Update to allow pre generated questions

app.py

@@ -2,17 +2,21 @@ import streamlit as st
 from streamlit_js_eval import streamlit_js_eval
 from azure.storage.blob import BlobServiceClient
 from azure.cosmos import CosmosClient, exceptions
+from PyPDF2 import PdfReader
+import io
+import openai
 import json
 import os
 import uuid
 import time
 import calendar
+import re
 
 connection_string = os.getenv("CONNECTION")
 blob_service_client = BlobServiceClient.from_connection_string(connection_string)
 
 
-def upload_blob(pdf_name, json_data, pdf_data_jobdescription,pdf_data_cvs):
+def upload_blob(pdf_name, json_data, pdf_data_jobdescription,pdf_data_cvs, pre_generated_bool, custom_questions):
     try:
         container_name = "jobdescriptions"
         # json_blob_name = f"{pdf_name}_jsondata.json"
@@ -26,7 +30,7 @@ def upload_blob(pdf_name, json_data, pdf_data_jobdescription,pdf_data_cvs):
         pdf_blob_client = container_client.get_blob_client(pdf_blob_name_jobdescription)
         pdf_blob_client.upload_blob(pdf_data_jobdescription, overwrite=True)
 
-        upload_job_db_item(pdf_name,len(pdf_data_cvs),json.loads(json_data))
+        upload_job_db_item(pdf_name,len(pdf_data_cvs),json.loads(json_data),pre_generated_bool, custom_questions)
 
         links = []
         names = []
@@ -53,7 +57,7 @@ def upload_blob(pdf_name, json_data, pdf_data_jobdescription,pdf_data_cvs):
         print(f"Fehler beim Hochladen der Daten: {str(e)}")
         return False
 
-def upload_job_db_item(id, number_of_applicants, data):
+def upload_job_db_item(id, number_of_applicants, data, pre_generated_bool, custom_questions):
     endpoint = "https://wg-candidate-data.documents.azure.com:443/"
     key = os.getenv("CONNECTION_DB")
     client = CosmosClient(endpoint, key)
@@ -69,6 +73,8 @@ def upload_job_db_item(id, number_of_applicants, data):
         "question_one": data["question_one"],
         "question_two": data["question_two"],
         "question_three": data["question_three"],
+        "pre_generated": pre_generated_bool,
+        "custom_questions": custom_questions
     }
     try:
         # Fügen Sie das Element in den Container ein
@@ -109,13 +115,6 @@ def upload_db_item(name, data, job_description_id, cv_id):
     except Exception as e:
         print(f"Allgemeiner Fehler: {str(e)}")
 
-# def clear_states():
-#     if len(st.session_state.title) > 0 and len(st.session_state.mail) > 0 and st.session_state.job and len(st.session_state.cvs)>0:
-#         st.session_state.title = ""
-#         st.session_state.mail = ""
-#         # st.session_state.job = None
-#         st.session_state.cvs = []
-
 st.markdown(
 """
 <style>
@@ -131,26 +130,122 @@ st.markdown(
 )
 col1, col2 = st.columns([2, 1])
 
+if "ai_questions" not in st.session_state:
+    st.session_state["ai_questions"] = None
+if "pdf_data_cvs" not in st.session_state:
+    st.session_state["pdf_data_cvs"] = None
+if "pdf_data_cvs_string" not in st.session_state:
+    st.session_state["pdf_data_cvs_string"] = None
+if "pdf_data_jobdescription" not in st.session_state:
+    st.session_state["pdf_data_jobdescription"] = None
+if "pdf_data_jobdescription_string" not in st.session_state:
+    st.session_state["pdf_data_jobdescription_string"] = None
+if "final_question_string" not in st.session_state:
+    st.session_state["final_question_string"] = []
+
+with open("sys_prompt_frontend.txt") as f:
+    sys_prompt = f.read()
 
