app.py

import spacy
import requests
import wikipedia
import gradio as gr

# 1) Load spaCy small English model (make sure to add en_core_web_sm in requirements.txt)
nlp = spacy.load("en_core_web_sm")

# 2) Helper: Overpass query for POIs
def fetch_osm(lat, lon, osm_filter, limit=5):
    overpass = """
    [out:json][timeout:25];
    (
      node{filt}(around:1000,{lat},{lon});
      way{filt}(around:1000,{lat},{lon});
      rel{filt}(around:1000,{lat},{lon});
    );
    out center {lim};
    """.format(filt=osm_filter, lat=lat, lon=lon, lim=limit)
    r = requests.post("https://overpass-api.de/api/interpreter", data={"data": overpass})
    elems = r.json().get("elements", [])
    results = []
    for el in elems:
        name = el.get("tags", {}).get("name")
        if name:
            results.append({"name": name, **({"info": el["tags"].get("cuisine")} if "cuisine" in el["tags"] else {})})
    return results

# 3) Geocode via Nominatim
def geocode(place: str):
    r = requests.get(
        "https://nominatim.openstreetmap.org/search",
        params={"q": place, "format": "json", "limit": 1},
        headers={"User-Agent":"iVoiceContext/1.0"}
    )
    data = r.json()
    if not data: return None
    return float(data[0]["lat"]), float(data[0]["lon"])

# 4) Main context extractor
def get_context(text):
    doc = nlp(text)
    out = {}
    # gather unique entities of interest
    for ent in {e.text for e in doc.ents if e.label_ in ("GPE","LOC","PERSON","ORG")}:
        label = next(e.label_ for e in doc.ents if e.text == ent)
        if label in ("GPE","LOC"):
            geo = geocode(ent)
            if not geo:
                out[ent] = {"type":"location","error":"could not geocode"}
            else:
                lat, lon = geo
                out[ent] = {
                    "type": "location",
                    "restaurants": fetch_osm(lat, lon, '["amenity"="restaurant"]'),
                    "attractions": fetch_osm(lat, lon, '["tourism"="attraction"]'),
                }
        else:  # PERSON or ORG
            try:
                summ = wikipedia.summary(ent, sentences=2)
            except Exception:
                summ = "No summary available"
            out[ent] = {"type":"wiki","summary": summ}
    if not out:
        return {"error":"no named entities found"}
    return out

# 5) Gradio interface
iface = gr.Interface(
    fn=get_context,
    inputs=gr.Textbox(lines=3, placeholder="Enter or paste your translated text…"),
    outputs="json",
    title="iVoice Context-Aware API",
    description="Extracts people, places, orgs from text and returns nearby POIs or Wikipedia summaries."
)

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