Update app.py

app.py

@@ -6,13 +6,15 @@ import time
 # Load datasets from CSV files
 @st.cache_resource
 def load_csv_datasets():
-    jobs_data = pd.read_csv("job_data.csv")
+    jobs_data = pd.read_csv("job_descriptions.csv")
     courses_data = pd.read_csv("courses_data.csv")
     return jobs_data, courses_data
 
 jobs_data, courses_data = load_csv_datasets()
+
 # Constants
 universities_url = "https://www.4icu.org/top-universities-world/"
+
 # Initialize the text generation pipeline
 @st.cache_resource
 def load_pipeline():
@@ -126,26 +128,64 @@ if "profile_data" in st.session_state:
             # Extracting user profile data
             profile = st.session_state.profile_data
             user_tech_skills = set(skill.strip().lower() for skill in profile["tech_skills"].split(","))
+            user_soft_skills = set(skill.strip().lower() for skill in profile["soft_skills"].split(","))
             user_interests = set(interest.strip().lower() for interest in profile["interests"].split(","))
-
-            # Job Recommendations using RAG technique
-            job_recommendations = jobs_data[jobs_data['Skills Required'].apply(
-                lambda skills: any(skill in skills.lower() for skill in user_tech_skills)
-            )]
+            user_answers = st.session_state.get('answers', {})
+
+            # Job Recommendations using refined scoring logic
+            def match_job_criteria(row, profile, user_answers):
+                job_title = row['Job Title'].lower()
+                job_description = row['Job Description'].lower()
+                qualifications = row['Qualifications'].lower()
+                skills = row['skills'].lower()
+                role = row['Role'].lower()
+
+                educational_background = profile['educational_background'].lower()
+                tech_skills = set(skill.strip().lower() for skill in profile["tech_skills"].split(","))
+                soft_skills = set(skill.strip().lower() for skill in profile["soft_skills"].split(","))
+                interests = set(interest.strip().lower() for interest in profile["interests"].split(","))
+                user_answers_text = ' '.join(user_answers.values()).lower()
+
+                score = 0
+
+                if educational_background in qualifications or educational_background in job_description:
+                    score += 2
+                if any(skill in skills for skill in tech_skills):
+                    score += 3
+                if any(skill in job_description or role for skill in soft_skills):
+                    score += 1
+                if any(interest in job_title or job_description for interest in interests):
+                    score += 2
+                if any(answer in job_description or qualifications for answer in user_answers_text.split()):
+                    score += 2
+
+                return score >= 5
+
+            # Get unique job recommendations
+            job_recommendations = jobs_data[jobs_data.apply(lambda row: match_job_criteria(row, profile, user_answers), axis=1)]
+            unique_jobs = job_recommendations.drop_duplicates(subset=['Job Title'])
+
+            # Display Job Recommendations in a table with bullet points
+            st.subheader("Job Recommendations")
+            if not unique_jobs.empty:
+                job_list = unique_jobs.head(5)[['Job Title', 'Job Description']].reset_index(drop=True)
+                job_list.index += 1  # Start index from 1
+                job_list.columns = ["Job Title", "Description"]
+                st.table(job_list)
+            else:
+                st.write("No specific job recommendations found matching your profile.")
+                st.write("Here are some general job recommendations:")
+                fallback_jobs = jobs_data.drop_duplicates(subset=['Job Title']).head(3)
+                fallback_list = fallback_jobs[['Job Title', 'Job Description']].reset_index(drop=True)
+                fallback_list.index += 1  # Start index from 1
+                fallback_list.columns = ["Job Title", "Description"]
+                st.table(fallback_list)
 
             # Course Recommendations using RAG technique
             course_recommendations = courses_data[courses_data['Course Name'].apply(
                 lambda name: any(interest in name.lower() for interest in user_interests)
             )]
 
-            # Display Job Recommendations
-            st.subheader("Job Recommendations")
-            if not job_recommendations.empty:
-                for _, row in job_recommendations.head(5).iterrows():
-                    st.write(f"- **{row['Job Title']}**: {row['Description']}")
-            else:
-                st.write("No specific job recommendations found matching your profile.")
-
             # Display Course Recommendations
             st.subheader("Recommended Courses")
             if not course_recommendations.empty:
@@ -155,7 +195,6 @@ if "profile_data" in st.session_state:
                 st.write("No specific course recommendations found matching your interests.")
                 st.write("Here are some general course recommendations aligned with your profile:")
 
-                # Suggest 3 fallback courses aligned with the user's educational background or technical skills
                 fallback_courses = courses_data[
                     courses_data['Course Name'].apply(
                         lambda name: any(
@@ -174,6 +213,6 @@ if "profile_data" in st.session_state:
 # University Recommendations Section
 st.header("Top Universities")
 st.write("For further education, you can explore the top universities worldwide:")
-st.write(f"[View Top Universities Rankings]({universities_url})")                    
+st.write(f"[View Top Universities Rankings]({universities_url})")
 
 st.write("Thank you for using the Career Counseling Application with RAG!")