Merge branch 'switch_jupyterlab_and_app' into 'main'

Switch jupyterlab and app

See merge request !7

.gitignore

@@ -5,3 +5,8 @@
 *.pyc
 *.txt
 config.txt
+
+run.sh
+build.sh
+run_my_app.sh
+run_streamlit_app.sh
\ No newline at end of file

Dockerfile

-FROM jupyter/datascience-notebook:hub-3.1.1
+FROM quay.io/jupyter/datascience-notebook:hub-5.2.0
 
 USER root
 
-COPY requirements.txt environment.yml /tmp/
-RUN conda env update -q -f /tmp/environment.yml && \
-    /opt/conda/bin/pip install -r /tmp/requirements.txt && \
-    conda clean -y --all && \
-    conda env export -n "root" && \
-    jupyter lab build 
+COPY requirements.txt .
+RUN pip install --no-cache-dir -r requirements.txt && rm requirements.txt
 
-COPY dash_proxy /tmp/dash_proxy/
-RUN pip install /tmp/dash_proxy/
+ENV HOME_PATH=/home/jovyan/
 
+# add the proxy for streamlit
+COPY streamlit_proxy /tmp/streamlit_proxy/
+RUN pip install /tmp/streamlit_proxy/
+
+# install some utilities for GPT
 COPY llm_utils /llm_utils/
 RUN pip install /llm_utils/
 ENV CONFIG_PATH=/home/jovyan/config.txt
 
-COPY app /dash/app/
-RUN chown -R jovyan /dash/app/
+# copy the apps into the container
+COPY app /streamlit/app/
+RUN chown -R jovyan /streamlit/app/
 
 # install some NLTK and spaCy data
 RUN python -m nltk.downloader stopwords
 RUN python -m nltk.downloader wordnet
 RUN python -m nltk.downloader punkt
-RUN python -m spacy download en_core_web_sm
+
+USER jovyan
\ No newline at end of file

app/app.py

-import sys 
-sys.path.append("/home/jovyan/")
+import os
+import streamlit as st
 
-import argparse
-import logging
+from llm_utils.client import get_openai_client
 
-from urllib.parse import urlparse, urljoin
 
-from dash import Dash
-
-from jupyter_server.serverapp import list_running_servers
-
-from layout import layout
-from callbacks import register_callbacks
-
-logging.basicConfig(level=logging.INFO)
-
-# weird trick to find base_url for the jupyterlab
-def find_jupyterlab_base_url():
-    servers = list_running_servers()
-    for server in servers:
-        if server["port"] == 8888:
-            return server['url']
-    return None
-
-
-# get the correct port from proxy
-parser = argparse.ArgumentParser()
-parser.add_argument("--port", type=int)
-args = parser.parse_args()
-port: int = args.port
+MODEL = "gpt-4o"
+client = get_openai_client(
+    model=MODEL,
+    config_path=os.environ.get("CONFIG_PATH")
+)
 
-if not port:
-    raise ValueError(f"Port of proxy server for Dash not found in {args}.")
-else: 
-    logging.debug(f"Dash app running on port {port}.")
 
 
-base_url = find_jupyterlab_base_url()
-if base_url is None:
-    raise ValueError("Base URL of Jupyterlab could not be detected.")
-logging.debug(f"Base URL: {base_url}")
+# STREAMLIT APP
 
-proxy_base_path = urlparse(urljoin(base_url + "/", f"proxy/{port}/")).path
-logging.debug(f"Proxy base path: {proxy_base_path}")
+st.title("ChatGPT in Streamlit")
 
-# define Dash app
-app = Dash(
-    name=__name__, 
-    requests_pathname_prefix=proxy_base_path
+client = get_openai_client(
+    model=MODEL,
+    config_path=os.environ.get("CONFIG_PATH")
 )
 
-# define layout
-app.layout = layout
-
-# register all callback functions
-register_callbacks(app=app)
-
-# Run Dash app in the notebook
-app.run(
-    jupyter_mode="jupyterlab", 
-    port=port, 
-    host="0.0.0.0",
-    debug=True
-)
+if "openai_model" not in st.session_state:
+    st.session_state["openai_model"] = "gpt-4o"
+
+if "messages" not in st.session_state:
+    st.session_state.messages = []
+
+for message in st.session_state.messages:
+    with st.chat_message(message["role"]):
+        st.markdown(message["content"])
+
+if prompt := st.chat_input("What is up?"):
+    st.session_state.messages.append({"role": "user", "content": prompt})
+    with st.chat_message("user"):
+        st.markdown(prompt)
+
+    with st.chat_message("assistant"):
+        stream = client.chat.completions.create(
+            model=st.session_state["openai_model"],
+            messages=[
+                {"role": m["role"], "content": m["content"]}
+                for m in st.session_state.messages
+            ],
+            stream=True,
+        )
+        response = st.write_stream(stream)
+    st.session_state.messages.append({"role": "assistant", "content": response})
\ No newline at end of file

