Update app.py

app.py

@@ -2,20 +2,54 @@ import gradio as gr
 import pandas as pd
 import datetime
 import numpy as np
-from transformers import pipeline
+import docx
+from PyPDF2 import PdfReader
+from sentence_transformers import SentenceTransformer, util
 
 class AIHRAgent:
     def __init__(self):
-        self.resume_scanner = pipeline("zero-shot-classification", model="facebook/bart-large-mnli")
+        # Advanced model for semantic similarity
+        self.resume_model = SentenceTransformer("sentence-transformers/all-MiniLM-L6-v2")
         self.employee_records = pd.DataFrame(columns=["Name", "Position", "Start Date", "Attendance", "Performance", "Leaves"])
         self.company_policies = "Employees are entitled to 24 annual leaves and must adhere to company policies regarding attendance and punctuality."
 
+    def extract_text_from_file(self, file_path):
+        """Extract text from uploaded file (PDF or DOCX)."""
+        try:
+            if file_path.name.endswith(".pdf"):
+                pdf_reader = PdfReader(file_path)
+                text = " ".join(page.extract_text() for page in pdf_reader.pages if page.extract_text())
+            elif file_path.name.endswith(".docx"):
+                doc = docx.Document(file_path)
+                text = " ".join(paragraph.text for paragraph in doc.paragraphs)
+            else:
+                raise ValueError("Unsupported file format. Please upload a PDF or DOCX file.")
+            return text
+        except Exception as e:
+            return f"Error extracting text from file: {e}"
+
     def screen_resume(self, resume_text, job_description):
-        results = self.resume_scanner(resume_text, candidate_labels=[job_description, "Not Relevant"])
-        return f"Relevance Score: {results['scores'][0]:.2f} for the position of {job_description}."
+        """Advanced resume screening using sentence embeddings."""
+        try:
+            if not resume_text or not job_description:
+                return "Please provide both the resume text and job description."
+
+            # Semantic similarity scoring
+            job_embedding = self.resume_model.encode(job_description, convert_to_tensor=True)
+            resume_embedding = self.resume_model.encode(resume_text, convert_to_tensor=True)
+            similarity = util.pytorch_cos_sim(job_embedding, resume_embedding).item()
+            return f"Relevance Score: {similarity:.2f} for the position of {job_description}."
+        except Exception as e:
+            return f"Error during resume screening: {e}"
 
     def onboarding_guide(self, employee_name, position):
-        return f"Welcome {employee_name}! As a {position}, your first day involves orientation, meeting the team, and setting up your work systems."
+        """Automated onboarding guide generation."""
+        return (f"Welcome {employee_name}!\n"
+                f"As a {position}, your onboarding plan includes:\n"
+                f"1. Orientation session.\n"
+                f"2. Team introductions.\n"
+                f"3. Work system setup.\n"
+                f"4. Initial training and goal setting.")
 
     def add_employee(self, name, position, start_date):
         new_employee = {
@@ -64,23 +98,28 @@ class AIHRAgent:
             return f"Exit interview recorded for {employee_name}. Feedback: {feedback}"
         return f"Employee {employee_name} not found."
 
+# AI HR Agent Instance
 ai_hr = AIHRAgent()
 
+# Gradio Interface
 def gradio_interface():
     with gr.Blocks() as interface:
-        gr.Markdown("# **AI HR Agent**")
-        gr.Markdown("Automate all HR functionalities with an intelligent AI agent.")
+        gr.Markdown("# **Advanced AI HR Agent**")
+        gr.Markdown("This AI automates all HR tasks and provides advanced features such as resume screening and policy management.")
 
