Spaces:

BitBasher
/

EduConnect

Sleeping

App Files Files Community

dtyago commited on Feb 22

Commit

5b1d21d

•

1 Parent(s): 0d44817

Cleanup code for better UX

Browse files

Files changed (9) hide show

Dockerfile +6 -6
app/admin/templates/admin_login.html +1 -1
app/admin/templates/data_management.html +1 -1
app/admin/templates/registration_success.html +1 -0
app/admin/templates/user_registration.html +1 -1
app/dependencies.py +8 -2
app/main.py +1 -1
app/utils/chat_rag.py +84 -84
app/utils/db.py +0 -1

Dockerfile CHANGED Viewed

@@ -23,9 +23,6 @@ ENV HOME=/home/user \
 # Set the non-root user's home directory as the working directory
 WORKDIR $HOME
-# Create a symbolic link from /data to /home/user/data
-RUN ln -s /data /home/user/data
 # Change to the non-root user
 USER user
@@ -44,6 +41,12 @@ WORKDIR $HOME/app
 # Copy the rest of the application files into the container
 COPY --chown=user:user ./app .
 # Make port 7860 available to the world outside this container
 EXPOSE 7860
@@ -56,8 +59,5 @@ RUN chmod +x /home/user/entrypoint.sh
 # Update the ENTRYPOINT command to use the full path to entrypoint.sh
 ENTRYPOINT ["/home/user/entrypoint.sh"]
-# Set Python path just to make sure
-ENV PYTHONPATH="/home/user/app:${PYTHONPATH}"
 # Adjust the CMD to ensure it correctly references the FastAPI app
 CMD ["uvicorn", "main:app", "--host", "0.0.0.0", "--port", "7860"]

 # Set the non-root user's home directory as the working directory
 WORKDIR $HOME
 # Change to the non-root user
 USER user
 # Copy the rest of the application files into the container
 COPY --chown=user:user ./app .
+# Create a symbolic link from /data to /home/user/data
+RUN ln -s /data /home/user/data
+# Set Python path just to make sure
+ENV PYTHONPATH="/home/user/app:${PYTHONPATH}"
 # Make port 7860 available to the world outside this container
 EXPOSE 7860
 # Update the ENTRYPOINT command to use the full path to entrypoint.sh
 ENTRYPOINT ["/home/user/entrypoint.sh"]
 # Adjust the CMD to ensure it correctly references the FastAPI app
 CMD ["uvicorn", "main:app", "--host", "0.0.0.0", "--port", "7860"]

app/admin/templates/admin_login.html CHANGED Viewed

@@ -2,7 +2,7 @@
   <head>
     <title>EduConnect Administration - login page</title>
     <meta name="viewport" content="width=device-width, initial-scale=1.0">
-    <link href="static/css/mvp.css" rel="stylesheet" media="screen">
   </head>
   <body>
     <div class="container">

   <head>
     <title>EduConnect Administration - login page</title>
     <meta name="viewport" content="width=device-width, initial-scale=1.0">
+    <link href="/static/css/mvp.css" rel="stylesheet" media="screen">
   </head>
   <body>
     <div class="container">

app/admin/templates/data_management.html CHANGED Viewed

@@ -3,7 +3,7 @@
 <head>
     <title>Econnect Data Management</title>
     <meta name="viewport" content="width=device-width, initial-scale=1.0">
-    <link href="static/css/mvp.css" rel="stylesheet" media="screen">
     <!-- Bootstrap CSS -->
     <link rel="stylesheet" href="https://stackpath.bootstrapcdn.com/bootstrap/4.3.1/css/bootstrap.min.css">
 </head>

 <head>
     <title>Econnect Data Management</title>
     <meta name="viewport" content="width=device-width, initial-scale=1.0">
+    <link href="/static/css/mvp.css" rel="stylesheet" media="screen">
     <!-- Bootstrap CSS -->
     <link rel="stylesheet" href="https://stackpath.bootstrapcdn.com/bootstrap/4.3.1/css/bootstrap.min.css">
 </head>

app/admin/templates/registration_success.html CHANGED Viewed

@@ -2,6 +2,7 @@
 <html>
 <head>
     <title>Registration Success</title>
 </head>
 <body>
     <h1>Registration Successful!</h1>

 <html>
 <head>
     <title>Registration Success</title>
+    <link href="/static/css/mvp.css" rel="stylesheet" media="screen">
 </head>
 <body>
     <h1>Registration Successful!</h1>

app/admin/templates/user_registration.html CHANGED Viewed

@@ -3,7 +3,7 @@
 <head>
     <title>Econnect User Registration</title>
     <meta name="viewport" content="width=device-width, initial-scale=1.0">
-    <link href="static/css/mvp.css" rel="stylesheet" media="screen">
 </head>
 <body>
     <div class=""container">