app/callbacks.py

-import os
-from datetime import datetime
-
-from dash import (
-    html, 
-    Dash
-)
-from dash.dependencies import (
-    Input, 
-    Output, 
-    State
-)
-
-from llm_utils.client import ChatGPT, get_openai_client
-
-
-def format_chat_messages(chat_history):
-    chat_messages = []
-    for message in chat_history:
-        chat_messages.append(html.Div([
-            html.P(f'{message["sender"]}: {message["message"]}'),
-            html.P(f'Sent at: {message["timestamp"]}')
-        ]))
-    return chat_messages
-
-
-def register_callbacks(app: Dash):
-    model="gpt4"
-    client = get_openai_client(
-        model=model,
-        config_path=os.environ.get("CONFIG_PATH")
-    )
-    chat_gpt = ChatGPT(
-        client=client,
-        model="gpt4"
-    )
-    
-    @app.callback(
-        [Output('chat-container', 'children'),
-        Output('chat-history', 'data')],
-        [Input('send-button', 'n_clicks')],
-        [State('user-input', 'value'),
-        State('chat-history', 'data')]
-    )
-    def update_chat(n_clicks, input_value, chat_history):
-        if chat_history is None:
-            chat_history = []
-        
-        if n_clicks > 0 and input_value:
-            chat_history.append({
-                'sender': 'User',
-                'message': input_value,
-                'timestamp': datetime.now().strftime("%Y-%m-%d %H:%M:%S")
-            })
-            
-            response = chat_gpt.chat_with_gpt(input_value)
-            
-            # Add response to chat history
-            chat_history.append({
-                'sender': 'Language Model',
-                'message': response,
-                'timestamp': datetime.now().strftime("%Y-%m-%d %H:%M:%S")
-            })
-        
-        return format_chat_messages(chat_history), chat_history

app/layout.py

-from dash import (
-    html, 
-    dcc
-)
-
-layout = html.Div([
-    dcc.Store(
-        id='chat-history', 
-        data=[]
-    ),
-    html.H1(
-        "Simple Chat App", 
-        style={'text-align': 'center'}
-    ),
-    html.Div(
-        id='chat-container', 
-        style={'overflowY': 'scroll', 'height': '70vh', 'padding': '10px'}
-    ),
-    html.Div([
-        dcc.Input(
-            id='user-input', 
-            type='text', 
-            placeholder='Type your message...', 
-            debounce=True
-        ),
-        html.Button(
-            'Send', 
-            id='send-button', 
-            n_clicks=0
-        )
-    ], style={
-        'display': 'flex', 
-        'alignItems': 'center', 
-        'justifyContent': 'center', 
-        'position': 'fixed', 
-        'bottom': 0, 
-        'width': '100%', 
-        'padding': '10px'
-    })
-], style={'position': 'relative'})

app/my_app.py

+import os
 import sys 
-sys.path.append("/home/jovyan/")
 
-import argparse
-import logging
+HOME_PATH = os.environ.get("HOME_PATH")
+sys.path.append(HOME_PATH)
 
-from urllib.parse import urlparse, urljoin
-
-from dash import Dash
-
-from jupyter_server.serverapp import list_running_servers
+print(HOME_PATH)
 
 try: 
-    from my_layout import layout
-    from my_callbacks import register_callbacks
-except ModuleNotFoundError:
-    # do not let Dash start
-    exit()
-
-
-logging.basicConfig(level=logging.INFO)
-
-# weird trick to find base_url for the jupyterlab
-def find_jupyterlab_base_url():
-    servers = list_running_servers()
-    for server in servers:
-        if server["port"] == 8888:
-            return server['url']
-    return None
-
-
-# get the correct port from proxy
-parser = argparse.ArgumentParser()
-parser.add_argument("--port", type=int)
-args = parser.parse_args()
-port: int = args.port
+    import streamlit_app # this runs the app
 
