Upload New File

fb6672b3 · Albar · 004e91de · fb6672b3
Commit fb6672b3 authored Dec 14, 2023 by Albar
Hide whitespace changes
Inline Side-by-side

Showing with 132 additions and 0 deletions

Generating Synthetic Data.py Generating Synthetic Data.py +132 -0

No files found.
--- a/Generating Synthetic Data.py
+++ b/Generating Synthetic Data.py
+#!/usr/bin/env python
+# coding: utf-8
+# In[ ]:
+# Installing required packages
+get_ipython().system('pip install openai')
+get_ipython().system('pip install nltk')
+get_ipython().system('pip install pandas')
+# In[1]:
+# Importing libraries
+import os
+import openai
+import io
+import pandas as pd
+# Download NLTK resources
+import nltk
+from nltk.sentiment import SentimentIntensityAnalyzer
+nltk.download('vader_lexicon')
+# Using Azure OpenAI, replace the keys below
+openai.api_type = "azure"
+openai.api_base = "INSERT URI HERE"
+openai.api_version = "INSERT API VERSION HERE"
+openai.api_key = "INSERT KEY HERE"
+# In[3]:
+# Example with a small dataset
+data = {
+ 'Full Name': ['Alice Johnson', 'Bob Smith', 'Charlie Brown', 'David Miller', 'Eva White'],
+ 'Address': ['123 Main St', '456 Oak Ave', '789 Pine Ln', '101 Elm Blvd', '202 Cedar Dr'],
+ 'Age': [25, 30, 22, 35, 28],
+ 'reviews': [
+ "The food at this restaurant was absolutely amazing! Each dish was a delightful experience, and the service was top-notch.",
+ "Unfortunately, my experience at this restaurant was less than satisfactory. The service was slow, and the food arrived cold. I wouldn't recommend it.",
+ "I enjoyed the atmosphere of the restaurant, but the portions were disappointingly small. I left feeling hungry despite the tasty dishes.",
+ "The service was excellent, and the dishes were delicious. I will definitely come back to this restaurant in the future.",
+ "My visit to this restaurant was terrible. The staff was rude, and the food had no flavor. I won't be returning."
+]
+}
+original_data = pd.DataFrame(data)
+# In[4]:
+original_data
+# In[5]:
+prompt = "You are an AI Assistant that generates 5 synthetic data points upon user request. Only return the markdown table. Do not return additional text. Create new reviews"
+query = f"Generate 5 new synthetic data with similar distribution, pattern and sentiment to the following data. Write new reviews:\n\n{original_data.to_markdown(index=False)}"
+# Make a request to OpenAI GPT-3.5. 
+response = openai.ChatCompletion.create(
+ engine="INSERT MODEL NAME HERE", 
+ messages=[
+ {"role": "system", "content": prompt},
+ {"role": "user", "content": query}
+ ],
+ temperature=0.7,
+ stop=None
+)
+# In[6]:
+# Function to get sentiment score
+def get_sentiment_score(text):
+ sid = SentimentIntensityAnalyzer()
+ sentiment_scores = sid.polarity_scores(text)
+ return sentiment_scores['compound']
+# Add a new column 'sentiment_score' to the DataFrame
+original_data['sentiment_score'] = original_data['reviews'].apply(get_sentiment_score)
+# Display the DataFrame
+original_data
+# In[7]:
+generated_text=response['choices'][0]['message']['content']
+# In[8]:
+generated_text
+# In[9]:
+cleaned_string = '\n'.join(['|' + '|'.join([cell.strip() for cell in row.split('|')[1:-1]]) + '|' for row in generated_text.strip().split('\n')])
+# Use pandas to read the formatted string into a DataFrame
+df = pd.read_csv(io.StringIO(cleaned_string), sep="|", index_col=False)
+# In[10]:
+df = df.dropna(axis=1, how='all')
+df=df = df.drop(0, axis=0)
+df['Age'] = pd.to_numeric(df['Age'], errors='coerce')
+# In[11]:
+df
+# In[14]:
+df['sentiment_score'] = df['reviews'].apply(get_sentiment_score)
+df