return response.strip()
#     except Exception as e:
#         return f"Error: {str(e)}"


# # Gradio UI
# translator = gr.Interface(
#     fn=translate,
#     inputs=[
#         gr.Textbox(label="Input Text", placeholder="Type English text here..."),
#         gr.Dropdown(list(LANGUAGES.keys()), label="Target Language"),
#     ],
#     outputs=gr.Textbox(label="Translation"),
#     title="NLLB-200 Translator",
#     description="Translate English to South African languages using Meta's NLLB model",
# )

# translator.launch()


# hate


# import requests
# import gradio as gr
# from dotenv import load_dotenv
# import os

# # Load Hugging Face token from .env
# load_dotenv()
# HF_TOKEN = os.getenv("HF_TOKEN")
# headers = {"Authorization": f"Bearer {HF_TOKEN}"}

# # NLLB model name
# MODEL_NAME = "facebook/nllb-200-3.3B"
# API_URL = f"https://api-inference.huggingface.co/models/{MODEL_NAME}"

# # Define supported language pairs and NLLB codes
# LANGUAGE_PAIRS = {
#     "English → Afrikaans": ("eng_Latn", "afr_Latn"),
#     "English → Xhosa": ("eng_Latn", "xho_Latn"),
#     "English → Zulu": ("eng_Latn", "zul_Latn"),
#     "English → Sesotho": ("eng_Latn", "sot_Latn"),
#     "English → Tswana": ("eng_Latn", "tsn_Latn"),
#     "English → Northern Sotho": ("eng_Latn", "nso_Latn"),
#     "English → Swati": ("eng_Latn", "ssw_Latn"),
#     "English → Tsonga": ("eng_Latn", "tso_Latn"),
#     "Afrikaans → English": ("afr_Latn", "eng_Latn"),
#     "Xhosa → English": ("xho_Latn", "eng_Latn"),
#     "Zulu → English": ("zul_Latn", "eng_Latn"),
#     "Sesotho → English": ("sot_Latn", "eng_Latn"),
#     "Tswana → English": ("tsn_Latn", "eng_Latn"),
#     "Northern Sotho → English": ("nso_Latn", "eng_Latn"),
#     "Swati → English": ("ssw_Latn", "eng_Latn"),
#     "Tsonga → English": ("tso_Latn", "eng_Latn"),
# }


# def translate(input_text, language_pair):
#     src_lang, tgt_lang = LANGUAGE_PAIRS[language_pair]

#     payload = {
#         "inputs": input_text,
#         "parameters": {
#             "src_lang": src_lang,
#             "tgt_lang": tgt_lang,
#         },
#         "options": {"wait_for_model": True},
#     }

#     response = requests.post(API_URL, headers=headers, json=payload)

#     if response.status_code != 200:
#         return f"[ERROR] {response.status_code}: {response.text}"

#     try:
#         output = response.json()
#         return output[0]["translation_text"]
#     except Exception as e:
#         return f"[ERROR] Failed to parse response: {e}"


# # Gradio UI
# translator = gr.Interface(
#     fn=translate,
#     inputs=[
#         gr.Textbox(label="Input Text", placeholder="Type text here..."),
#         gr.Dropdown(choices=list(LANGUAGE_PAIRS.keys()), label="Select Language Pair"),
#     ],
#     outputs=gr.Textbox(label="Translation"),
#     title="Translademia (NLLB Edition)",
#     description="Translate between English and official South African languages using Meta's NLLB-200 model.",
# )

# translator.launch(share=True)


# ///////////////////////////////////////////////////////////////////////////////////////////////////////////////
# Using Unesco API

# import requests
# import gradio as gr
# from dotenv import load_dotenv
# import os

# # Load Hugging Face token from .env
# load_dotenv()
# HF_TOKEN = os.getenv("HF_TOKEN")
# headers = {"Authorization": f"Bearer {HF_TOKEN}"}

# # NLLB model endpoint
# MODEL_NAME = "facebook/nllb-200-3.3B"
# API_URL = f"https://api-inference.huggingface.co/models/{MODEL_NAME}"

# # Define supported language pairs and NLLB codes
# LANGUAGE_PAIRS = {
#     "English → Afrikaans": ("eng_Latn", "afr_Latn"),
#     "English → Xhosa": ("eng_Latn", "xho_Latn"),
#     "English → Zulu": ("eng_Latn", "zul_Latn"),
#     "English → Sesotho": ("eng_Latn", "sot_Latn"),
#     "English → Tswana": ("eng_Latn", "tsn_Latn"),
#     "English → Northern Sotho": ("eng_Latn", "nso_Latn"),
#     "English → Swati": ("eng_Latn", "ssw_Latn"),
#     "English → Tsonga": ("eng_Latn", "tso_Latn"),
#     "Afrikaans → English": ("afr_Latn", "eng_Latn"),
#     "Xhosa → English": ("xho_Latn", "eng_Latn"),
#     "Zulu → English": ("zul_Latn", "eng_Latn"),
#     "Sesotho → English": ("sot_Latn", "eng_Latn"),
#     "Tswana → English": ("tsn_Latn", "eng_Latn"),
#     "Northern Sotho → English": ("nso_Latn", "eng_Latn"),
#     "Swati → English": ("ssw_Latn", "eng_Latn"),
#     "Tsonga → English": ("tso_Latn", "eng_Latn"),
# }


# def translate(input_text, language_pair):
#     if not input_text.strip():
#         return "[ERROR] Please enter some text to translate."

#     # Get source and target language codes
#     src_lang, tgt_lang = LANGUAGE_PAIRS[language_pair]

#     # Prepend target language token to the input
#     formatted_input = f">>{tgt_lang}<< {input_text.strip()}"

#     # Send request to Hugging Face Inference API
#     payload = {
#         "inputs": formatted_input,
#         "options": {"wait_for_model": True},
#     }

#     response = requests.post(API_URL, headers=headers, json=payload)

#     if response.status_code != 200:
#         return f"[ERROR] {response.status_code}: {response.text}"

#     try:
#         output = response.json()
#         return output[0]["translation_text"]
#     except Exception as e:
#         return f"[ERROR] Failed to parse response: {e}"


# # Gradio UI
# translator = gr.Interface(
#     fn=translate,
#     inputs=[
#         gr.Textbox(label="Input Text", placeholder="Type text here..."),
#         gr.Dropdown(choices=list(LANGUAGE_PAIRS.keys()), label="Select Language Pair"),
#     ],
#     outputs=gr.Textbox(label="Translation"),
#     title="Translademia (NLLB Edition)",
#     description="Translate between English and South African languages using Meta's NLLB-200 multilingual model.",
# )

# translator.launch(share=True)