 <head>
     <title>Econnect User Registration</title>
     <meta name="viewport" content="width=device-width, initial-scale=1.0">
+    <link href="/static/css/mvp.css" rel="stylesheet" media="screen">
 </head>
 <body>
     <div class=""container">

app/dependencies.py CHANGED Viewed

@@ -13,6 +13,12 @@ async def get_current_user(token: str = Depends(oauth2_scheme)):
         headers={"WWW-Authenticate": "Bearer"},
     )
     # Utilize the centralized JWT decoding and catch any JWT-related errors
     try:
         payload = decode_jwt(token)
@@ -23,9 +29,9 @@ async def get_current_user(token: str = Depends(oauth2_scheme)):
     if user_id is None:
         raise credentials_exception
-    # Verify if the token is stored and valid
     if not tinydb_helper.query_token(user_id, token):
-        raise credentials_exception
     # Payload is already obtained and validated, so just return it or its specific parts as needed
     return {"user_id": user_id, "name": payload.get("name"), "role": payload.get("role")}

         headers={"WWW-Authenticate": "Bearer"},
     )
+    expiry_exception = HTTPException(
+        status_code=status.HTTP_401_UNAUTHORIZED,
+        detail="No session found, please login again",
+        headers={"WWW-Authenticate": "Bearer"},
+    )
     # Utilize the centralized JWT decoding and catch any JWT-related errors
     try:
         payload = decode_jwt(token)
     if user_id is None:
         raise credentials_exception
+    # Verify if the token is stored and valid as active session
     if not tinydb_helper.query_token(user_id, token):
+        raise expiry_exception
     # Payload is already obtained and validated, so just return it or its specific parts as needed
     return {"user_id": user_id, "name": payload.get("name"), "role": payload.get("role")}

app/main.py CHANGED Viewed

@@ -49,7 +49,7 @@ async def startup_event():
     print(f"MODEL_PATH in main.py = {os.getenv('MODEL_PATH')} ")
 # Mount static files
-app.mount("/home/user/app/static", StaticFiles(directory="static"), name="static")
 # Setup Jinja2Templates to point to the templates directory
 templates = Jinja2Templates(directory="admin/templates")

     print(f"MODEL_PATH in main.py = {os.getenv('MODEL_PATH')} ")
 # Mount static files
+app.mount("/static", StaticFiles(directory="static"), name="static")
 # Setup Jinja2Templates to point to the templates directory
 templates = Jinja2Templates(directory="admin/templates")

app/utils/chat_rag.py CHANGED Viewed

@@ -126,90 +126,90 @@ async def load_llm():
-#step 5, to instantiate once to create default_chain,router_chain,destination_chains into chain and set vectordb. so will not re-create per prompt
-async def default_chain(llm, user_collection_name):
-    # Get Chromadb location
-    CHROMADB_LOC = os.getenv('CHROMADB_LOC')
-    vectordb = await get_vectordb_for_user_cached(user_collection_name)  # Use the dynamic vectordb based on user_id
-    sum_template = """
-    As a machine learning education specialist, our expertise is pivotal in deepening the comprehension of complex machine learning concepts for both educators and students.
-    our role entails:
-    Providing Detailed Explanations: Deliver comprehensive answers to these questions, elucidating the underlying technical principles.
-    Assisting in Exam Preparation: Support educators in formulating sophisticated exam and quiz questions, including MCQs, accompanied by thorough explanations.
-    Summarizing Course Material: Distill key information from course materials, articulating complex ideas within the context of advanced machine learning practices.
-    Objective: to summarize and explain the key points.
-    Here the question:
-    {input}"""
-    mcq_template = """
-    As a machine learning education specialist, our expertise is pivotal in deepening the comprehension of complex machine learning concepts for both educators and students.
-    our role entails:
-    Crafting Insightful Questions: Develop thought-provoking questions that explore the intricacies of machine learning topics.
-    Generating MCQs: Create MCQs for each machine learning topic, comprising a question, four choices (A-D), and the correct answer, along with a rationale explaining the answer.
-    Objective: to create multiple choice question in this format
-    [question:
-    options A:
-    options B:
-    options C:
-    options D:
-    correct_answer:
-    explanation:]
-    Here the question:
-    {input}"""
-    prompt_infos = [
-        {
-            "name": "SUMMARIZE",
-            "description": "Good for summarizing and explaination ",
-            "prompt_template": sum_template,
-        },
-        {
-            "name": "MCQ",
-            "description": "Good for creating multiple choices questions",
-            "prompt_template": mcq_template,
-        },
-    ]
-    destination_chains = {}
-    for p_info in prompt_infos:
-        name = p_info["name"]
-        prompt_template = p_info["prompt_template"]
-        #vectordb=p_info["vector"]
-        prompt = PromptTemplate(template=prompt_template, input_variables=["input"])
-        embeddings = HuggingFaceEmbeddings(model_name='sentence-transformers/all-MiniLM-L6-v2') #new
-        vectordb= Chroma(persist_directory = CHROMADB_LOC, embedding_function = embeddings) #new
-        retriever = vectordb.as_retriever()#new
-        memory = VectorStoreRetrieverMemory(retriever=retriever) #new
-        chain = LLMChain(llm=llm, prompt=prompt, verbose=True, memory=memory) #new memory=memory
-        destination_chains[name] = chain
-    #default_chain = ConversationChain(llm=llm, output_key="text")
-    #memory = ConversationBufferMemory(memory_key='chat_history', return_messages=True)
-    default_chain = ConversationalRetrievalChain.from_llm(llm=llm,
-                                                retriever=vectordb.as_retriever(search_kwargs={'k': 3}),
-                                                verbose=True, output_key="text" )
-    destinations = [f"{p['name']}: {p['description']}" for p in prompt_infos]
-    destinations_str = "\n".join(destinations)
-    router_template = MULTI_PROMPT_ROUTER_TEMPLATE.format(destinations=destinations_str)
-    router_prompt = PromptTemplate(
-        template=router_template,
-        input_variables=["input"],
-        output_parser=RouterOutputParser(),
-    )
-    router_chain = LLMRouterChain.from_llm(llm, router_prompt)
-    return default_chain,router_chain,destination_chains
 # Adjust llm_infer to accept user_id and use it for user-specific processing
 async def llm_infer(user_collection_name, prompt):