 col1.title("Job description upload")
 col2.image("https://www.workgenius.com/wp-content/uploads/2023/03/WorkGenius_navy-1.svg")
 
 st.write("Please upload the job description and resume(s) as PDF and enter the job title for the position. To receive the evaluation of the potential candidate(s), please provide your email address.")
 upload_success = True    
+
+#This container represents the form 
 with st.container():
+
+    #Form section for the files, names, title and mail
     uploaded_file_jobdescription = st.file_uploader("Upload the job description:", type=["pdf"], key="job")
     job_title = st.text_input("Enter the job title:", key="title")
     email = st.text_input("Enter the email:" , key="mail")
     uploaded_file_cvs = st.file_uploader("Upload the resume(s):", type=["pdf"],accept_multiple_files=True,  key="cvs")
     for i,cv in enumerate(st.session_state["cvs"]):
         st.text_input(label="Enter the name of the "+str(i+1)+". CV (File: "+cv.name+")", value=cv.name,key="cv-"+str(i+1))
-    with st.expander("Enter up to three predefined questions if needed. Otherwise leave it blank:"):
-            question_one = st.text_input("Enter the first question:")
-            question_two = st.text_input("Enter the second question:")
-            question_three = st.text_input("Enter the third question:")
+    
+    #Form section for the interview mode (pre generated or not) and additional questions
+    if len(job_title) > 0 and len(email) > 0 and uploaded_file_jobdescription and len(uploaded_file_cvs)>0:
+        st.write("Activate the toggle to generate and select the questions in advance. Otherwise the questions will be generated automatically during the interview.")
+        if not st.session_state["pdf_data_cvs"] and not st.session_state["pdf_data_cvs_string"] and not st.session_state["pdf_data_jobdescription"] and not st.session_state["pdf_data_jobdescription_string"]:
+            pdf_data_jobdescription = uploaded_file_jobdescription.read()
+            pdf_data_jobdescription_string = ""
+            pdf_reader_job = PdfReader(io.BytesIO(pdf_data_jobdescription))
+            for page_num in range(len(pdf_reader_job.pages)):
+                    page = pdf_reader_job.pages[page_num]
+                    pdf_data_jobdescription_string += page.extract_text()
+            pdf_data_cvs = []
+            pdf_data_cvs_string = ""
+            for i,cv in enumerate(st.session_state["cvs"]):
+                print(cv.name)
+                # print(cv.name)
+                # print(cv.size)
+                cv_data_bytes = cv.read()
+                # print(len(cv_data_bytes))
+                pdf_data_cvs.append(cv_data_bytes)
+                pdf_reader_cvs = PdfReader(io.BytesIO(cv_data_bytes))
+                pdf_data_cvs_string += "CV "+str(i+1)+": "
+                for page_num in range(len(pdf_reader_cvs.pages)):
+                    page = pdf_reader_cvs.pages[page_num]
+                    pdf_data_cvs_string += page.extract_text()
+                pdf_data_cvs_string += "\n"
+            st.session_state["pdf_data_cvs"] = pdf_data_cvs
+            st.session_state["pdf_data_cvs_string"] = pdf_data_cvs_string
+            st.session_state["pdf_data_jobdescription"] = pdf_data_jobdescription
+            st.session_state["pdf_data_jobdescription_string"] = pdf_data_jobdescription_string
+        pre_generate = st.toggle("Activate to pre generate questions", key="pre_toggle")
+        if pre_generate:
+            system = sys_prompt.format(job=st.session_state["pdf_data_jobdescription_string"], resume=st.session_state["pdf_data_cvs_string"], n=15)
+            if not st.session_state["ai_questions"]:
+                try:
+                    st.write("The questions are generated. This may take a short moment...")
+                    res = openai.ChatCompletion.create(
+                        model="gpt-4",
+                        temperature=0.2,
+                        messages=[
+                            {
+                                "role": "system",
+                                "content": system,
+                            },
+                        ],
+                        )
+                    st.session_state["ai_questions"] = res.choices[0]["message"]["content"].split("\n")
