Upload 6 files

PwsResumeClassifier.py

@@ -25,6 +25,8 @@ class PwsResumeClassifier:
     paragraph) the reason why the resume was deemed as a good match and why it got the ranking it did.'''
 
     def __init__(self, organization_id, project_id, api_key):
+        self.thread_chatbot = None
+        self.assistant_chatbot = None
         self.client = OpenAI(
             organization=organization_id,
             project=project_id,
@@ -37,7 +39,6 @@ class PwsResumeClassifier:
             instructions=self.instructions,
             model='gpt-4-turbo',
             tools=[{"type": "file_search"}],
-            temperature=0.1
         )
 
         self.assistant_id = self.assistant.id
@@ -49,18 +50,18 @@ class PwsResumeClassifier:
                         file_name in
                         file_names]
 
-        vector_store = self.client.beta.vector_stores.create(
+        self.vector_store = self.client.beta.vector_stores.create(
             name="Resumes",
         )
 
         file_batch = self.client.beta.vector_stores.file_batches.upload_and_poll(
-            vector_store_id=vector_store.id,
+            vector_store_id=self.vector_store.id,
             files=file_streams
         )
 
         self.assistant = self.client.beta.assistants.update(
             assistant_id=self.assistant_id,
-            tool_resources={"file_search": {"vector_store_ids": [vector_store.id]}},
+            tool_resources={"file_search": {"vector_store_ids": [self.vector_store.id]}},
         )
 
     def get_best_resumes(self, job_description) -> str:
@@ -80,28 +81,63 @@ class PwsResumeClassifier:
 
         regex_patter = r"【.*?】"
         message_content = re.sub(regex_patter, '', message_content)
-        return message_content
 
-    def enter_live_thread(self):
-        thread = self.client.beta.threads.create()
-        while True:
-            text = input("Enter your query:")
+        description_chatbot = r'''You are an assistant that answers questions regarding a set of resumes and a job 
+        posting.'''
+
+        instructions_chatbot = r'''You are provided multiple resumes of applicants applying to a job posting. The job 
+        description of the posting is: ''' + job_description + '''. The criteria 
+        for the judging the best matching resume are the following:\n\n
+        
+        1. How well the applicant's years of experiences matches the years of experience requirement 
+        in the job posting.\n
+        2. How well the applicant's educational degree(s) match the degree requirements in the job posting.\n
+        3. How well the applicant's skills compare to the skills desired in the job posting.\n
+        4. How well the applicant's previous work experiences have relation to the job posting.\n\n
+        
+        Not all of the criteria above may be relevant to the job posting, but use as many criteria as possible.\n\n. You 
+        will provide answers to the user's prompts based on the resumes provides, job description, and criteria outlined above.'''
+
+        self.assistant_chatbot = self.client.beta.assistants.create(
+            name="Resume chatbot",
+            description=description_chatbot,
+            instructions=instructions_chatbot,
+            model="gpt-4o",
+            tools=[{"type": "file_search"}],
+            tool_resources={"file_search": {"vector_store_ids": [self.vector_store.id]}}
+        )
+        self.thread_chatbot = self.client.beta.threads.create()
+
+        return message_content
 
-            if text == 'quit':
-                break
+    def query_assistant(self, user_input):
 
-            message = self.client.beta.threads.messages.create(
-                thread_id=thread.id,
+        try:
+            messages = self.client.beta.threads.messages.create(
+                thread_id=self.thread_chatbot.id,
                 role="user",
-                content=text,
+                content=user_input
             )
 
             run = self.client.beta.threads.runs.create_and_poll(
-                thread_id=thread.id, assistant_id=self.assistant_id
+                thread_id=self.thread_chatbot.id,
+                assistant_id=self.assistant_chatbot.id,
             )
 