+# #step 5, to instantiate once to create default_chain,router_chain,destination_chains into chain and set vectordb. so will not re-create per prompt
+# async def default_chain(llm, user_collection_name):
+#     # Get Chromadb location
+#     CHROMADB_LOC = os.getenv('CHROMADB_LOC')
+#     vectordb = await get_vectordb_for_user_cached(user_collection_name)  # Use the dynamic vectordb based on user_id
+#     sum_template = """
+#     As a machine learning education specialist, our expertise is pivotal in deepening the comprehension of complex machine learning concepts for both educators and students.
+#     our role entails:
+#     Providing Detailed Explanations: Deliver comprehensive answers to these questions, elucidating the underlying technical principles.
+#     Assisting in Exam Preparation: Support educators in formulating sophisticated exam and quiz questions, including MCQs, accompanied by thorough explanations.
+#     Summarizing Course Material: Distill key information from course materials, articulating complex ideas within the context of advanced machine learning practices.
+#     Objective: to summarize and explain the key points.
+#     Here the question:
+#     {input}"""
+#     mcq_template = """
+#     As a machine learning education specialist, our expertise is pivotal in deepening the comprehension of complex machine learning concepts for both educators and students.
+#     our role entails:
+#     Crafting Insightful Questions: Develop thought-provoking questions that explore the intricacies of machine learning topics.
+#     Generating MCQs: Create MCQs for each machine learning topic, comprising a question, four choices (A-D), and the correct answer, along with a rationale explaining the answer.
+#     Objective: to create multiple choice question in this format
+#     [question:
+#     options A:
+#     options B:
+#     options C:
+#     options D:
+#     correct_answer:
+#     explanation:]
+#     Here the question:
+#     {input}"""
+#     prompt_infos = [
+#         {
+#             "name": "SUMMARIZE",
+#             "description": "Good for summarizing and explaination ",
+#             "prompt_template": sum_template,
+#         },
+#         {
+#             "name": "MCQ",
+#             "description": "Good for creating multiple choices questions",
+#             "prompt_template": mcq_template,
+#         },
+#     ]
+#     destination_chains = {}
+#     for p_info in prompt_infos:
+#         name = p_info["name"]
+#         prompt_template = p_info["prompt_template"]
+#         #vectordb=p_info["vector"]
+#         prompt = PromptTemplate(template=prompt_template, input_variables=["input"])
+#         embeddings = HuggingFaceEmbeddings(model_name='sentence-transformers/all-MiniLM-L6-v2') #new
+#         vectordb= await get_vectordb_for_user_cached(user_collection_name)
+#         retriever = vectordb.as_retriever()#new
+#         memory = VectorStoreRetrieverMemory(retriever=retriever) #new
+#         chain = LLMChain(llm=llm, prompt=prompt, verbose=True, memory=memory) #new memory=memory
+#         destination_chains[name] = chain
+#     #default_chain = ConversationChain(llm=llm, output_key="text")
+#     #memory = ConversationBufferMemory(memory_key='chat_history', return_messages=True)
+#     default_chain = ConversationalRetrievalChain.from_llm(llm=llm,
+#                                                 retriever=vectordb.as_retriever(search_kwargs={'k': 3}),
+#                                                 verbose=True, output_key="text" )
+#     destinations = [f"{p['name']}: {p['description']}" for p in prompt_infos]
+#     destinations_str = "\n".join(destinations)
+#     router_template = MULTI_PROMPT_ROUTER_TEMPLATE.format(destinations=destinations_str)
+#     router_prompt = PromptTemplate(
+#         template=router_template,
+#         input_variables=["input"],
+#         output_parser=RouterOutputParser(),
+#     )
+#     router_chain = LLMRouterChain.from_llm(llm, router_prompt)
+#     return default_chain,router_chain,destination_chains
 # Adjust llm_infer to accept user_id and use it for user-specific processing
 async def llm_infer(user_collection_name, prompt):

app/utils/db.py CHANGED Viewed

@@ -1,4 +1,3 @@
 import os
 from tinydb import TinyDB, Query, where
 from tinydb.storages import MemoryStorage

 import os
 from tinydb import TinyDB, Query, where
 from tinydb.storages import MemoryStorage