Update app.py

app.py

@@ -1,77 +1,85 @@
-import spacy
+# app.py
+
 import requests
 import wikipedia
 import gradio as gr
+from transformers import pipeline
+
+# 1) Load the BERT NER pipeline
+ner = pipeline(
+    "ner",
+    model="dslim/bert-base-NER-uncased",
+    grouped_entities=True
+)
 
-# 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) Geocode via Nominatim
+def geocode(place: str):
+    resp = requests.get(
+        "https://nominatim.openstreetmap.org/search",
+        params={"q": place, "format": "json", "limit": 1},
+        headers={"User-Agent": "iVoiceContext/1.0"}
+    )
+    data = resp.json()
+    if not data:
+        return None
+    return float(data[0]["lat"]), float(data[0]["lon"])
 
-# 2) Helper: Overpass query for POIs
+# 3) Fetch POIs via Overpass
 def fetch_osm(lat, lon, osm_filter, limit=5):
-    overpass = """
+    query = f"""
     [out:json][timeout:25];
     (
-      node{filt}(around:1000,{lat},{lon});
-      way{filt}(around:1000,{lat},{lon});
-      rel{filt}(around:1000,{lat},{lon});
+      node{osm_filter}(around:1000,{lat},{lon});
+      way{osm_filter}(around:1000,{lat},{lon});
+      rel{osm_filter}(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})
+    out center {limit};
+    """
+    r = requests.post("https://overpass-api.de/api/interpreter", data={"data": query})
     elems = r.json().get("elements", [])
-    results = []
-    for el in elems:
-        name = el.get("tags", {}).get("name")
+    items = []
+    for e in elems:
+        name = e.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"])
+            items.append({"name": name})
+    return items
 
-# 4) Main context extractor
-def get_context(text):
-    doc = nlp(text)
+# 4) Main function
+def get_context(text: str):
+    results = ner(text)
+    # de-duplicate entities by text
+    ents = {ent["word"]: ent["entity_group"] for ent in results}
     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)
+    for word, label in ents.items():
+        if label == "LOC":
+            geo = geocode(word)
             if not geo:
-                out[ent] = {"type":"location","error":"could not geocode"}
+                out[word] = {"type":"location", "error":"could not geocode"}
             else:
                 lat, lon = geo
-                out[ent] = {
+                out[word] = {
                     "type": "location",
                     "restaurants": fetch_osm(lat, lon, '["amenity"="restaurant"]'),
                     "attractions": fetch_osm(lat, lon, '["tourism"="attraction"]'),
                 }
-        else:  # PERSON or ORG
+        else:
+            # for PERSON, ORG, MISC, etc → Wikipedia
             try:
-                summ = wikipedia.summary(ent, sentences=2)
+                summary = wikipedia.summary(word, sentences=2)
             except Exception:
-                summ = "No summary available"
-            out[ent] = {"type":"wiki","summary": summ}
+                summary = "No summary available."
+            out[word] = {"type": "wiki", "summary": summary}
     if not out:
-        return {"error":"no named entities found"}
+        return {"error": "No entities found"}
     return out
 
-# 5) Gradio interface
+# 5) Gradio UI
 iface = gr.Interface(
     fn=get_context,
-    inputs=gr.Textbox(lines=3, placeholder="Enter or paste your translated text…"),
+    inputs=gr.Textbox(lines=3, placeholder="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."
+    title="iVoice Context-Aware",
+    description="BERT NER → geocode LOC → Overpass POIs → Wikipedia for others"
 )
 
 if __name__ == "__main__":