-            self.client.beta.threads.delete(thread.id)
-            messages = list(self.client.beta.threads.messages.list(thread_id=thread.id, run_id=run.id))
-            message_content = messages[0].content[0].text
-            print("Response: \n")
-            print(f"{message_content.value}\n")
+            if run.status == 'completed':
+                messages = self.client.beta.threads.messages.list(
+                    thread_id=self.thread_chatbot.id
+                )
+
+            messages = self.client.beta.threads.messages.list(
+                thread_id=self.thread_chatbot.id
+            )
+
+            response_text = messages.data[0].content[0].text.value
+            regex_patter = r"【.*?】"
+            response_text = re.sub(regex_patter, '', response_text)
+            return response_text
+
+        except Exception as e:
+            print(e)
+            return 'My apologies. We could not fulfill your request due to an error.'

app.py

@@ -1,135 +1,95 @@
-from PwsResumeClassifier import PwsResumeClassifier as pws
-import os
-import shutil
-import glob
-import gradio as gr
-from pathlib import Path
-from datetime import datetime
-
-
-all_jobs = []
-job_names = []
-job_descriptions = []
-try:
-    ai = pws(organization_id=os.getenv('OPENAIORG'), project_id=os.getenv('PROJECT'), api_key=os.getenv('APIKEY'))
-except Exception as e:
-    print(e)
-    print('Could not connect to openai client')
-
-
-def check_password(password):
-    if password == os.getenv('MILPASSWORD'):
-
-        return [gr.Textbox(visible=False), gr.Button(visible=False), gr.Row.update(visible=False), gr.Row.update(visible=True)]
-    else:
-        return [gr.Textbox(visible=True), gr.Button(visible=True), gr.Row.update(visible=True), gr.Row.update(visible=False)]
-
-
-def add_resume(resume_list):
-    for resume in resume_list:
-        shutil.copy2(os.path.abspath(resume), os.path.abspath('Resumes'))
-
-
-def remove_resumes(file_explorer):
-    for file in file_explorer:
-        os.remove(file)
-
-
-def job_selected(selected_name):
-
-    if selected_name in job_names:
-        return [selected_name, job_descriptions[job_names.index(selected_name)], gr.Row.update(visible=True),
-                gr.Markdown()]
-    elif selected_name == 'Custom':
-        return ['', '', gr.Row.update(visible=True), gr.Markdown()]
-    else:
-        return ['', '', gr.Row.update(visible=False), gr.Markdown()]
-
-
-def send_to_openai(name, description):
-    try:
-        print('Sending to AI')
-        ai.set_vector_store(os.path.abspath('Resumes'))
-        resumes_output = ai.get_best_resumes('Job Name: ' + name + '. Job Description: ' + description)
-        response = gr.Markdown('<p>' + resumes_output + '</p>')
-        return [response, gr.Row.update(visible=True)]
-    except Exception as e:
-        response = gr.Markdown('<p>An error occurred. Please try again.</p>')
-        return [response, gr.Row.update(visible=True)]
-
-
-with gr.Blocks(theme=gr.themes.Soft()) as demo:
-    files = glob.glob(os.path.abspath('Jobs') + "/*")
-    for f in files:
-        with open(f, 'r') as f_job:
-            job_description = f_job.read()
-            job_name = Path(f).stem
-
-            job_names.append(job_name)
-            job_descriptions.append(job_description)
-
-    gr.Markdown("""<h1><center>Resume Processing</center></h1>""")
-    password_row = gr.Row()
-    password_text_row = gr.Row(visible=False)
-    content_row = gr.Row(visible=False)
-    confirmation_row = gr.Row(visible=False)
-
-    MIL_Password = gr.Textbox(type='password', label="Enter the Password", visible=True)
-    log_in = gr.Button("Log In", visible=True)
