src/streamlit_app.py

import tempfile
import urllib.parse
import requests
import streamlit as st
from gtts import gTTS
from translatepy import Translator

# -------------------- CONFIG --------------------
REQUEST_TIMEOUT = 20
HEADERS = {
    "User-Agent": "WikiHealthNavigator/1.0 (educational-project)"
}
translator = Translator()

# -------------------- FETCH SUMMARY (SAFE + FALLBACK) --------------------
@st.cache_data(ttl=3600)
def fetch_summary(query, lang):
    try:
        title = urllib.parse.quote(query.strip())
        url = f"https://{lang}.wikipedia.org/api/rest_v1/page/summary/{title}"
        r = requests.get(url, headers=HEADERS, timeout=REQUEST_TIMEOUT)

        # Handle missing pages (404 etc.)
        if r.status_code != 200:
            return None, None

        data = r.json()
        return (
            data.get("extract"),
            data.get("content_urls", {}).get("desktop", {}).get("page")
        )
    except Exception:
        return None, None

# -------------------- FETCH IMAGES (WIKIMEDIA COMMONS – SAFE) --------------------
@st.cache_data(ttl=3600)
def fetch_images(query, max_images=4):
    api = "https://commons.wikimedia.org/w/api.php"
    params = {
        "action": "query",
        "format": "json",
        "generator": "search",
        "gsrsearch": query,
        "gsrlimit": max_images,
        "gsrnamespace": 6,  # File namespace
        "prop": "imageinfo",
        "iiprop": "url"
    }

    try:
        r = requests.get(api, params=params, headers=HEADERS, timeout=REQUEST_TIMEOUT)

        if r.status_code != 200:
            return []

        if "application/json" not in r.headers.get("Content-Type", ""):
            return []

        data = r.json()
        pages = data.get("query", {}).get("pages", {})
        images = []

        for page in pages.values():
            if "imageinfo" in page:
                images.append(page["imageinfo"][0]["url"])

        return images
    except Exception:
        return []

# -------------------- NARRATION --------------------
def narrate(text, lang):
    try:
        tts = gTTS(text=text, lang=lang)
        with tempfile.NamedTemporaryFile(delete=False, suffix=".mp3") as f:
            tts.save(f.name)
            st.audio(f.name)
    except Exception as e:
        st.error(f"Audio error: {e}")

# -------------------- SESSION STATE INIT --------------------
for key in ["summary", "images", "url", "lang"]:
    if key not in st.session_state:
        st.session_state[key] = None

# -------------------- UI --------------------
st.set_page_config(page_title="WikiHealth Navigator", layout="centered")
st.title("🩺 WikiHealth Navigator")
st.markdown(
    "Health topic search using **Wikipedia** text and "
    "**** images"
)

languages = {
    "English": "en",
    "हिन्दी (Hindi)": "hi",
    "తెలుగు (Telugu)": "te",
    "தமிழ் (Tamil)": "ta"
}

lang_name = st.selectbox("Select Language", list(languages.keys()))
lang_code = languages[lang_name]

query = st.text_input("Enter health topic")

# -------------------- SEARCH --------------------
if st.button("🔍 Search") and query:
    # 1️⃣ Try selected language
    summary, url = fetch_summary(query, lang_code)

    # 2️⃣ Fallback to English if page not found
    if not summary:
        summary, url = fetch_summary(query, "en")

        # 3️⃣ Translate English → selected language
        if summary and lang_code != "en":
            summary = translator.translate(summary, lang_code).result

    st.session_state.summary = summary
    st.session_state.images = fetch_images(query)
    st.session_state.url = url
    st.session_state.lang = lang_code

# -------------------- DISPLAY --------------------
if st.session_state.summary:
    st.subheader("📝 Summary")
    st.write(st.session_state.summary)

    st.subheader("🖼️ Related Images")
    if st.session_state.images:
        cols = st.columns(len(st.session_state.images))
        for col, img in zip(cols, st.session_state.images):
            col.image(img, use_container_width=True)
    else:
        st.info("Images temporarily unavailable. Please try again.")

    if st.button("🔊 Narrate Summary"):
        narrate(st.session_state.summary, st.session_state.lang)

    if st.session_state.url:
        st.markdown(f"[Read full article on Wikipedia]({st.session_state.url})")