Delete generate_schema.py
Browse files- generate_schema.py +0 -45
generate_schema.py
DELETED
@@ -1,45 +0,0 @@
|
|
1 |
-
import requests
|
2 |
-
import json
|
3 |
-
import os
|
4 |
-
import os
|
5 |
-
from dotenv import load_dotenv
|
6 |
-
import streamlit as st
|
7 |
-
|
8 |
-
API_KEY = st.secrets["hf_token"]
|
9 |
-
|
10 |
-
|
11 |
-
|
12 |
-
def generate_schema(user_prompt):
|
13 |
-
""" Generates a synthetic dataset schema using Hugging Face API. """
|
14 |
-
|
15 |
-
system_prompt = """
|
16 |
-
You are an expert data scientist designing synthetic datasets.
|
17 |
-
For any given dataset description, generate:
|
18 |
-
- Column names
|
19 |
-
- Data types (string, int, float, date)
|
20 |
-
- Approximate row count
|
21 |
-
|
22 |
-
Output in **pure JSON** format like:
|
23 |
-
{
|
24 |
-
"columns": ["PatientID", "Age", "Gender", "Diagnosis"],
|
25 |
-
"types": ["int", "int", "string", "string"],
|
26 |
-
"size": 500
|
27 |
-
}
|
28 |
-
"""
|
29 |
-
|
30 |
-
payload = {
|
31 |
-
"inputs": system_prompt + "\n\nUser request: " + user_prompt,
|
32 |
-
"options": {"wait_for_model": True}
|
33 |
-
}
|
34 |
-
|
35 |
-
response = requests.post(HF_MODEL_URL, headers={"Authorization": f"Bearer {API_KEY}"}, json=payload)
|
36 |
-
|
37 |
-
if response.status_code == 200:
|
38 |
-
try:
|
39 |
-
output = response.json()[0]['generated_text']
|
40 |
-
schema = json.loads(output.strip()) # Convert to JSON
|
41 |
-
return schema
|
42 |
-
except json.JSONDecodeError:
|
43 |
-
return {"error": "Invalid JSON output from model. Try again."}
|
44 |
-
else:
|
45 |
-
return {"error": f"API request failed. Status Code: {response.status_code}"}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|