-    log_in.click(fn=check_password, inputs=MIL_Password,
-                 outputs=[MIL_Password, log_in, password_text_row, content_row])
-
-    with password_text_row:
-        gr.Markdown("""<h3><center>Incorrect Password. Access Denied.</center></h3>""")
-
-    with content_row:
-        with gr.Tab('Resumes'):
-            gr.Markdown("""<h3><center>Resumes to Process:</center></h3>""")
-            with gr.Row():
-                with gr.Column():
-                    upload_resume = gr.UploadButton('Upload Resume', file_count="multiple")
-                with gr.Column():
-                    delete_resume = gr.Button('Delete')
-
-            with gr.Row():
-
-                @gr.render(triggers=[upload_resume.upload, delete_resume.click, log_in.click])
-                def file_exp():
-                    resume_file_explorer = gr.FileExplorer(
-                        root_dir=os.path.abspath("Resumes"),
-                        label='Resumes',
-                        interactive=True,
-                        elem_id='explorer',
-                    )
-
-                    upload_resume.upload(fn=add_resume, inputs=upload_resume)
-                    delete_resume.click(fn=remove_resumes, inputs=resume_file_explorer)
-        with gr.Tab('OpenAI'):
-            with gr.Row():
-                job_select = gr.Dropdown(label='Select Job', choices=job_names + ['Custom'], allow_custom_value=True)
-            with gr.Row(visible=False) as job_details:
-                with gr.Column():
-                    job_name_input = gr.Textbox(label='Job Name', lines=1, interactive=True)
-                    job_description_input = gr.Textbox(label='Job Description', lines=5, interactive=True)
-                    get_resumes = gr.Button('Find best resume')
-            with gr.Row(visible=True) as gpt_response:
-                best_resumes = gr.Markdown()
-
-        job_select.select(fn=job_selected, inputs=job_select,
-                          outputs=[job_name_input, job_description_input, job_details, best_resumes])
-        get_resumes.click(send_to_openai, inputs=[job_name_input, job_description_input],
-                          outputs=[best_resumes, gpt_response])
-
-
-if __name__ == '__main__':
-    # Remove all current resumes from the Resumes folder
-    files = glob.glob(os.path.abspath('Resumes') + "/*")
-    for f in files:
-        os.remove(f)
-
-    # Copy all the defaults resumes to the resume folder
-    files = glob.glob(os.path.abspath('Default Resumes') + "/*")
-    for f in files:
-        shutil.copy(f, os.path.abspath('Resumes'))
-
+import app_funtions as appfun
+import os
+import shutil
+import glob
+import gradio as gr
+from pathlib import Path
+from datetime import datetime
+
+with gr.Blocks(theme=gr.themes.Soft()) as demo:
+
+    gr.Markdown("""<h1><center>Resume Processing</center></h1>""")
+    password_row = gr.Row()
+    password_text_row = gr.Row(visible=False)
+    content_row = gr.Row(visible=False)
+    confirmation_row = gr.Row(visible=False)
+
+    MIL_Password = gr.Textbox(type='password', label="Enter the Password", visible=True)
+    log_in = gr.Button("Log In", visible=True)
+    log_in.click(fn=appfun.check_password, inputs=MIL_Password,
+                 outputs=[MIL_Password, log_in, password_text_row, content_row])
+
+    with password_text_row:
+        gr.Markdown("""<h3><center>Incorrect Password. Access Denied.</center></h3>""")
+
+    with content_row:
+        with gr.Tab('Resume Input'):
+            gr.Markdown("""<h3><center>Resumes Input</center></h3>""")
+            gr.Markdown("""<p><center>Upload or delete any resumes you would like to have the AI use:</center></p>""")
