Update app.py

app.py

@@ -1,5 +1,5 @@
 import streamlit as st
-from transformers import pipeline, TFAutoModelForSequenceClassification, AutoTokenizer
+from transformers import pipeline, TFAutoModelForSequenceClassification, AutoTokenizer, TFAutoModelForTokenClassification
 from datasets import load_dataset
 import pandas as pd
 
@@ -12,7 +12,7 @@ def load_data():
     return dataset
 
 dataset = load_data()
-st.title("Türkçe Sentiment Analizi Uygulaması")
+st.title("Türkçe Sentiment Analizi ve Entity Tanıma Uygulaması")
 
 # Örnek veri gösterme
 st.subheader("Örnek Veri")
@@ -21,7 +21,7 @@ st.write(sample_df.head())
 
 # Model seçim kısmı
 model_list = ['WhiteAngelss/entity-word-sentiment-analysis']
-st.sidebar.header("Sentiment Analizi Modeli Seçimi")
+st.sidebar.header("Model Seçimi")
 model_checkpoint = st.sidebar.radio("", model_list)
 
 st.sidebar.write("Model detayları için: 'https://huggingface.co/WhiteAngelss/entity-word-sentiment-analysis'")
@@ -43,22 +43,44 @@ elif input_method == "Yeni Metin Yaz veya Yapıştır":
 # Model ve tokenizer'ı yükleme
 @st.cache_resource
 def set_model(model_checkpoint):
-    model = TFAutoModelForSequenceClassification.from_pretrained(model_checkpoint, from_tf=True)  # TensorFlow modelini yükleyin
+    sentiment_model = TFAutoModelForSequenceClassification.from_pretrained(model_checkpoint, from_tf=True)
     tokenizer = AutoTokenizer.from_pretrained(model_checkpoint)
-    return pipeline('sentiment-analysis', model=model, tokenizer=tokenizer)
+    
+    # Named Entity Recognition (NER) için model
+    ner_model = TFAutoModelForTokenClassification.from_pretrained('dbmdz/bert-large-cased-finetuned-conll03-english', from_tf=True)
+    ner_tokenizer = AutoTokenizer.from_pretrained('dbmdz/bert-large-cased-finetuned-conll03-english')
+    
+    return {
+        'sentiment_pipeline': pipeline('sentiment-analysis', model=sentiment_model, tokenizer=tokenizer),
+        'ner_pipeline': pipeline('ner', model=ner_model, tokenizer=ner_tokenizer)
+    }
 
 # Analiz butonu
 run_button = st.button("Analiz Et")
 
 if run_button and input_text:
-    sentiment_pipeline = set_model(model_checkpoint)
-    output = sentiment_pipeline(input_text)
+    pipelines = set_model(model_checkpoint)
     
-    st.subheader("Sentiment Analizi Sonuçları")
-    df = pd.DataFrame(output)
-    st.dataframe(df)
-
-    # Sonuçları kullanıcı dostu bir formatta gösterme
-    sentiment = output[0]['label']
-    score = output[0]['score']
-    st.write(f"Sentiment: {sentiment} (Skor: {score:.2f})")
+    # Named Entity Recognition (NER) yapma
+    ner_pipeline = pipelines['ner_pipeline']
+    ner_output = ner_pipeline(input_text)
+    
+    # Sentiment analizi yapma
+    sentiment_pipeline = pipelines['sentiment_pipeline']
+    
+    # Entity bazında sonuçları toplama
+    results = []
+    for entity in ner_output:
+        entity_text = entity['word']
+        sentiment_output = sentiment_pipeline(entity_text)
+        
+        results.append({
+            'Entity': entity_text,
+            'Entity Sınıfı': entity['entity'],
+            'Sentiment': sentiment_output[0]['label'],
+            'Skor': sentiment_output[0]['score']
+        })
+    
+    st.subheader("Entity ve Sentiment Analizi Sonuçları")
+    results_df = pd.DataFrame(results)
+    st.write(results_df)