+                    for i,q in enumerate(res.choices[0]["message"]["content"].split("\n")):
+                        st.session_state["disable_row_"+str(i)] = False
+                    st.rerun()
+                except Exception as e:
+                    print(f"Fehler beim generieren der Fragen: {str(e)}")
+                    st.error("An error has occurred. Please reload the page or contact the admin.", icon="🚨")
+            else:
+                for i,question in enumerate(st.session_state["ai_questions"]):
+                    cols = st.columns([5,1])
+                    with cols[1]:
+                        if st.button("Accept",use_container_width=True,key="btn_accept_row_"+str(i)):
+                            print("accept")
+                            pattern = re.compile(r"^[1-9][0-9]?\.")
+                            questions_length = len(st.session_state["final_question_string"])
+                            question_from_text_area = st.session_state["text_area_"+str(i)]
+                            question_to_append = str(questions_length+1)+"."+re.sub(pattern, "", question_from_text_area)
+                            st.session_state["final_question_string"].append(question_to_append)
+                            st.session_state["disable_row_"+str(i)] = True
+                            st.rerun()
+                        if st.button("Delete",use_container_width=True,key="btn_del_row_"+str(i)):
+                            print("delete")
+                            st.session_state["ai_questions"].remove(question)
+                            st.rerun()
+                    with cols[0]:
+                        st.text_area(label="Question "+str(i+1)+":",value=question,label_visibility="collapsed",key="text_area_"+str(i),disabled=st.session_state["disable_row_"+str(i)])
+        else:
+            with st.expander("Enter up to three predefined questions if needed. Otherwise leave it blank:"):
+                    question_one = st.text_input("Enter the first question:")
+                    question_two = st.text_input("Enter the second question:")
+                    question_three = st.text_input("Enter the third question:")
+    
+    #Form section for Submit and Clear
     col_submit_btn, col_empty, col_clear_btn = st.columns([1,4, 1])
     if col_clear_btn.button("Clear " ,use_container_width=True):
         streamlit_js_eval(js_expressions="parent.window.location.reload()")
+    
+    #Code to handle the input
     if col_submit_btn.button("Submit", use_container_width=True):
         if len(job_title) > 0 and len(email) > 0 and uploaded_file_jobdescription and len(uploaded_file_cvs)>0:
             data = {
@@ -160,12 +255,13 @@ with st.container():
                 "question_two": "",
                 "question_three": "",
             }
-            if question_one:
-                data["question_one"] = question_one
-            if question_two:
-                data["question_two"] = question_two
-            if question_three:
-                data["question_three"] = question_three
+            if not st.session_state["pre_toggle"]:
+                if question_one:
+                    data["question_one"] = question_one
+                if question_two:
+                    data["question_two"] = question_two
+                if question_three:
+                    data["question_three"] = question_three
                 
             json_data = json.dumps(data, ensure_ascii=False)
 
@@ -177,14 +273,11 @@ with st.container():
 
             pdf_name = uuid_string
 
-            pdf_data_jobdescription = uploaded_file_jobdescription.read()
-            pdf_data_cvs = []
-            for i,cv in enumerate(st.session_state["cvs"]):
-                print(cv.name)
-                pdf_data_cvs.append(cv.read())
+            print(st.session_state["final_question_string"])
+            
             # pdf_data_cv = uploaded_file_cv.read()
 
-            upload_success = upload_blob(pdf_name, json_data, pdf_data_jobdescription,pdf_data_cvs)
+            upload_success = upload_blob(pdf_name, json_data, st.session_state["pdf_data_jobdescription"],st.session_state["pdf_data_cvs"],st.session_state["pre_toggle"],st.session_state["final_question_string"])
         else:
             st.write("Please fill out both fields and upload a PDF file.")