Upload app.py

app.py

@@ -0,0 +1,89 @@
+import tensorflow as tf
+from huggingface_hub import hf_hub_download
+from datasets import load_dataset
+import numpy as np
+import gradio as gr
+from sklearn.preprocessing import MinMaxScaler
+import pandas as pd
+
+model_path = hf_hub_download(repo_id='neuronetties/bezrabotica', filename='BEZRAB.keras')
+model = tf.keras.models.load_model(model_path)
+
+dataset = load_dataset("Vilyam888/BEZRAB_DATA")
+data = pd.DataFrame(dataset['train'])
+
+columns_to_keep = [
+    'territory',
+    'num_economactivepopulation_all',
+    'employed_num_all',
+    'unemployed_num_all',
+    'eactivity_lvl',
+    'employment_lvl',
+    'unemployment_lvl',
+    'dis_unagegroup_30-39',
+    'dis_emagegroup_30-39',
+    'num_unagegroup_30-39',
+    'num_emagegroup_30-39',
+    'year'
+]
+data = data[columns_to_keep]
+data.fillna(data.mean(numeric_only=True), inplace=True)
+
+def clean_territory(value):
+    if isinstance(value, str) and len(value) > 100:
+        return value[:100]
+    return value
+
+data['territory'] = data['territory'].apply(clean_territory)
+data['territory'] = data['territory'].astype(str).str.strip()
+territory_mapping = {territory: idx for idx, territory in enumerate(data['territory'].unique())}
+data['territory'] = data['territory'].map(territory_mapping)
+
+scaler = MinMaxScaler()
+normalized_data = scaler.fit_transform(data.drop(['year'], axis=1))
+normalized_df = pd.DataFrame(normalized_data, columns=[col for col in data.columns if col != 'year'])
+normalized_df['year'] = data['year'].values
+
+def predict(territory, year, num_economactivepopulation_all, employed_num_all, unemployed_num_all,
+            eactivity_lvl, employment_lvl, dis_unagegroup_30_39, dis_emagegroup_30_39, 
+            num_unagegroup_30_39, num_emagegroup_30_39):
+    input_data = pd.DataFrame({
+        'territory': [territory_mapping[territory]],
+        'num_economactivepopulation_all': [num_economactivepopulation_all],
+        'employed_num_all': [employed_num_all],
+        'unemployed_num_all': [unemployed_num_all],
+        'eactivity_lvl': [eactivity_lvl],
+        'employment_lvl': [employment_lvl],
+        'dis_unagegroup_30-39': [dis_unagegroup_30_39],
+        'dis_emagegroup_30-39': [dis_emagegroup_30_39],
+        'num_unagegroup_30-39': [num_unagegroup_30_39],
+        'num_emagegroup_30-39': [num_emagegroup_30_39]
+    })
+
+    input_normalized = scaler.transform(input_data)
+    input_sequence = np.expand_dims(input_normalized, axis=0)
+
+    prediction = model.predict(input_sequence)
+    return round(prediction[0][0], 2)
+
+interface = gr.Interface(
+    fn=predict,
+    inputs=[
+        gr.Textbox(label="Territory", placeholder="Введите название территории"),
+        gr.Number(label="Year"),
+        gr.Number(label="Economic Active Population"),
+        gr.Number(label="Employed Population"),
+        gr.Number(label="Unemployed Population"),
+        gr.Number(label="Economic Activity Level"),
+        gr.Number(label="Employment Level"),
+        gr.Number(label="Disunemployed Age Group 30-39"),
+        gr.Number(label="Disemployed Age Group 30-39"),
+        gr.Number(label="Unemployed Age Group 30-39"),
+        gr.Number(label="Employed Age Group 30-39")
+    ],
+    outputs=gr.Textbox(label="Predicted Value"),
+    title="Unemployment Prediction Model",
+    description="Введите значения для прогноза уровня безработицы."
+)
+
+interface.launch()