Update app.py

app.py

@@ -1,6 +1,206 @@
 
 
 
+# import gradio as gr
+# import pandas as pd 
+# import numpy as np 
+# import pickle
+# import sklearn
+# from datasets import load_dataset
+# import joblib
+
+
+# # Read the data
+# data = pd.read_csv("mldata.csv")
+
+# # Function to load model based on selection
+# def load_model(model_choice):
+#     if model_choice == "Random Forest":
+#         with open('rfweights (1).pkl', 'rb') as pickleFile:
+#             return pickle.load(pickleFile)
+#     elif model_choice == "Decision Tree":
+#         with open('dtreeweights.pkl', 'rb') as pickleFile:
+#             return pickle.load(pickleFile)
+#     elif model_choice == "Sequential Model":
+#         try:
+#             # Try loading the Sequential model saved using joblib
+#             with open('my_seq_model_second.pkl', 'rb') as pickleFile:
+#                 return joblib.load(pickleFile)
+#         except:
+#             # If joblib loading fails, fallback to TensorFlow loading
+#             return load_model('my_seq_model (1)')
+#     else:
+#         raise ValueError("Invalid model selection")
+
+# # Prepare categorical data (same as original code)
+# categorical_cols = data[[
+#     'certifications',
+#     'workshops',
+#     'Interested subjects',
+#     'interested career area ',
+#     'Type of company want to settle in?',
+#     'Interested Type of Books'
+# ]]
+
+# # Assign category codes
+# for i in categorical_cols:
+#     data[i] = data[i].astype('category')
+#     data[i] = data[i].cat.codes
+
+# # Create reference dictionaries for embeddings (same as original code)
+# def create_embedding_dict(column):
+#     unique_names = list(categorical_cols[column].unique())
+#     unique_codes = list(data[column].unique())
+#     return dict(zip(unique_names, unique_codes))
+
+# certificates_references = create_embedding_dict('certifications')
+# workshop_references = create_embedding_dict('workshops')
+# subjects_interest_references = create_embedding_dict('Interested subjects')
+# career_interest_references = create_embedding_dict('interested career area ')
+# company_intends_references = create_embedding_dict('Type of company want to settle in?')
+# book_interest_references = create_embedding_dict('Interested Type of Books')
+
+# # Prediction function (modified to accept model choice)
+# def rfprediction(model_choice, name, logical_thinking, hackathon_attend, coding_skills, public_speaking_skills,
+#                  self_learning, extra_course, certificate_code, worskhop_code, read_writing_skill, memory_capability,
+#                  subject_interest, career_interest, company_intend, senior_elder_advise, book_interest, introvert_extro,
+#                  team_player, management_technical, smart_hardworker):
+#     # Load the selected model
+#     rfmodel = load_model(model_choice)
+    
+#     # Create DataFrame (same as original code)
+#     df = pd.DataFrame.from_dict(
+#         {
+#             "logical_thinking": [logical_thinking],
+#             "hackathon_attend": [hackathon_attend],
+#             "coding_skills": [coding_skills],
+#             "public_speaking_skills": [public_speaking_skills],
+#             "self_learning": [self_learning],
+#             "extra_course": [extra_course],
+#             "certificate": [certificate_code],
+#             "workshop": [worskhop_code],
+#             "read_writing_skills": [
+#                     (0 if "poor" in read_writing_skill else 1 if "medium" in read_writing_skill else 2)
+#                 ],
+#             "memory_capability": [
+#                     (0 if "poor" in memory_capability else 1 if "medium" in memory_capability else 2)   
+#                 ],
+#             "subject_interest": [subject_interest],
+#             "career_interest": [career_interest],
+#             "company_intend": [company_intend],
+#             "senior_elder_advise": [senior_elder_advise],
+#             "book_interest": [book_interest],
+#             "introvert_extro": [introvert_extro],
+#             "team_player": [team_player],
+#             "management_technical":[management_technical],
+#             "smart_hardworker": [smart_hardworker]
+#         }
+#     )
+    
+#     # Replace string values with numeric representations
+#     df = df.replace({
+#         "certificate": certificates_references, 
+#         "workshop": workshop_references,
+#         "subject_interest": subjects_interest_references,
+#         "career_interest": career_interest_references,
+#         "company_intend": company_intends_references,
+#         "book_interest": book_interest_references
+#     })
+    
+#     # Dummy encoding (same as original code)
+#     userdata_list = df.values.tolist()
+    
+#     # Management-Technical dummy encoding
+#     if(df["management_technical"].values == "Management"):
+#         userdata_list[0].extend([1])
+#         userdata_list[0].extend([0])
+#         userdata_list[0].remove('Management')
+#     elif(df["management_technical"].values == "Technical"):
+#         userdata_list[0].extend([0])
+#         userdata_list[0].extend([1])
+#         userdata_list[0].remove('Technical')
+#     else: 
+#         return "Error in Management-Technical encoding"
+
+#     # Smart-Hard worker dummy encoding
+#     if(df["smart_hardworker"].values == "smart worker"):
+#         userdata_list[0].extend([1])
+#         userdata_list[0].extend([0])
+#         userdata_list[0].remove('smart worker')
+#     elif(df["smart_hardworker"].values == "hard worker"):
