import gradio as gr import pandas as pd def nlp_pipeline(original_df): original_df['Sum'] = df['a'] + df['b'] return original_df def process_excel(file): try: # Ensure the file path is correct file_path = file.name if hasattr(file, 'name') else file # Read the Excel file df = pd.read_excel(file_path) # Perform any processing on the DataFrame here # Example: adding a new column with the sum of two other columns # df['Sum'] = df['Column1'] + df['Column2'] result_df = nlp_pipeline(original_df) return result_df # Return the first few rows as an example except Exception as e: return str(e) # Return the error message # Define the Gradio interface interface = gr.Interface( fn=process_excel, # The function to process the uploaded file inputs=gr.File(type="filepath", label="Upload Excel File"), # File upload input outputs="dataframe", # Display the output as a DataFrame title="Excel File Uploader", description="Upload an Excel file to see the first few rows." ) # Launch the interface if __name__ == "__main__": interface.launch() # #!/usr/bin/env python # # coding: utf-8 # import pandas as pd # import string # import nltk # import seaborn as sns # import matplotlib.pyplot as plt # from nltk.corpus import stopwords # from nltk.tokenize import word_tokenize # from nltk.sentiment import SentimentIntensityAnalyzer # from sklearn.feature_extraction.text import TfidfVectorizer # from sklearn.cluster import KMeans # from transformers import T5ForConditionalGeneration, T5Tokenizer # from datasets import Dataset # # Load the data # file_responses = pd.read_excel("#TaxDirection (Responses).xlsx") # # Process financial allocations # def process_allocations(df, col_name): # return pd.to_numeric(df[col_name], errors='coerce').fillna(0) # columns_to_process = [ # '''Your financial allocation for Problem 1: # Mention the percentage of your Tax Amount which you wish the Government would allocate through their annual budget, to implement a specific solution for your 1st problem.''', # '''Your financial allocation for Problem 2: # Mention the percentage of your Tax Amount which you wish the Government would allocate through their annual budget, to implement a solution specifically to your 2nd problem.''', # '''Your financial allocation for Problem 3: # Mention the percentage of your Tax Amount which you wish the Government would allocate through their annual budget, to implement a solution specifically to your 3rd problem.''' # ] # for col in columns_to_process: # file_responses[col] = process_allocations(file_responses, col) # file_responses['How much was your latest Tax payment (in U$D)?'] = pd.to_numeric( # file_responses['How much was your latest Tax payment (in U$D)?'], errors='coerce').fillna(0) # # Compute total allocation and financial weights # file_responses['Total Allocation'] = file_responses[columns_to_process].apply(lambda x: x.clip(lower=10)).sum(axis=1) # for i in range(1, 4): # file_responses[f'Financial Token Weight for Problem {i}'] = ( # file_responses['How much was your latest Tax payment (in U$D)?'] * # file_responses[columns_to_process[i - 1]] / # file_responses['Total Allocation'] # ) # # Create initial datasets # initial_datasets = [] # for i in range(1, 4): # initial_datasets.append( # file_responses[[f'''Describe Problem {i}: # Enter the context of the problem. # What are the difficulties you are facing personally or as a part of an organization? # You may briefly propose a solution idea as well.''', # f'''Problem {i}: Geographical Location : # Where is the location you are facing this problem? # You may mention the nearby geographical area of the proposed solution as: # City/Town, State/Province, Country.''', # f'Financial Token Weight for Problem {i}']] # ) # # Rename columns # for idx, df in enumerate(initial_datasets): # initial_datasets[idx] = df.rename(columns={ # df.columns[0]: 'Problem_Description', # df.columns[1]: 'Geographical_Location', # df.columns[2]: 'Financial_Weight' # }) # # Merge datasets # merged_dataset = pd.concat(initial_datasets, ignore_index=True) # # Preprocess text # nltk.download('stopwords') # nltk.download('punkt') # nltk.download('omw-1.4') # def preprocess_text(text): # translator = str.maketrans("", "", string.punctuation) # text = text.translate(translator) # tokens = word_tokenize(text) # stop_words = set(stopwords.words('english')) # tokens = [word for word in tokens if word.lower() not in stop_words] # return ' '.join(tokens) # merged_dataset['Problem_Description'] = merged_dataset['Problem_Description'].astype(str).apply(preprocess_text) # merged_dataset['Problem_Description'] = merged_dataset['Problem_Description'].str.replace(r'\d+', '', regex=True) # merged_dataset['Geographical_Location'] = merged_dataset['Geographical_Location'].str.replace(r'\d+', '', regex=True) # merged_dataset['Problem_Description'] = merged_dataset['Problem_Description'].replace(r'http\S+', '', regex=True).replace(r'www\S+', '', regex=True) # merged_dataset['Geographical_Location'] = merged_dataset['Geographical_Location'].replace(r'http\S+', '', regex=True).replace(r'www\S+', '', regex=True) # # Lemmatize