         with gr.Tab("Recruitment and Onboarding"):
-            resume_input = gr.Textbox(label="Paste Resume Text")
-            job_description_input = gr.Textbox(label="Job Description")
-            resume_output = gr.Textbox(label="Screening Result")
-            screen_button = gr.Button("Screen Resume")
-
-            onboarding_name = gr.Textbox(label="Employee Name")
-            onboarding_position = gr.Textbox(label="Position")
-            onboarding_output = gr.Textbox(label="Onboarding Guide")
-            onboarding_button = gr.Button("Generate Onboarding Guide")
+            with gr.Row():
+                with gr.Column():
+                    resume_upload = gr.File(label="Upload Resume (PDF/DOCX)")
+                    job_description_input = gr.Textbox(label="Job Description")
+                    resume_screen_output = gr.Textbox(label="Screening Result")
+                    screen_button = gr.Button("Screen Resume")
+
+                with gr.Column():
+                    onboarding_name = gr.Textbox(label="Employee Name")
+                    onboarding_position = gr.Textbox(label="Position")
+                    onboarding_output = gr.Textbox(label="Onboarding Guide")
+                    onboarding_button = gr.Button("Generate Onboarding Guide")
 
         with gr.Tab("Employee Management"):
             add_name = gr.Textbox(label="Employee Name")
@@ -99,42 +138,26 @@ def gradio_interface():
             payroll_output = gr.Textbox(label="Payroll Result")
             payroll_button = gr.Button("Process Payroll")
 
-        with gr.Tab("Employee Engagement"):
-            pulse_output = gr.Textbox(label="Pulse Survey")
-            pulse_button = gr.Button("Get Pulse Survey")
-
-            feedback_scores = gr.Textbox(label="Feedback Scores (comma-separated)")
-            feedback_output = gr.Textbox(label="Feedback Analysis Result")
-            feedback_button = gr.Button("Analyze Feedback")
-
-        with gr.Tab("Performance Management"):
-            review_name = gr.Textbox(label="Employee Name")
-            review_score = gr.Number(label="Review Score")
-            review_output = gr.Textbox(label="Review Result")
-            review_button = gr.Button("Update Performance Review")
-
-        with gr.Tab("Compliance and Policy Management"):
-            policy_output = gr.Textbox(label="Company Policies")
-            policy_button = gr.Button("View Policies")
-
         with gr.Tab("Exit Management"):
             exit_name = gr.Textbox(label="Employee Name")
             exit_feedback = gr.Textbox(label="Exit Feedback")
             exit_output = gr.Textbox(label="Exit Interview Result")
             exit_button = gr.Button("Record Exit Interview")
 
-        screen_button.click(ai_hr.screen_resume, inputs=[resume_input, job_description_input], outputs=resume_output)
+        # Button Actions
+        screen_button.click(
+            lambda file, job_desc: ai_hr.screen_resume(ai_hr.extract_text_from_file(file), job_desc) if file else "No resume file uploaded.",
+            inputs=[resume_upload, job_description_input],
+            outputs=resume_screen_output,
+        )
         onboarding_button.click(ai_hr.onboarding_guide, inputs=[onboarding_name, onboarding_position], outputs=onboarding_output)
         add_button.click(ai_hr.add_employee, inputs=[add_name, add_position, add_start_date], outputs=add_output)
         attendance_button.click(ai_hr.track_attendance, inputs=attendance_name, outputs=attendance_output)
         payroll_button.click(ai_hr.process_payroll, inputs=[payroll_name, payroll_salary], outputs=payroll_output)
-        pulse_button.click(lambda: ai_hr.pulse_survey(), outputs=pulse_output)
-        feedback_button.click(lambda scores: ai_hr.feedback_analysis(list(map(int, scores.split(',')))), inputs=feedback_scores, outputs=feedback_output)
-        review_button.click(ai_hr.performance_review, inputs=[review_name, review_score], outputs=review_output)
-        policy_button.click(lambda: ai_hr.get_policy(), outputs=policy_output)
         exit_button.click(ai_hr.exit_interview, inputs=[exit_name, exit_feedback], outputs=exit_output)
 
     return interface
 
+# Launch Interface
 interface = gradio_interface()
 interface.launch(share=True)