-if not port:
-    raise ValueError(f"Port of proxy server for Dash not found in {args}.")
-else: 
-    logging.debug(f"Dash app running on port {port}.")
-
-
-base_url = find_jupyterlab_base_url()
-if base_url is None:
-    raise ValueError("Base URL of Jupyterlab could not be detected.")
-logging.debug(f"Base URL: {base_url}")
-
-proxy_base_path = urlparse(urljoin(base_url + "/", f"proxy/{port}/")).path
-logging.debug(f"Proxy base path: {proxy_base_path}")
-
-# define Dash app
-app = Dash(
-    name=__name__, 
-    requests_pathname_prefix=proxy_base_path
-)
-
-# define layout
-app.layout = layout
-
-# register all callback functions
-register_callbacks(app=app)
-
-# Run Dash app in the notebook
-app.run(
-    jupyter_mode="jupyterlab", 
-    port=port, 
-    host="0.0.0.0",
-    debug=True
-)
+except ModuleNotFoundError as e:
+    # do not let start
+    print(e)
+    exit()

app/streamlit-favicon.svg

+<svg width="512" height="512" viewBox="0 0 512 512" fill="none" xmlns="http://www.w3.org/2000/svg">
+<path d="M255.968 288.494L166.211 241.067L10.4062 158.753C10.2639 158.611 9.97951 158.611 9.83728 158.611C4.14838 155.909 -1.68275 161.596 0.450591 167.283L79.8393 369.685L79.8535 369.728C79.9388 369.927 80.0099 370.126 80.0953 370.325C83.3522 377.874 90.4633 382.537 98.2002 384.371C98.8544 384.513 99.3225 384.643 100.108 384.799C100.89 384.973 101.983 385.21 102.922 385.281C103.078 385.295 103.221 385.295 103.377 385.31H103.491C103.605 385.324 103.718 385.324 103.832 385.338H103.989C104.088 385.352 104.202 385.352 104.302 385.352H104.486C104.6 385.366 104.714 385.366 104.828 385.366L167.175 392.161C226.276 398.602 285.901 398.602 345.002 392.161L407.35 385.366C408.558 385.366 409.739 385.31 410.877 385.196C411.246 385.153 411.602 385.111 411.958 385.068C412 385.054 412.057 385.054 412.1 385.039C412.342 385.011 412.583 384.968 412.825 384.926C413.181 384.883 413.536 384.812 413.892 384.741C414.603 384.585 414.926 384.471 415.891 384.139C416.856 383.808 418.458 383.228 419.461 382.745C420.464 382.261 421.159 381.798 421.999 381.272C423.037 380.618 424.024 379.948 425.025 379.198C425.457 378.868 425.753 378.656 426.066 378.358L425.895 378.258L255.968 288.494Z" fill="#FF2B2B"/>
+<path d="M501.789 158.755H501.647L345.784 241.07L432.426 370.058L511.616 167.285V167.001C513.607 161.03 507.492 155.627 501.789 158.755" fill="#7D353B"/>
+<path d="M264.274 119.615C260.292 113.8 251.616 113.8 247.776 119.615L166.211 241.068L255.968 288.495L426.067 378.357C427.135 377.312 427.991 376.293 428.897 375.217C430.177 373.638 431.372 371.947 432.424 370.056L345.782 241.068L264.274 119.615Z" fill="#BD4043"/>
+</svg>

dash_proxy/dash_proxy.py

-def setup_dash_proxy():
-    command = [
-        'python',
-        '/dash/app/app.py',
-        '--port',
-        '{port}'
-    ]
-    
-    return {
-        "command": command,
-        "new_browser_tab": False,
-        "launcher_entry": {
-            "enabled": True,
-            'title': 'Dash'
-        }
-    }

environment.yml

-name: "base"
-channels:
-  - defaults
-# dependencies:
-# - add packages here
-# - one per line
-prefix: "/opt/conda"

llm_utils/src/llm_utils/client.py

@@ -8,6 +8,7 @@ from enum import Enum
 class OpenAIModels(Enum):
     GPT_3 = "gpt3"
     GPT_4 = "gpt4"
+    GPT_4o = "gpt-4o"
     EMBED = "embed"
 
     @classmethod
@@ -43,7 +44,7 @@ def get_openai_client(
     client = AzureOpenAI(
         api_key=AZURE_OPENAI_API_KEY,
         azure_endpoint=AZURE_OPENAI_ENDPOINT,
-        api_version=OPENAI_API_VERSION, 
+        # api_version=OPENAI_API_VERSION, 
         azure_deployment=model
     )
     return client

my_callbacks.py

-from dash.dependencies import (
-    Input, 
-    Output
-)
-from dash import html
-
-
-def register_callbacks(app):
-    @app.callback(
-        Output('output-container-button', 'children'),
-        [Input('submit-btn', 'n_clicks')],
-        [Input('input-text', 'value')]
-    )
-    def update_output(n_clicks, input_value):
-        if n_clicks > 0:
-            return html.Div([
-                html.Label("You entered:"),
-                html.P(input_value)
-            ])
-        else:
-            return ''