+#         userdata_list[0].extend([0])
+#         userdata_list[0].extend([1])
+#         userdata_list[0].remove('hard worker')
+#     else: 
+#         return "Error in Smart-Hard worker encoding"
+
+#     # Prediction
+#     prediction_result_all = rfmodel.predict_proba(userdata_list)
+    
+#     # Create result dictionary
+#     result_list = {
+#         "Applications Developer": float(prediction_result_all[0][0]),
+#         "CRM Technical Developer": float(prediction_result_all[0][1]),
+#         "Database Developer": float(prediction_result_all[0][2]),
+#         "Mobile Applications Developer": float(prediction_result_all[0][3]),
+#         "Network Security Engineer": float(prediction_result_all[0][4]),
+#         "Software Developer": float(prediction_result_all[0][5]),
+#         "Software Engineer": float(prediction_result_all[0][6]),
+#         "Software Quality Assurance (QA)/ Testing": float(prediction_result_all[0][7]),
+#         "Systems Security Administrator": float(prediction_result_all[0][8]),
+#         "Technical Support": float(prediction_result_all[0][9]),
+#         "UX Designer": float(prediction_result_all[0][10]),
+#         "Web Developer": float(prediction_result_all[0][11]),
+#     }
+#     return result_list
+
+# # Lists for dropdown menus (same as original code)
+# cert_list = ["app development", "distro making", "full stack", "hadoop", "information security", "machine learning", "python", "r programming", "shell programming"]
+# workshop_list = ["cloud computing", "data science", "database security", "game development", "hacking", "system designing", "testing", "web technologies"]
+# skill = ["excellent", "medium", "poor"]
+# subject_list = ["cloud computing", "Computer Architecture", "data engineering", "hacking", "IOT", "Management", "networks", "parallel computing", "programming", "Software Engineering"]
+# career_list = ["Business process analyst", "cloud computing", "developer", "security", "system developer", "testing"]
+# company_list = ["BPA", "Cloud Services", "Finance", "Product based", "product development", "SAaS services", "Sales and Marketing", "Service Based", "Testing and Maintainance Services", "Web Services"]
+# book_list = ["Action and Adventure", "Anthology", "Art", "Autobiographies", "Biographies", "Childrens", "Comics","Cookbooks","Diaries","Dictionaries","Drama","Encyclopedias","Fantasy","Guide","Health","History","Horror","Journals","Math","Mystery","Poetry","Prayer books","Religion-Spirituality","Romance","Satire","Science","Science fiction","Self help","Series","Travel","Trilogy"]
+# Choice_list = ["Management", "Technical"]
+# worker_list = ["hard worker", "smart worker"]
+
+# # Create Gradio interface (modified to include model selection)
+# demo = gr.Interface(
+#     fn=rfprediction, 
+#     inputs=[
+#         gr.Dropdown(["Random Forest", "Decision Tree","Sequential Model"], label="Select Machine Learning Model"),
+#         gr.Textbox(placeholder="What is your name?", label="Name"),
+#         gr.Slider(minimum=1, maximum=9, value=3, step=1, label="Are you a logical thinking person?", info="Scale: 1 - 9"),
+#         gr.Slider(minimum=0, maximum=6, value=0, step=1, label="Do you attend any Hackathons?", info="Scale: 0 - 6 | 0 - if not attended any"),
+#         gr.Slider(minimum=1, maximum=9, value=5, step=1, label="How do you rate your coding skills?", info="Scale: 1 - 9"),
+#         gr.Slider(minimum=1, maximum=9, value=3, step=1, label="How do you rate your public speaking skills/confidency?", info="Scale: 1 - 9"),
+#         gr.Radio({"Yes", "No"}, type="index", label="Are you a self-learning person? *"),
+#         gr.Radio({"Yes", "No"}, type="index", label="Do you take extra courses in uni (other than IT)? *"),
+#         gr.Dropdown(cert_list, label="Select a certificate you took!"),
+#         gr.Dropdown(workshop_list, label="Select a workshop you attended!"),
+#         gr.Dropdown(skill, label="Select your read and writing skill"),
+#         gr.Dropdown(skill, label="Is your memory capability good?"),
+#         gr.Dropdown(subject_list, label="What subject you are interested in?"),
+#         gr.Dropdown(career_list, label="Which IT-Career do you have interests in?"),
+#         gr.Dropdown(company_list, label="Do you have any interested company that you intend to settle in?"),
+#         gr.Radio({"Yes", "No"}, type="index", label="Do you ever seek any advices from senior or elders? *"),
+#         gr.Dropdown(book_list, label="Select your interested genre of book!"),
+#         gr.Radio({"Yes", "No"}, type="index", label="Are you an Introvert?| No - extrovert *"),
+#         gr.Radio({"Yes", "No"}, type="index", label="Ever worked in a team? *"),
+#         gr.Dropdown(Choice_list, label="Which area do you prefer: Management or Technical?"),
+#         gr.Dropdown(worker_list, label="Are you a Smart worker or Hard worker?")
+#     ],
+#     outputs=gr.Label(num_top_classes=5), 
+#     title="IT-Career Recommendation System: TMI4033 Colletive Intelligence, Group 12",
+#     description="Members: Derrick Lim Kin Yeap 74597, Jason Jong Sheng Tat 75125, Jason Ng Yong Xing 75127, Muhamad Hazrie Bin Suhkery 73555 "
+# )
+
+# # Main execution
+# if __name__ == "__main__":
+#     demo.launch(share=True)
+
+
 import gradio as gr
 import pandas as pd 
 import numpy as np 
@@ -8,7 +208,7 @@ import pickle
 import sklearn
 from datasets import load_dataset
 import joblib
-
+import tensorflow as tf
 
