import gradio as gr import pandas as pd from realtabformer import REaLTabFormer from scipy.io import arff import os rtf_model = REaLTabFormer( model_type="tabular", epochs=25, # Default is 200 gradient_accumulation_steps=4) def generate_data(file, num_samples): if '.arff' in file.name: data = arff.loadarff(open(file.name,'rt')) df = pd.DataFrame(data[0]) elif '.csv' in file.name: df = pd.read_csv(file.name) rtf_model.fit(df, num_bootstrap=100) # Default is 500 # Generate synthetic data samples = rtf_model.sample(n_samples=num_samples) return samples def generate_relational_data(parent_file, child_file, join_on): parent_df = pd.read_csv(parent_file.name) child_df = pd.read_csv(child_file.name) #Make sure join_on column exists in both assert ((join_on in parent_df.columns) and (join_on in child_df.columns)) rtf_model.fit(parent_df.drop(join_on, axis=1), num_bootstrap=100) pdir = Path("rtf_parent/") rtf_model.save(pdir) # # Get the most recently saved parent model, # # or a specify some other saved model. # parent_model_path = pdir / "idXXX" parent_model_path = sorted([ p for p in pdir.glob("id*") if p.is_dir()], key=os.path.getmtime)[-1] child_model = REaLTabFormer( model_type="relational", parent_realtabformer_path=parent_model_path, epochs = 25, output_max_length=None, train_size=0.8) child_model.fit( df=child_df, in_df=parent_df, join_on=join_on, num_bootstrap=100) # Generate parent samples. parent_samples = rtf_model.sample(5) # Create the unique ids based on the index. parent_samples.index.name = join_on parent_samples = parent_samples.reset_index() # Generate the relational observations. child_samples = child_model.sample( input_unique_ids=parent_samples[join_on], input_df=parent_samples.drop(join_on, axis=1), gen_batch=5) return parent_samples, child_samples, gr.update(visible = True) css = """ .gradio-container { font-family: 'IBM Plex Sans', sans-serif; } .gr-button { color: white; border-color: black; background: black; } input[type='range'] { accent-color: black; } .dark input[type='range'] { accent-color: #dfdfdf; } .container { max-width: 430px; margin: auto; padding-top: 1.5rem; } #gallery { min-height: 22rem; margin-bottom: 15px; margin-left: auto; margin-right: auto; border-bottom-right-radius: .5rem !important; border-bottom-left-radius: .5rem !important; } #gallery>div>.h-full { min-height: 20rem; } .details:hover { text-decoration: underline; } .gr-button { white-space: nowrap; } .gr-button:focus { border-color: rgb(147 197 253 / var(--tw-border-opacity)); outline: none; box-shadow: var(--tw-ring-offset-shadow), var(--tw-ring-shadow), var(--tw-shadow, 0 0 #0000); --tw-border-opacity: 1; --tw-ring-offset-shadow: var(--tw-ring-inset) 0 0 0 var(--tw-ring-offset-width) var(--tw-ring-offset-color); --tw-ring-shadow: var(--tw-ring-inset) 0 0 0 calc(3px var(--tw-ring-offset-width)) var(--tw-ring-color); --tw-ring-color: rgb(191 219 254 / var(--tw-ring-opacity)); --tw-ring-opacity: .5; } #advanced-btn { font-size: .7rem !important; line-height: 19px; margin-top: 12px; margin-bottom: 12px; padding: 2px 8px; border-radius: 14px !important; } #advanced-options { display: none; margin-bottom: 20px; } .footer { margin-bottom: 45px; margin-top: 35px; text-align: center; border-bottom: 1px solid #e5e5e5; } .footer>p { font-size: .8rem; display: inline-block; padding: 0 10px; transform: translateY(10px); background: white; } .dark .footer { border-color: #303030; } .dark .footer>p { background: #0b0f19; } """ with gr.Blocks(css = css) as demo: gr.Markdown(""" ## REaLTabFormer: Generating Realistic Relational and Tabular Data using Transformers """) gr.HTML('''

This is an unofficial demo for REaLTabFormer, an approach that can be used to generate synthetic data from single tabular data using GPT. The demo is based on the Github implementation provided by the authors.

''') with gr.Column(): #gr.Markdown(""" ### Record audio """) # with gr.Tab("Record Audio"): # audio_input_r = gr.Audio(label = 'Record Audio Input',source="microphone",type="filepath") # transcribe_audio_r = gr.Button('Transcribe') with gr.Tab("Upload Data as File: Tabular Data"): data_input_u = gr.File(label = 'Upload Data File (Currently supports CSV and ARFF)', file_types=[".csv", ".arff"]) num_samples = gr.Slider(label="Number of Samples", minimum=5, maximum=100, value=5, step=10) generate_data_btn = gr.Button('Generate Synthetic Data') with gr.Tab("Upload Data as File: Relational Data"): data_input_parent = gr.File(label = 'Upload Data File for Parent Dataset', file_types=[ ".csv"]) data_input_child = gr.File(label = 'Upload Data File for Child Dataset', file_types=[ ".csv"]) join_on = gr.Textbox(label = 'Column name to join on') generate_data_btn_relational = gr.Button('Generate Synthetic Data') with gr.Row(): #data_sample = gr.Dataframe(label = "Original Data") data_output = gr.Dataframe(label = "Synthetic Data") with gr.Row(visible = False) as child_sample: data_output_child = gr.Dataframe(label = "Synthetic Data for Child Dataset") generate_data_btn.click(generate_data, inputs = [data_input_u,num_samples], outputs = [data_output]) generate_data_btn_relational.click(generate_relational_data, inputs = [data_input_parent,data_input_child,join_on], outputs = [data_output, data_output_child]) examples = gr.Examples(examples=[['diabetes.arff',5], ["titanic.csv", 15]],inputs = [data_input_u,num_samples], outputs = [data_output], cache_examples = True, fn = generate_data) demo.launch()