my_layout.py

-from dash import html
-from dash import dcc
-
-
-layout = html.Div([
-    html.H1("Yeay, my app!"),
-    html.Div([
-        html.Label("Enter your text:"),
-        dcc.Input(id='input-text', type='text', value=''),
-        html.Button('Submit', id='submit-btn', n_clicks=0),
-    ]),
-    html.Div(id='output-container-button')
-])

python_startup/01_load_db.py

-import os
-from dotenv import load_dotenv
-import sqlalchemy
-import pandas as pd
-from urllib.parse import quote_plus
-
-load_dotenv("/usr/src/app/.env")
-
-class Database:
-    def __init__(self, db_name):
-        self.conn = db_connect(db_name)
-
-    def __getattr__(self, table_name):
-        return pd.read_sql_table(table_name, self.conn)
-
-    def list_tables(self):
-        inspector = sqlalchemy.inspect(self.conn)
-        table_names = inspector.get_table_names()
-        return table_names
-
-def db_connect(db_name):
-    hostname=os.getenv("DB_HOST")
-    user=os.getenv("DB_USER")
-    password=quote_plus(os.getenv("DB_PASSWORD"))
-    conn = sqlalchemy.create_engine(f'postgresql+psycopg2://{user}:{password}@{hostname}/{db_name}')
-    return conn
-
-def get_table(db_name, table_name):
-    conn = db_connect(db_name)
-    dat = pd.read_sql_table(table_name, conn)
-    return dat
-
-def get_all_tables(db_name):
-    db_obj = Database(db_name)
-    return db_obj
-

requirements.txt

-jupyter-server-proxy==4.0.0
-jupyterlab-git==0.42.0
-jupyter_server>=2.0
-
-flake8
-
-dash
-dash-bootstrap-components
-plotly
+jupyter-server-proxy==4.4.0
 
+streamlit
 openai
 rapidfuzz
 nltk
-spacy==3.4.4
-numpy==1.21.6
-scikit-learn==1.1.3
 
 pdfplumber
 python-dotenv

streamlit_app.py

+import streamlit as st 
+
+
+st.title("This is my app!")

dash_proxy/my_app_proxy.py → streamlit_proxy/my_app_proxy.py

 def setup_my_app_proxy():
     command = [
-        'python',
-        '/dash/app/my_app.py',
-        '--port',
-        '{port}'
+        "/streamlit/app/run_my_app.sh",
+        "--browser.gatherUsageStats", "false",
+        "--browser.serverAddress", "0.0.0.0",
+        "--server.port", "{port}",
+        "--server.headless", "true",
+        "--server.enableCORS", "false",
+        "--server.enableXsrfProtection", "false",
     ]
-    
+
     return {
         "command": command,
+        "timeout": 20,
         "new_browser_tab": False,
         "launcher_entry": {
             "enabled": True,
-            'title': 'MyApp'
+            "title": "MyApp",
+            "icon_path": "/streamlit/app/streamlit-favicon.svg"
         }
-    }
+    }
\ No newline at end of file

dash_proxy/setup.py → streamlit_proxy/setup.py

@@ -3,15 +3,14 @@ import setuptools
 setuptools.setup(
     author="Julian Rasch",
     author_email="julian.rasch@fh-muenster.de",
-    description="A small module to run Dash inside a dockerized Jupyterlab.",
-    name="jupyter-dash-proxy",
-    py_modules=["dash_proxy", "my_app_proxy"],
+    description="A small module to run Streamlit inside a dockerized JupyterLab.",
+    name="jupyter-streamlit-proxy",
+    py_modules=["streamlit_proxy", "my_app_proxy"],
     entry_points={
         "jupyter_serverproxy_servers": [
-            # name = packagename:function_name
-            "Dash = dash_proxy:setup_dash_proxy",
+            "Streamlit = streamlit_proxy:setup_streamlit_proxy",
             "MyApp = my_app_proxy:setup_my_app_proxy"
         ]
     },
     install_requires=["jupyter-server-proxy==4.0.0"],
-)
+)
\ No newline at end of file

streamlit_proxy/streamlit_proxy.py

+import os
+
+
+def setup_streamlit_proxy():
+    command = [
+        "/streamlit/app/run_streamlit_app.sh",
+        "--browser.gatherUsageStats", "false",
+        "--browser.serverAddress", "0.0.0.0",
+        "--server.port", "{port}",
+        "--server.headless", "true",
+        "--server.enableCORS", "false",
+        "--server.enableXsrfProtection", "false",
+    ]
+
+
+    return {
+        "command": command,
+        "timeout": 20,
+        "new_browser_tab": False,
+        "launcher_entry": {
+            "enabled": True,
+            "title": "Streamlit App",
+            "icon_path": "/streamlit/app/streamlit-favicon.svg"
+        }
+    }