Upload app.py

app.py

@@ -0,0 +1,139 @@
+import gradio as gr
+import pandas as pd
+import requests
+import json
+import time
+import base64
+from io import StringIO
+import getpass
+import os
+
+pd.set_option('display.max_columns', None)
+pd.set_option('display.max_rows', None)
+
+
+DATABRICKS_INSTANCE = os.getenv("DATABRICS_INSTANCE_")
+ACCESS_TOKEN = os.getenv("ACCESS_TOKEN_")
+JOB_ID = os.getenv("JOB_ID_")
+
+
+def get_run_status(run_id):
+    url = f"{DATABRICKS_INSTANCE}/api/2.1/jobs/runs/get?run_id={run_id}"
+    headers = {
+        "Authorization": f"Bearer {ACCESS_TOKEN}"
+    }
+    response = requests.get(url, headers=headers)
+    return response.json()
+
+def get_databricks_file(file_path):
+    # Construct the DBFS URL for the file
+    dbfs_url = f"{DATABRICKS_INSTANCE}/api/2.0/dbfs/read"
+    headers = {
+        "Authorization": f"Bearer {ACCESS_TOKEN}",
+        "Content-Type": "application/json"
+    }
+    data = {
+        "path": file_path
+    }
+    response = requests.get(dbfs_url, headers=headers, json=data)
+    #response.raise_for_status()  # Raise an exception for HTTP errors
+    content = response.json()
+    file_data = base64.b64decode(content['data'])
+    return file_data
+
+def get_month_label(i):
+    return f'M{i+1}'
+
+data_predictions_ = get_databricks_file("dbfs:/dbfs/FileStore/forecast_alleviatetax_predictions.csv")
+df = pd.read_csv(StringIO(data_predictions_.decode('utf-8')))
+df['revenue'] = (df['vintage_unique_cases'] / df['predicted_monthly_payment_rate']).round(2)
+df['Month'] = [get_month_label(i) for i in range(len(df))]
+    
+def update_table(start_date, end_date, window):
+    #############################################################################################
+    try:
+        DATABRICKS_INSTANCE = os.getenv("DATABRICS_INSTANCE_")
+        ACCESS_TOKEN = os.getenv("ACCESS_TOKEN_")
+        JOB_ID = os.getenv("JOB_ID_")
+        
+        # Parameters to pass to the job
+        params = {
+            "events_api_base_url": "https://beta.datascore.ai",
+            "month_train_starting": str(start_date),
+            "month_for_prediction": str(end_date),
+            "moving_average_period": str(window)
+        }
+        
+        url = f"{DATABRICKS_INSTANCE}/api/2.1/jobs/run-now"
+        headers = {
+            "Authorization": f"Bearer {ACCESS_TOKEN}",
+            "Content-Type": "application/json"
+        }
+        data = {
+            "job_id": JOB_ID,
+            "notebook_params": params  # Pass parameters here
+        }
+        response = requests.post(url, headers=headers, data=json.dumps(data))
+        #return response.json()
+        # Run the Job
+        #run_info = post_run_job(JOB_ID, params)
+        run_info = response.json()
+        run_id = run_info['run_id']
+        #print(f"Run ID: {run_id}")
+        
+        while True:
+            status = get_run_status(run_id)
+            state = status['state']['life_cycle_state']
+            if state in ['TERMINATED', 'SKIPPED', 'FAILED', 'ERROR']:
+                print(f"Job finished with state: {state}")
+                break
+            elif state in ['PROGRES']:
+                print(f"Job progress: {state}")
+            time.sleep(30)
+        
+        print(f"Run ID: {run_id}")
+        data_payments = get_databricks_file("dbfs:/dbfs/FileStore/forecast_alleviatetax_payments.csv")
+        data_metrics = get_databricks_file("dbfs:/dbfs/FileStore/forecast_alleviatetax_metrics.csv")
+        data_predictions = get_databricks_file("dbfs:/dbfs/FileStore/forecast_alleviatetax_predictions.csv")
+        df_payments = pd.read_csv(StringIO(data_payments.decode('utf-8')))
+        new_columns = ['vintage', 'vintage_unique_cases'] + [f'M{col}' for col in df_payments.columns[2:]] 
+        df_metrics = pd.read_csv(StringIO(data_metrics.decode('utf-8')))
+        df_predictions = pd.read_csv(StringIO(data_predictions.decode('utf-8')))
+        df_predictions['revenue'] = (df_predictions['vintage_unique_cases'] / df_predictions['predicted_monthly_payment_rate']).round(2)
+        df_predictions['Month'] = [get_month_label(i) for i in range(len(df_predictions))]
+    except Exception as e:
+        return {error: str(e)}
+    #############################################################################################
+    global df  # Use global variable
+    #df = pd.read_csv('ok_csv.csv')
+    #df = df_payments.copy()
+    df = df_predictions.copy()
+    return df
+
+
+def save_csv(file_name):
+    global df  # Use global variable
+    if file_name:
+        df.to_csv(file_name, index=False)
+        return f"File '{file_name}' saved!"
+    else:
+        return "Please set file name!"
+
+
+with gr.Blocks() as demo:
+    with gr.Row():
+        start_input = gr.Textbox(label="Start Date:", placeholder="Enter date (e.g., 2023-09-01)", max_lines=1)
+        prediction_input = gr.Text(label="Month Prediction:", placeholder="Enter date (e.g., 2024-10-31)", max_lines=1)
+        range_input = gr.Slider(1, 100, 6, label="Window / Moving Average Period")
+    
+    table_ = gr.DataFrame(value=df, label="Predictions (consult 'revenue' column):")
+    btn_actualizar = gr.Button("Run Forecast")
+
+    with gr.Row():
+        file_name_input = gr.Textbox(label="CSV Name", placeholder="Predictions.csv")
+        btn_guardar = gr.Button("Save CSV")
+    
+    btn_actualizar.click(fn=update_table, inputs=[start_input, prediction_input, range_input], outputs=table_)
+    btn_guardar.click(fn=save_csv, inputs=file_name_input, outputs=gr.Textbox(label="File Status:"))
+
+demo.launch()