app1.py

import random
import re
import gradio as gr
import emoji

# -------------------------
# Config
# -------------------------

ELIGIBLE = ("NOUN", "ADJ")  # for clarity; used by our simple tagger

WORD2EMOJI = {
    # Nouns
    "heart":"❤️","love":"❤️","car":"🚗","bike":"🚲","bicycle":"🚲","bus":"🚌","train":"🚆","plane":"✈️",
    "pizza":"🍕","burger":"🍔","coffee":"☕","tea":"🍵","book":"📚","laptop":"💻","computer":"🖥️",
    "phone":"📱","camera":"📷","money":"💰","cash":"💵","rain":"🌧️","sun":"☀️","moon":"🌙","star":"⭐",
    "fire":"🔥","party":"🎉","gift":"🎁","music":"🎵","dog":"🐶","cat":"🐱","bird":"🐦","flower":"🌸",
    "tree":"🌳","house":"🏠","home":"🏠","school":"🏫","office":"🏢","hospital":"🏥","food":"🍽️",
    "cake":"🍰","cookie":"🍪","chocolate":"🍫","icecream":"🍨","beach":"🏖️","mountain":"⛰️","city":"🏙️",
    "rocket":"🚀","earth":"🌍","globe":"🌍","world":"🌍","mail":"✉️","letter":"✉️","football":"🏈",
    "soccer":"⚽","basketball":"🏀","tennis":"🎾","medal":"🏅","trophy":"🏆","alarm":"⏰","clock":"⏰",
    "idea":"💡","lightbulb":"💡","check":"✅","cross":"❌","warning":"⚠️",
    # Adjectives
    "happy":"😊","sad":"😢","angry":"😡","tired":"😴","sleepy":"😴","cool":"😎","loveable":"🥰",
    "lovely":"🥰","beautiful":"😍","pretty":"😍","handsome":"🤵","smart":"🧠","brilliant":"🧠",
    "fast":"⚡","quick":"⚡","slow":"🐢","hot":"🔥","cold":"❄️","sweet":"🍬","salty":"🧂",
    "spicy":"🌶️","fresh":"🫧","strong":"💪","weak":"🫤","shiny":"✨","clean":"🫧","dirty":"🧹",
    "busy":"📈","rich":"💰","poor":"🪙","lucky":"🍀","fun":"🎉","romantic":"💘","scary":"😱"
}

ADJECTIVE_WORDS = {
    "happy","sad","angry","tired","sleepy","cool","loveable","lovely","beautiful","pretty",
    "handsome","smart","brilliant","fast","quick","slow","hot","cold","sweet","salty","spicy",
    "fresh","strong","weak","shiny","clean","dirty","busy","rich","poor","lucky","fun","romantic","scary"
}

# Heuristic suffixes
_ADJ_SUFFIXES = ("y","ive","ful","less","ous","al","ic","able","ible","ish")
_NOUN_SUFFIXES = ("tion","ment","ness","ity","er","or","ship","age","ance","ence","ism","ist")

# -------------------------
# Utilities
# -------------------------

def try_emojize_alias(word):
    """Try alias-based emojization without using 'strict' (for older emoji libs)."""
    lemma = word.lower()
    for cand in (f":{lemma}:", f":{lemma}_face:", f":{lemma}_emoji:", f":{lemma}_button:"):
        try:
            em = emoji.emojize(cand, language="alias")
        except TypeError:
            # very old emoji versions: no language kw; still OK
            em = emoji.emojize(cand)
        if em != cand:
            return em
    return None

def tokenize(text):
    # Keep punctuation as separate tokens; include emojis as tokens
    return re.findall(r"\w+|[^\w\s]", text, flags=re.UNICODE)

def simple_tag(tokens):
    """Return list of (token, tag) where tag is 'NOUN', 'ADJ', or 'OTHER'."""
    tagged = []
    for t in tokens:
        # Non-letters (punct/emoji) -> OTHER
        if not re.search(r"[A-Za-z]", t):
            tagged.append((t, "OTHER"))
            continue

        low = t.lower()
        if low in WORD2EMOJI:
            tagged.append((t, "ADJ" if low in ADJECTIVE_WORDS else "NOUN"))
            continue

        # Heuristics
        if low.endswith(_ADJ_SUFFIXES):
            tagged.append((t, "ADJ"))
        elif low.endswith(_NOUN_SUFFIXES) or t[:1].isupper():
            tagged.append((t, "NOUN"))
        else:
            tagged.append((t, "OTHER"))
    return tagged

# -------------------------
# Core function
# -------------------------

def translate_text(text, mode="Replace", replacement_rate=1.0, keep_case=True, dedupe=True):
    try:
        text = (text or "").strip()
        if not text:
            return ""

        tokens = tokenize(text)
        tagged = simple_tag(tokens)

        out = []
        for tok, tag in tagged:
            eligible = (tag in ELIGIBLE) and re.search(r"[A-Za-z]", tok)
            if eligible and random.random() <= replacement_rate:
                key = tok.lower()
                em = WORD2EMOJI.get(key) or try_emojize_alias(key)
                if em:
                    if mode == "Replace":
                        out.append(em)
                    else:
                        out.append(f"{tok} {em}" if not dedupe else f"{tok} {em}")
                    continue
            out.append(tok)

        # No space before , . ! ? ; : ) ] }
        result = ""
        for i, t in enumerate(out):
            result += t
            if i < len(out) - 1:
                nxt = out[i + 1]
                if not re.match(r"^[,\.!\?;:\)\]\}]+$", nxt):
                    result += " "

        if not keep_case:
            result = re.sub(r"[A-Za-z]+", lambda m: m.group(0).lower(), result)

        return result
    except Exception as e:
        # Show error message in UI instead of throwing
        return f"[Error in translate_text: {e}]"

# -------------------------
# UI
# -------------------------

EXAMPLES = [
    "I love spicy pizza and hot coffee on a rainy day.",
    "The smart dog chased a happy cat around the house.",
    "We had a beautiful view of the city from the mountain.",
    "She wrote a brilliant idea in her notebook.",
    "This is a strong team with a cool product.",
]

with gr.Blocks(title="Emoji Translator 😎➡️🙂🍕🚀") as demo:
    gr.Markdown(
        """
# Emoji Translator 😎➡️🙂🍕🚀
Type a sentence — I’ll replace **nouns/adjectives** with matching emojis.

**Mode**
- **Replace**: swap eligible words with emojis  
- **Append**: keep words and add emojis right after

**Tips**
- Use the *Replacement rate* to control how many words get emojis.
- Try examples below!
        """
    )

    with gr.Row():
        with gr.Column():
            inp = gr.Textbox(label="Your sentence", placeholder="Type here…", lines=4)
            mode = gr.Radio(choices=["Replace", "Append"], value="Replace", label="Mode")
            rate = gr.Slider(0.0, 1.0, value=1.0, step=0.05, label="Replacement rate")
            keep_case = gr.Checkbox(value=True, label="Preserve original capitalization")
            dedupe = gr.Checkbox(value=True, label="Avoid duplicate emojis when appending")
            btn = gr.Button("Translate ✨", variant="primary")
        with gr.Column():
            out = gr.Textbox(label="Emoji-fied text", lines=4)
            gr.Examples(EXAMPLES, inputs=inp, label="Examples")

    btn.click(translate_text, inputs=[inp, mode, rate, keep_case, dedupe], outputs=out)

if __name__ == "__main__":
    demo.launch()