Create app.py
Browse files
app.py
ADDED
@@ -0,0 +1,151 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
from __future__ import annotations
|
2 |
+
from typing import Iterable
|
3 |
+
import gradio as gr
|
4 |
+
from gradio.themes.base import Base
|
5 |
+
from gradio.themes.utils import colors, fonts, sizes
|
6 |
+
import time
|
7 |
+
|
8 |
+
import pandas as pd
|
9 |
+
import matplotlib.pyplot as plt
|
10 |
+
import numpy as np
|
11 |
+
|
12 |
+
import torch
|
13 |
+
from chronos import ChronosPipeline
|
14 |
+
import warnings
|
15 |
+
warnings.filterwarnings("ignore")
|
16 |
+
|
17 |
+
class Seafoam(Base):
|
18 |
+
def __init__(
|
19 |
+
self,
|
20 |
+
*,
|
21 |
+
primary_hue: colors.Color | str = colors.emerald,
|
22 |
+
secondary_hue: colors.Color | str = colors.blue,
|
23 |
+
neutral_hue: colors.Color | str = colors.blue,
|
24 |
+
spacing_size: sizes.Size | str = sizes.spacing_md,
|
25 |
+
radius_size: sizes.Size | str = sizes.radius_md,
|
26 |
+
text_size: sizes.Size | str = sizes.text_lg,
|
27 |
+
font: fonts.Font
|
28 |
+
| str
|
29 |
+
| Iterable[fonts.Font | str] = (
|
30 |
+
fonts.GoogleFont("Quicksand"),
|
31 |
+
"ui-sans-serif",
|
32 |
+
"sans-serif",
|
33 |
+
),
|
34 |
+
font_mono: fonts.Font
|
35 |
+
| str
|
36 |
+
| Iterable[fonts.Font | str] = (
|
37 |
+
fonts.GoogleFont("IBM Plex Mono"),
|
38 |
+
"ui-monospace",
|
39 |
+
"monospace",
|
40 |
+
),
|
41 |
+
):
|
42 |
+
super().__init__(
|
43 |
+
primary_hue=primary_hue,
|
44 |
+
secondary_hue=secondary_hue,
|
45 |
+
neutral_hue=neutral_hue,
|
46 |
+
spacing_size=spacing_size,
|
47 |
+
radius_size=radius_size,
|
48 |
+
text_size=text_size,
|
49 |
+
font=font,
|
50 |
+
font_mono=font_mono,
|
51 |
+
)
|
52 |
+
super().set(
|
53 |
+
body_background_fill="repeating-linear-gradient(45deg, *primary_200, *primary_200 10px, *primary_50 10px, *primary_50 20px)",
|
54 |
+
body_background_fill_dark="repeating-linear-gradient(45deg, *primary_800, *primary_800 10px, *primary_900 10px, *primary_900 20px)",
|
55 |
+
button_primary_background_fill="linear-gradient(90deg, *primary_300, *secondary_400)",
|
56 |
+
button_primary_background_fill_hover="linear-gradient(90deg, *primary_200, *secondary_300)",
|
57 |
+
button_primary_text_color="white",
|
58 |
+
button_primary_background_fill_dark="linear-gradient(90deg, *primary_600, *secondary_800)",
|
59 |
+
slider_color="*secondary_300",
|
60 |
+
slider_color_dark="*secondary_600",
|
61 |
+
block_title_text_weight="600",
|
62 |
+
block_border_width="3px",
|
63 |
+
block_shadow="*shadow_drop_lg",
|
64 |
+
button_primary_shadow="*shadow_drop_lg",
|
65 |
+
button_large_padding="32px",
|
66 |
+
)
|
67 |
+
|
68 |
+
seafoam = Seafoam()
|
69 |
+
|
70 |
+
|
71 |
+
import numpy as np
|
72 |
+
import matplotlib.ticker as ticker
|
73 |
+
|
74 |
+
def process_data(csv_file):
|
75 |
+
try:
|
76 |
+
# Read the CSV file
|
77 |
+
df = pd.read_csv(csv_file.name)
|
78 |
+
|
79 |
+
df['date'] = pd.to_datetime(df['date'])
