Update app.py

app.py

@@ -1,6 +1,5 @@
 import gradio as gr
 import requests
-from bs4 import BeautifulSoup
 
 from transformers import (
     AutoTokenizer,
@@ -34,53 +33,65 @@ viru_model = AutoModelForSeq2SeqLM.from_pretrained(
 )
 
 # =========================
-# SYNONYM FUNCTIONS
+# WIKTIONARY LOOKUP
 # =========================
 
-def get_ru_synonyms(word):
+def get_russian_info(word):
 
     try:
 
-        url = f"https://sinonim.org/s/{word}"
+        url = (
+            "https://ru.wiktionary.org/w/api.php"
+        )
 
-        headers = {
-            "User-Agent": "Mozilla/5.0"
+        params = {
+            "action": "query",
+            "prop": "extracts",
+            "exintro": True,
+            "explaintext": True,
+            "titles": word,
+            "format": "json"
         }
 
         response = requests.get(
             url,
-            headers=headers,
+            params=params,
             timeout=10
         )
 
-        soup = BeautifulSoup(
-            response.text,
-            "html.parser"
-        )
+        data = response.json()
 
-        syns = []
+        pages = data["query"]["pages"]
 
-        for a in soup.select("a"):
+        page = next(iter(pages.values()))
 
-            text = a.text.strip()
+        extract = page.get(
+            "extract",
+            "Không tìm thấy định nghĩa"
+        )
 
-            if (
-                text
-                and text.lower() != word.lower()
-                and len(text) > 1
-            ):
-                syns.append(text)
+        lines = extract.split("\n")
 
-        syns = list(dict.fromkeys(syns))
+        definition = (
+            lines[0]
+            if len(lines) > 0
+            else "Không tìm thấy"
+        )
 
-        if len(syns) == 0:
-            return "Không tìm thấy"
+        example = (
+            lines[1]
+            if len(lines) > 1
+            else "Không có ví dụ"
+        )
 
-        return ", ".join(syns[:10])
+        return definition, example
 
-    except Exception:
-        return "Lỗi tìm synonym"
+    except:
 
+        return (
+            "Không tìm thấy định nghĩa",
+            "Không có ví dụ"
+        )
 
 # =========================
 # TRANSLATION FUNCTIONS
@@ -100,10 +111,15 @@ def ru_to_vi(word):
         skip_special_tokens=True
     )
 
-    syns = get_ru_synonyms(word)
-
-    return translated, syns
+    definition, example = (
+        get_russian_info(word)
+    )
 
+    return (
+        translated,
+        definition,
+        example
+    )
 
 def vi_to_ru(word):
 
@@ -121,7 +137,6 @@ def vi_to_ru(word):
 
     return translated
 
-
 # =========================
 # UI
 # =========================
@@ -146,8 +161,12 @@ with gr.Blocks() as demo:
             label="🇻🇳 Nghĩa tiếng Việt"
         )
 
-        ru_syn_output = gr.Textbox(
-            label="📚 Từ đồng nghĩa"
+        definition_output = gr.Textbox(
+            label="📖 Định nghĩa"
+        )
+
+        example_output = gr.Textbox(
+            label="💬 Ví dụ"
         )
 
         ru_btn = gr.Button("Tra cứu")
@@ -157,7 +176,8 @@ with gr.Blocks() as demo:
             inputs=ru_input,
             outputs=[
                 vi_output,
-                ru_syn_output
+                definition_output,
+                example_output
             ]
         )