Spaces:
Sleeping
Sleeping
| import gradio as gr | |
| from llama_cpp import Llama | |
| from transformers import pipeline | |
| import json | |
| # Load the Llama model | |
| try: | |
| llm = Llama.from_pretrained( | |
| repo_id="HuggingFaceTB/SmolLM2-360M-Instruct-GGUF", | |
| filename="smollm2-360m-instruct-q8_0.gguf" # Replace with the correct path to your GGUF file | |
| ) | |
| except Exception as e: | |
| raise RuntimeError(f"Failed to load model: {e}") | |
| # Load summarization model | |
| summarizer = pipeline("summarization") | |
| # Summarize text to fit within token limits | |
| def summarize_text(text, max_length=100): | |
| # Use the summarizer to condense the text | |
| summary = summarizer(text, max_length=max_length, min_length=25, do_sample=False) | |
| return summary[0]["summary_text"] | |
| # Function to match CV to job descriptions with debug information | |
| def match_cv_to_jobs(cv_text, job_descriptions): | |
| debug_info = "Debug Info:\n" | |
| results = [] | |
| # Summarize `cv_text` and `job_descriptions` to manage token limits | |
| summarized_cv = summarize_text(cv_text, max_length=400) | |
| debug_info += f"Summarized CV Text: {summarized_cv}\n" | |
| descriptions = job_descriptions.strip().split("\n") | |
| for description in descriptions: | |
| summarized_description = summarize_text(description, max_length=100) | |
| debug_info += f"\nSummarized Job Description: {summarized_description}\n" | |
| # Create a prompt to compare the summarized CV with each summarized job description | |
| prompt = ( | |
| f"Compare the following job description with this resume. Job Description: {summarized_description}. " | |
| f"Resume: {summarized_cv}. Provide a match score and a brief analysis." | |
| ) | |
| debug_info += f"\nGenerated Prompt: {prompt}\n" | |
| # Generate response from the model | |
| try: | |
| response = llm.create_chat_completion( | |
| messages=[ | |
| { | |
| "role": "user", | |
| "content": prompt | |
| } | |
| ] | |
| ) | |
| # Extract the analysis text | |
| response_content = response["choices"][0]["message"]["content"] | |
| debug_info += f"Model Response: {response_content}\n" | |
| # Attempt to parse as JSON; if not JSON, use the raw text | |
| try: | |
| response_data = json.loads(response_content) | |
| results.append(response_data) | |
| except json.JSONDecodeError: | |
| results.append({ | |
| "Job Description": description, | |
| "Analysis": response_content | |
| }) | |
| except Exception as e: | |
| debug_info += f"Error: {str(e)}\n" | |
| results.append({"Job Description": description, "Error": str(e)}) | |
| return results, debug_info | |
| # Gradio interface | |
| with gr.Blocks() as demo: | |
| gr.Markdown("# CV and Job Description Matcher with Summarization and Debugging") | |
| # Input fields for CV and job descriptions | |
| cv_text = gr.Textbox(label="CV Text", placeholder="Enter the CV text here", lines=10) | |
| job_descriptions = gr.Textbox(label="Job Descriptions (one per line)", placeholder="Enter each job description on a new line", lines=5) | |
| # Button and output area | |
| match_button = gr.Button("Match CV to Job Descriptions") | |
| output = gr.JSON(label="Match Results") | |
| debug_output = gr.Textbox(label="Debug Info", lines=10) # Add a debug box to display debug info | |
| # Set button click to run the function | |
| match_button.click(fn=match_cv_to_jobs, inputs=[cv_text, job_descriptions], outputs=[output, debug_output]) | |
| demo.launch() | |