 # Read the data
 data = pd.read_csv("mldata.csv")
@@ -24,15 +224,15 @@ def load_model(model_choice):
     elif model_choice == "Sequential Model":
         try:
             # Try loading the Sequential model saved using joblib
-            with open('my_seq_model_second.pkl', 'rb') as pickleFile:
-                return joblib.load(pickleFile)
+            model = joblib.load('my_seq_model_second.pkl')
         except:
             # If joblib loading fails, fallback to TensorFlow loading
-            return load_model('my_seq_model (1)')
+            model = tf.keras.models.load_model('my_seq_model (1)')
+        return model
     else:
         raise ValueError("Invalid model selection")
 
-# Prepare categorical data (same as original code)
+# Prepare categorical data
 categorical_cols = data[[
     'certifications',
     'workshops',
@@ -47,7 +247,7 @@ for i in categorical_cols:
     data[i] = data[i].astype('category')
     data[i] = data[i].cat.codes
 
-# Create reference dictionaries for embeddings (same as original code)
+# Create reference dictionaries for embeddings
 def create_embedding_dict(column):
     unique_names = list(categorical_cols[column].unique())
     unique_codes = list(data[column].unique())
@@ -60,7 +260,7 @@ career_interest_references = create_embedding_dict('interested career area ')
 company_intends_references = create_embedding_dict('Type of company want to settle in?')
 book_interest_references = create_embedding_dict('Interested Type of Books')
 