+            with gr.Row():
+                with gr.Column():
+                    upload_resume = gr.UploadButton('Upload Resume', file_count="multiple")
+                with gr.Column():
+                    delete_resume = gr.Button('Delete')
+
+            with gr.Row():
+
+                @gr.render(triggers=[upload_resume.upload, delete_resume.click, log_in.click])
+                def file_exp():
+                    resume_file_explorer = gr.FileExplorer(
+                        root_dir=os.path.abspath("Resumes"),
+                        label='Resumes',
+                        interactive=True,
+                        elem_id='explorer',
+                    )
+                    upload_resume.upload(fn=appfun.add_resume, inputs=upload_resume)
+                    delete_resume.click(fn=appfun.remove_resumes, inputs=resume_file_explorer)
+        with gr.Tab('Job Input'):
+            gr.Markdown("""<h3><center>Job Input</center></h3>""")
+            gr.Markdown("""<p><center>Upload a document you would like to extract job description from. Supported file types are: .pdf, 
+                        .docx, .pptx, .txt.</center></p>""")
+            upload_file = gr.File()
+            extract_button = gr.Button('Extract')
+            extract_completion_message = gr.Markdown("""<p><center></center></p>""")
+        with gr.Tab('Rank Resumes'):
+            gr.Markdown("""<h3><center>Rank Resumes</center></h3>""")
+            gr.Markdown("""<p><center>Choose the job posting of which you would like to get the top 3 resumes for.</center></p>""")
+            job_select = gr.Dropdown(label='Select Job', choices=appfun.job_names + ['Custom'], allow_custom_value=True)
+            with gr.Row(visible=False) as job_details:
+                with gr.Column():
+                    job_name_input = gr.Textbox(label='Job Name', lines=1, interactive=True)
+                    job_description_input = gr.Textbox(label='Job Description', lines=5, interactive=True)
+                    get_resumes = gr.Button('Rank Resumes')
+            with gr.Row(visible=True) as gpt_response:
+                best_resumes = gr.Markdown()
+        with gr.Tab("Chatbot"):
+            chatbot = gr.Chatbot(avatar_images=("user.jpeg", "gpt.jpg"), height=750)
+            state = gr.State()
+            chatbot_textbox = gr.Textbox(label="Input", info="", lines=1,
+                                         placeholder="Please process resumes", scale=1,
+                                         interactive=False)
+            chatbot_submit = gr.Button("SEND", interactive=False, scale=1)
+            chatbot_submit.click(appfun.my_chatbot, inputs=[chatbot_textbox, state],
+                                 outputs=[chatbot, state, chatbot_textbox])
+
+
+
+        job_select.select(fn=appfun.job_selected, inputs=job_select,
+                          outputs=[job_name_input, job_description_input, job_details, best_resumes])
+        get_resumes.click(appfun.send_to_openai, inputs=[job_name_input, job_description_input],
+                          outputs=[best_resumes, gpt_response, chatbot_textbox, chatbot_submit])
+        extract_button.click(fn=appfun.extract_jobs, inputs=upload_file, outputs=[extract_completion_message, job_select])
+
+
+if __name__ == '__main__':
+    # Remove all current resumes from the Resumes folder
+    files = glob.glob(os.path.abspath('Resumes') + "/*")
+    for f in files:
+        os.remove(f)
+
+    # Copy all the defaults resumes to the resume folder
+    files = glob.glob(os.path.abspath('Default Resumes') + "/*")
+    for f in files:
+        shutil.copy(f, os.path.abspath('Resumes'))
+
     demo.launch()