|
80 |
+
df['month'] = df['date'].dt.month
|
81 |
+
df['year'] = df['date'].dt.year
|
82 |
+
|
83 |
+
monthly_sales = df.groupby(['year', 'month'])['sold_qty'].sum().reset_index()
|
84 |
+
monthly_sales = monthly_sales.rename(columns={'year': 'year', 'month': 'month', 'sold_qty': 'y'})
|
85 |
+
|
86 |
+
pipeline = ChronosPipeline.from_pretrained(
|
87 |
+
"amazon/chronos-t5-base",
|
88 |
+
device_map="cpu",
|
89 |
+
torch_dtype=torch.float32,
|
90 |
+
)
|
91 |
+
context = torch.tensor(monthly_sales["y"])
|
92 |
+
prediction_length = 12
|
93 |
+
forecast = pipeline.predict(context, prediction_length)
|
94 |
+
|
95 |
+
# Prepare forecast data
|
96 |
+
forecast_index = range(len(monthly_sales), len(monthly_sales) + prediction_length)
|
97 |
+
low, median, high = np.quantile(forecast[0].numpy(), [0.1, 0.5, 0.9], axis=0)
|
98 |
+
|
99 |
+
# Visualization
|
100 |
+
plt.figure(figsize=(30, 10))
|
101 |
+
plt.plot(monthly_sales["y"], color="royalblue", label="Historical Data", linewidth=2)
|
102 |
+
plt.plot(forecast_index, median, color="tomato", label="Median Forecast", linewidth=2)
|
103 |
+
plt.fill_between(forecast_index, low, high, color="tomato", alpha=0.3, label="80% Prediction Interval")
|
104 |
+
plt.title("Sales Forecasting Visualization", fontsize=16)
|
105 |
+
plt.xlabel("Months", fontsize=20)
|
106 |
+
plt.ylabel("Sold Qty", fontsize=20)
|
107 |
+
|
108 |
+
plt.xticks(fontsize=18)
|
109 |
+
plt.yticks(fontsize=18)
|
110 |
+
|
111 |
+
ax = plt.gca()
|
112 |
+
ax.xaxis.set_major_locator(ticker.MultipleLocator(3))
|
113 |
+
ax.yaxis.set_major_locator(ticker.MultipleLocator(5))
|
114 |
+
ax.grid(which='major', linestyle='--', linewidth=1.2, color='gray', alpha=0.7)
|
115 |
+
|
116 |
+
plt.legend(fontsize=18)
|
117 |
+
plt.grid(linestyle='--', linewidth=1.2, color='gray', alpha=0.7)
|
118 |
+
plt.tight_layout()
|
119 |
+
|
120 |
+
return plt.gcf()
|
121 |
+
|
122 |
+
except Exception as e:
|
123 |
+
print(f"Error: {str(e)}")
|
124 |
+
return None
|
125 |
+
|
126 |
+
# Create Gradio interface
|
127 |
+
with gr.Blocks(theme=seafoam) as demo:
|
128 |
+
gr.Markdown("# Chronos Forecasting - Tops Infosolution Pvt. Ltd")
|
129 |
+
gr.Markdown("Upload a CSV file and click 'Forecast' to generate sales forecast for next 12 months .")
|
130 |
+
|
131 |
+
with gr.Row():
|
132 |
+
file_input = gr.File(label="Upload CSV File", file_types=[".csv"])
|
133 |
+
|
134 |
+
with gr.Row():
|
135 |
+
visualize_btn = gr.Button("Forecast", variant="primary")
|
136 |
+
|
137 |
+
with gr.Row():
|
138 |
+
plot_output = gr.Plot(label="Chronos Forecasting Visualization")
|
139 |
+
|
140 |
+
with gr.Row():
|
141 |
+
plot_output = gr.Plot(label="Chronos Forecasting Visualization")
|
142 |
+
|
143 |
+
visualize_btn.click(
|
144 |
+
fn=process_data,
|
145 |
+
inputs=[file_input],
|
146 |
+
outputs=[plot_output]
|
147 |
+
)
|
148 |
+
|
149 |
+
# Launch the app
|
150 |
+
if __name__ == "__main__":
|
151 |
+
demo.launch()
|