Delete app.py

app.py

@@ -1,83 +0,0 @@
-from transformers import AutoModelForCausalLM, AutoTokenizer
-import gradio as gr
-import torch
-import tempfile  # ✅ Import tempfile to create temp files
-
-# ✅ Load the fastest model on CPU
-model_name = "Salesforce/codegen-350M-mono"  # Fastest model for code review
-tokenizer = AutoTokenizer.from_pretrained(model_name)
-model = AutoModelForCausalLM.from_pretrained(model_name).to("cpu")  # Force CPU mode
-
-import tempfile
-
-import tempfile
-
-import tempfile
-
-def review_code(code_snippet):
-    print("✅ Received Code:", code_snippet)  # Debugging log
-    
-    # ✅ Improved prompt with strict format rules
-    prompt = f"""
-### Instruction:
-You are a Python code reviewer. Your job is to analyze and fix errors in the provided Python code.
-Make necessary corrections such as adding missing return statements, fixing syntax errors, and correcting logical mistakes. 
-ONLY return the corrected function definition—DO NOT generate any new function calls or explanations.
-
-### Input Code:
-{code_snippet}
-
-### Corrected Function:
-"""
-
-    # Process input
-    inputs = tokenizer(prompt, return_tensors="pt").to("cpu")  # Move to CPU
-    outputs = model.generate(
-        **inputs,
-        max_new_tokens=50,  # ✅ Limits AI output to prevent hallucinations
-        do_sample=False,
-        num_beams=4,  # ✅ Higher beams for better correction quality
-        repetition_penalty=3.0  # ✅ Stronger penalty to prevent repeated tokens
-    )
-
-    # Check if the model generated output
-    if outputs is None:
-        print("❌ Model did not generate output!")  # Debugging log
-        return "Error: Model did not generate output."
-
-    reviewed_code = tokenizer.decode(outputs[0], skip_special_tokens=True).strip()
-    print("✅ Generated Code:", reviewed_code)  # Debugging log
-
-    # ✅ Ensure AI output starts with "def" to guarantee it returned a function
-    if not reviewed_code.startswith("def"):
-        return "Error: AI did not return a valid function."
-
-    # ✅ Write reviewed code to a temporary file for download
-    temp_file_path = tempfile.NamedTemporaryFile(delete=False, suffix=".txt").name
-    with open(temp_file_path, "w") as temp_file:
-        temp_file.write(reviewed_code)
-
-    return reviewed_code, temp_file_path  # ✅ Return reviewed code & file path
-
-
-# ✅ Handle user input and return reviewed code
-def check_code(input_code):
-    reviewed_code, file_path = review_code(input_code)
-    return input_code, reviewed_code, file_path  # ✅ Correctly return file path
-
-# ✅ Gradio UI with Side-by-Side Comparison & Fixed Download Option
-interface = gr.Interface(
-    fn=check_code,
-    inputs=gr.Textbox(label="Enter Python Code"),
-    outputs=[
-        gr.Textbox(label="Original Code", interactive=False),  # Left side
-        gr.Textbox(label="Reviewed Code", interactive=False),  # Right side
-        gr.File(label="Download Reviewed Code")  # ✅ Fixed Download Button
-    ],
-    title="🚀 AI Code Reviewer",
-    description="📌 Enter Python code and get a reviewed version. Download the reviewed code as a file.",
-    allow_flagging="never"
-)
-
-# ✅ Launch app (Fixes font issues and removes `share=True`)
-interface.launch(server_name="0.0.0.0", server_port=7860, show_error=True)