app5_selectbox/evaluation_analysis.py

@@ -1,18 +1,19 @@
-import gspread
+# import gspread
 import pandas as pd
-from oauth2client.service_account import ServiceAccountCredentials
+# from oauth2client.service_account import ServiceAccountCredentials
 from transformers import AutoTokenizer, AutoModelForSequenceClassification
 import torch
 import re
 import streamlit as st
 import hydralit_components as hc
 from matplotlib import pyplot as plt
-import numpy as np
+# import numpy as np
 from wordcloud import WordCloud
 import plotly.graph_objs as go
-import plotly.express as px
-import plotly.figure_factory as ff
+# import plotly.express as px
+# import plotly.figure_factory as ff
 # from PIL import ImageFont
+
 # from app5_selectbox.langchain_llama_gpu import llm_chain
 from app5_selectbox.g4f_prompt import g4f_prompt
 # from app5_selectbox.llama2_prompt import llama_prompt
@@ -61,55 +62,55 @@ def clean_text(text_list):
     return cleaned_samples
 
 
-# local model
-def classify_sentiments(text_samples, tokenizer, model):
-    instructor_comments = []
-    predicted_sentiments = []
-    predicted_sentiments_scores = []
+# # local model
+# def classify_sentiments(text_samples, tokenizer, model):
+#     instructor_comments = []
+#     predicted_sentiments = []
+#     predicted_sentiments_scores = []
     
-    # Iterate through the text samples and classify the sentiment
-    for idx, text_sample in enumerate(text_samples):
-        # Tokenize the text sample
-        inputs = tokenizer(text_sample, return_tensors="pt")
+#     # Iterate through the text samples and classify the sentiment
+#     for idx, text_sample in enumerate(text_samples):
+#         # Tokenize the text sample
+#         inputs = tokenizer(text_sample, return_tensors="pt")
 
-        # Perform sentiment classification
-        outputs = model(**inputs)
+#         # Perform sentiment classification
+#         outputs = model(**inputs)
 
-        # Get the predicted sentiment (positive/negative)
-        predicted_class = torch.argmax(outputs.logits, dim=1).item()
+#         # Get the predicted sentiment (positive/negative)
+#         predicted_class = torch.argmax(outputs.logits, dim=1).item()
 
-        # Get the probabilities for each class
-        probabilities = torch.softmax(outputs.logits, dim=1).tolist()[0]
+#         # Get the probabilities for each class
+#         probabilities = torch.softmax(outputs.logits, dim=1).tolist()[0]
 
-        # Store results
-        instructor_comments.append(text_sample)
-        predicted_sentiments.append("positive" if predicted_class == 1 else "negative")
-        predicted_sentiments_scores.append({"positive": probabilities[1]*100, "negative": probabilities[0]*100})
-    return instructor_comments, predicted_sentiments, predicted_sentiments_scores
+#         # Store results
+#         instructor_comments.append(text_sample)
+#         predicted_sentiments.append("positive" if predicted_class == 1 else "negative")
+#         predicted_sentiments_scores.append({"positive": probabilities[1]*100, "negative": probabilities[0]*100})
+#     return instructor_comments, predicted_sentiments, predicted_sentiments_scores
 
 
-# # inference
-# def classify_sentiments(text_samples, model):
-#     instructor_comments = []
-#     predicted_sentiments = []
-#     predicted_sentiments_scores = []
+# inference
+def classify_sentiments(text_samples, model):
+    instructor_comments = []
+    predicted_sentiments = []
+    predicted_sentiments_scores = []
 
-#     # text = ["i love this", "nice one!", "happy!"]
-#     selected_model = model
-#     results = [analyze_sintement(t, selected_model) for t in text_samples]
+    # text = ["i love this", "nice one!", "happy!"]
+    selected_model = model
+    results = [analyze_sintement(t, selected_model) for t in text_samples]
     
     
     
-#     for idx, result in enumerate(results):
-#         # st.text(result[0])
-#         # predicted_class, probabilities = analyze_sintement(text_sample, model)
-#         # Store results
-#         instructor_comments.append(text_samples[idx])
-#         predicted_sentiments.append("positive" if result[0] == "LABEL_1" else "negative")
-#         predicted_sentiments_scores.append({"positive": result[1]*100, "negative": 100-(result[1]*100)})
+    for idx, result in enumerate(results):
+        # st.text(result[0])
+        # predicted_class, probabilities = analyze_sintement(text_sample, model)
+        # Store results
+        instructor_comments.append(text_samples[idx])
+        predicted_sentiments.append("positive" if result[0] == "LABEL_1" else "negative")
+        predicted_sentiments_scores.append({"positive": result[1]*100, "negative": 100-(result[1]*100)})
 
-#     # st.write(instructor_comments)
-#     return instructor_comments, predicted_sentiments, predicted_sentiments_scores
+    # st.write(instructor_comments)
+    return instructor_comments, predicted_sentiments, predicted_sentiments_scores