text # lemmatizer = nltk.WordNetLemmatizer() # merged_dataset['Problem_Description'] = merged_dataset['Problem_Description'].apply(lambda x: ' '.join([lemmatizer.lemmatize(word) for word in x.split()])) # # Clustering # corpus = merged_dataset['Problem_Description'].tolist() # tfidf_vectorizer = TfidfVectorizer(max_features=77000) # tfidf_matrix = tfidf_vectorizer.fit_transform(corpus) # problem_cluster_count = 77 # kmeans = KMeans(n_clusters=problem_cluster_count) # kmeans.fit(tfidf_matrix) # terms = tfidf_vectorizer.get_feature_names_out() # ordered_centroids = kmeans.cluster_centers_.argsort()[:, ::-1] # cluster_representations = {} # for i in range(kmeans.n_clusters): # cluster_representations[i] = [terms[ind] for ind in ordered_centroids[i, :17]] # merged_dataset['Problem_Category_Numeric'] = kmeans.labels_ # merged_dataset['Problem_Category_Words'] = [cluster_representations[label] for label in kmeans.labels_] # # Clustering geographical locations # geographical_data = merged_dataset['Geographical_Location'].tolist() # tfidf_vectorizer_geography = TfidfVectorizer(max_features=3000) # tfidf_matrix_geography = tfidf_vectorizer_geography.fit_transform(geographical_data) # location_cluster_count = 33 # kmeans_locations = KMeans(n_clusters=location_cluster_count) # kmeans_locations.fit(tfidf_matrix_geography) # terms_geography = tfidf_vectorizer_geography.get_feature_names_out() # ordered_centroids_geography = kmeans_locations.cluster_centers_.argsort()[:, ::-1] # cluster_representations_geography = {} # for i in range(kmeans_locations.n_clusters): # cluster_representations_geography[i] = [terms_geography[ind] for ind in ordered_centroids_geography[i, :5]] # merged_dataset['Location_Category_Numeric'] = kmeans_locations.labels_ # merged_dataset['Location_Category_Words'] = [cluster_representations_geography[label] for label in kmeans_locations.labels_] # # Create 2D matrices for problem descriptions and financial weights # matrix2Dfinances = [[[] for _ in range(location_cluster_count)] for _ in range(problem_cluster_count)] # matrix2Dproblems = [[[] for _ in range(location_cluster_count)] for _ in range(problem_cluster_count)] # for index, row in merged_dataset.iterrows(): # location_index = row['Location_Category_Numeric'] # problem_index = row['Problem_Category_Numeric'] # problem_description = row['Problem_Description'] # financial_wt = row['Financial_Weight'] # matrix2Dproblems[problem_index][location_index].append(problem_description) # matrix2Dfinances[problem_index][location_index].append(financial_wt) # # Aggregating financial weights # aggregated_Financial_wts = {} # un_aggregated_Financial_wts = {} # for Financ_wt_index, Financ_wt_row in enumerate(matrix2Dfinances): # aggregated_Financial_wts[Financ_wt_index] = {} # un_aggregated_Financial_wts[Financ_wt_index] = {} # for location_index, cell_finances in enumerate(Financ_wt_row): # cell_sum = sum(cell_finances) # aggregated_Financial_wts[Financ_wt_index][location_index] = cell_sum # un_aggregated_Financial_wts[Financ_wt_index][location_index] = cell_finances # matrix2Dfinances_df = pd.DataFrame(aggregated_Financial_wts) # matrix2Dfinances_df.to_excel('matrix2Dfinances_HeatMap.xlsx', index=True) # unagregated_finances_df = pd.DataFrame(un_aggregated_Financial_wts) # unagregated_finances_df.to_excel('UNaggregated Financial Weights.xlsx', index=True) # # Create heatmaps # plt.figure(figsize=(15, 7)) # sns.heatmap(matrix2Dfinances_df, annot=False, cmap='RdYlGn') # plt.title('Project Financial Weights') # plt.ylabel('Location Clusters') # plt.xlabel('Problem Clusters') # plt.savefig('Project Financial Weights_HeatMap_GreenHigh.png') # plt.show() # plt.figure(figsize=(14, 6)) # sns.heatmap(matrix2Dfinances_df, annot=False, cmap='RdYlGn_r') # plt.title('Project Financial Weights') # plt.ylabel('Location Clusters') # plt.xlabel('Problem Clusters') # plt.savefig('Project Financial Weights_HeatMap_RedHigh.png') # plt.show() # # Summarizing problems using T5 # model = T5ForConditionalGeneration.from_pretrained('t5-small') # tokenizer = T5Tokenizer.from_pretrained('t5-small') # def t5_summarize(text): # input_text = "summarize: " + text # inputs = tokenizer.encode(input_text, return_tensors="pt", max_length=512, truncation=True) # summary_ids = model.generate(inputs, max_length=150, min_length=40, length_penalty=2.0, num_beams=4, early_stopping=True) # return tokenizer.decode(summary_ids[0], skip_special_tokens=True) # summarized_problems = [[t5_summarize(" ".join(cell)) for cell in row] for row in matrix2Dproblems] # # Save summarized problems # with open('summarized_problems.txt', 'w') as file: # for problem_row in summarized_problems: # file.write("\t".join(problem_row) + "\n")