app_funtions.py

@@ -0,0 +1,150 @@
+import re
+
+from PwsResumeClassifier import PwsResumeClassifier as pws
+from openai import OpenAI
+import app_funtions as appfun
+import os
+import shutil
+import glob
+import gradio as gr
+import json
+from pathlib import Path
+
+os.environ['PROJECT'] = 'proj_ct77YbVBWyY3cNDoGA4Xkpzk'
+os.environ['OPENAIORG'] = 'org-Qc0wKzpN6ewg8NpSc9s7cTtS'
+os.environ['MILPASSWORD'] = 'MILRESUME'
+os.environ['APIKEY'] = 'sk-proj-24nTWywXfoK0f4fEd2jdT3BlbkFJLaAZTDMkNcY8HP20Asf2'
+
+job_names = []
+job_descriptions = []
+
+client_pws = pws(organization_id=os.getenv('OPENAIORG'), project_id=os.getenv('PROJECT'), api_key=os.getenv('APIKEY'))
+
+client = OpenAI(
+    organization=os.getenv('OPENAIORG'),
+    project=os.getenv('PROJECT'),
+    api_key=os.getenv('APIKEY')
+)
+
+def check_password(password):
+    if password == os.getenv('MILPASSWORD'):
+
+        return [gr.Textbox(visible=False), gr.Button(visible=False), gr.Row.update(visible=False), gr.Row.update(visible=True)]
+    else:
+        return [gr.Textbox(visible=True), gr.Button(visible=True), gr.Row.update(visible=True), gr.Row.update(visible=False)]
+
+
+def add_resume(resume_list):
+    for resume in resume_list:
+        shutil.copy2(os.path.abspath(resume), os.path.abspath('Resumes'))
+
+
+def remove_resumes(file_explorer):
+    for file in file_explorer:
+        os.remove(file)
+
+
+def job_selected(selected_name):
+    print(job_names)
+    if selected_name in job_names:
+        return [selected_name, job_descriptions[job_names.index(selected_name)], gr.Row.update(visible=True),
+                gr.Markdown()]
+    elif selected_name == 'Custom':
+        return ['', '', gr.Row.update(visible=True), gr.Markdown()]
+    else:
+        return ['', '', gr.Row.update(visible=False), gr.Markdown()]
+
+
+def send_to_openai(name, description):
+
+    try:
+        print('Sending to AI')
+        client_pws.set_vector_store(os.path.abspath('Resumes'))
+        resumes_output = client_pws.get_best_resumes('Job Name: ' + name + '. Job Description: ' + description)
+        response = gr.Markdown('<p>' + resumes_output + '</p>')
+        return [response, gr.Row.update(visible=True),
+                gr.Textbox(label="Input", info="", lines=1, placeholder="Ask the chatbot.", interactive=True, scale=1),
+                gr.Button("SEND", interactive=True, scale=1)]
+    except Exception as e:
+        response = gr.Markdown('<p>An error occurred. Please try again.</p>')
+        return [response, gr.Row.update(visible=True),
+                gr.Textbox(label="Input", info="", lines=1, placeholder="Please process resumes", interactive=False, scale=1),
+                gr.Button("SEND", interactive=False, scale=1)]
+
+
+def extract_jobs(filepath):
+    try:
+        description = r"""You are a document information extraction assistant that extract job names and job description 
+        precisely and accurately from a provided file."""
+
+        instructions = r"""You will be provided a document that contains information about one or more job positions. 
+        You will extract all of these job positions and return them in a JSON Array. Each item in the JSON Array will 
+        have a field "job_name" for the name of the job position and "job_description" for the description and 
+        requirements of the job position. Your response 
+        should only be a JSON Array with all items."""
+
+        file_streams = [open(filepath, "rb")]
+        vector_store = client.beta.vector_stores.create(
+            name="Job Description File",
+        )
+
+        file_batch = client.beta.vector_stores.file_batches.upload_and_poll(
+            vector_store_id=vector_store.id,
+            files=file_streams
+        )
+
+        assistant = client.beta.assistants.create(
+            name="Job Description Extract Assistant",
+            description=description,
+            instructions=instructions,
+            model="gpt-4o",
+            tools=[{"type": "file_search"}],
+            tool_resources={"file_search": {"vector_store_ids": [vector_store.id]}},
+            temperature=0
+        )
+        thread = client.beta.threads.create()
+
+        messages = client.beta.threads.messages.create(
+            thread_id=thread.id,
+            role="user",
+            content="Please extract all job descriptions"
+        )
+
+        run = client.beta.threads.runs.create_and_poll(
+            thread_id=thread.id,
+            assistant_id=assistant.id,
+        )
+
+        if run.status == 'completed':
+            messages = client.beta.threads.messages.list(
+                thread_id=thread.id
+            )
+
+        messages = client.beta.threads.messages.list(
+            thread_id=thread.id
+        )
+        print(messages.data[0].content[0].text.value)
+        json_array = messages.data[0].content[0].text.value
+        json_array = re.sub(r"^```json", "", json_array)
+        json_array = re.sub(r"```$", "", json_array)
+        json_array = re.sub(r"【.*?】", "", json_array)
+        print(json_array)
+        print(json.loads(json_array))
+
+        for job in json.loads(json_array):
+            job_names.append(job['job_name'])
+            job_descriptions.append(job['job_description'])
+        return [gr.Markdown("""<p><center>Extraction completed successfully</center></p>"""),
+                gr.Dropdown(label='Select Job', choices=job_names + ['Custom'], allow_custom_value=True)]
+    except Exception as e:
+        return [gr.Markdown("""<p><center>Parsing failed. Error message: """ + repr(e) + """</center></p>"""),
+                gr.Dropdown(label='Select Job', choices=appfun.job_names + ['Custom'], allow_custom_value=True)]
+
+
+def my_chatbot(user_input, history):
+    text = ""
+    history = history or []
+    output = client_pws.query_assistant(user_input)
+    history.append((user_input, output))
+
+    return history, history, text