-# Prediction function (modified to accept model choice)
+# Prediction function
 def rfprediction(model_choice, name, logical_thinking, hackathon_attend, coding_skills, public_speaking_skills,
                  self_learning, extra_course, certificate_code, worskhop_code, read_writing_skill, memory_capability,
                  subject_interest, career_interest, company_intend, senior_elder_advise, book_interest, introvert_extro,
@@ -68,7 +268,7 @@ def rfprediction(model_choice, name, logical_thinking, hackathon_attend, coding_
     # Load the selected model
     rfmodel = load_model(model_choice)
     
-    # Create DataFrame (same as original code)
+    # Create DataFrame
     df = pd.DataFrame.from_dict(
         {
             "logical_thinking": [logical_thinking],
@@ -107,7 +307,7 @@ def rfprediction(model_choice, name, logical_thinking, hackathon_attend, coding_
         "book_interest": book_interest_references
     })
     
-    # Dummy encoding (same as original code)
+    # Dummy encoding
     userdata_list = df.values.tolist()
     
     # Management-Technical dummy encoding
@@ -134,27 +334,29 @@ def rfprediction(model_choice, name, logical_thinking, hackathon_attend, coding_
     else: 
         return "Error in Smart-Hard worker encoding"
 
-    # Prediction
-    prediction_result_all = rfmodel.predict_proba(userdata_list)
+    # Prediction handling for different model types
+    if model_choice in ["Random Forest", "Decision Tree"]:
+        prediction_result_all = rfmodel.predict_proba(userdata_list)
+    else:  # Sequential Model (Keras)
+        prediction_result_all = rfmodel.predict(userdata_list)
     
     # Create result dictionary
-    result_list = {
-        "Applications Developer": float(prediction_result_all[0][0]),
-        "CRM Technical Developer": float(prediction_result_all[0][1]),
-        "Database Developer": float(prediction_result_all[0][2]),
-        "Mobile Applications Developer": float(prediction_result_all[0][3]),
-        "Network Security Engineer": float(prediction_result_all[0][4]),
-        "Software Developer": float(prediction_result_all[0][5]),
-        "Software Engineer": float(prediction_result_all[0][6]),
-        "Software Quality Assurance (QA)/ Testing": float(prediction_result_all[0][7]),
-        "Systems Security Administrator": float(prediction_result_all[0][8]),
-        "Technical Support": float(prediction_result_all[0][9]),
-        "UX Designer": float(prediction_result_all[0][10]),
-        "Web Developer": float(prediction_result_all[0][11]),
-    }
+    careers = [
+        "Applications Developer", "CRM Technical Developer", "Database Developer", 
+        "Mobile Applications Developer", "Network Security Engineer", "Software Developer", 
+        "Software Engineer", "Software Quality Assurance (QA)/ Testing", 
+        "Systems Security Administrator", "Technical Support", "UX Designer", "Web Developer"
+    ]
+    
+    # Handle probability extraction based on model type
+    if model_choice in ["Random Forest", "Decision Tree"]:
+        result_list = {career: float(prediction_result_all[0][i]) for i, career in enumerate(careers)}
+    else:  # Sequential Model
+        result_list = {career: float(prediction_result_all[0][i]) for i, career in enumerate(careers)}
+    
     return result_list
 
-# Lists for dropdown menus (same as original code)
+# Lists for dropdown menus
 cert_list = ["app development", "distro making", "full stack", "hadoop", "information security", "machine learning", "python", "r programming", "shell programming"]
 workshop_list = ["cloud computing", "data science", "database security", "game development", "hacking", "system designing", "testing", "web technologies"]
 skill = ["excellent", "medium", "poor"]
@@ -165,7 +367,7 @@ book_list = ["Action and Adventure", "Anthology", "Art", "Autobiographies", "Bio
 Choice_list = ["Management", "Technical"]
 worker_list = ["hard worker", "smart worker"]
 
-# Create Gradio interface (modified to include model selection)
+# Create Gradio interface
 demo = gr.Interface(
     fn=rfprediction, 
     inputs=[
@@ -199,4 +401,3 @@ demo = gr.Interface(
 # Main execution
 if __name__ == "__main__":
     demo.launch(share=True)
-