notes.txt

# Creating Requirements text file
- Use pip freeze => List of all dependencies installed
- Use grep => to search for the one, wish to save in the req.txt file
- Ex. : pip freeze | grep streamlit

# Using pipreqs command to generate Requirements.txt
- It is bit finicky as it will update the non-breaking dependencies automatically
- This command has tendency to 

# Conda is used for Virtual Enviroment python
- Venv is more lightweight vs Conda
- Conda is package manager, allows more flexibility
- conda activate chat-with-website
___________________________________________________

# How Streamlit works ?
- Any Event or interaction happens on the Streamlit => Streamlit runs the full code again
- Chat_history in the above code => Needs to be in persistent state, shouldnt reintialize when streamlit runs everything again
- You can achieve this in streamlit, using streamlit.session_state

# Note 1:
- In this code, we are not using just the user query to do the semantic search on the chunk embeddings of the documnet
- but instead we are using both user query + the entire conversation history and then do a semantic search on the chunk embeddings of the document

# Note 2:
- llm could be part of either Langchain community or openai

# Note 3:
- In Langchain, every vector store have a method called 'as_retriever' --> same goes for Chroma vector store too !
- Allows to use the vector store to be used it to retrieve relevant text based on the query

# Note 4:
- In Langchain -->  chain invoke requires variables to be passed on for the internal prompt as key-value pairs

# In Production Grade App:
- We wont expose the chains to the frontend, and put them all in the backend

# Note 5:
- This application, adds wayy to much convo context, if you kept asking new questions, which might be an issue

# Note 6:
- One Issue, I found was running into was:
- this model's maximum context length is 4097 tokens. However, your messages resulted in 4561 tokens. Please reduce the length of the messages
___________________________________________________
# Comments from the Code:
# Line: 60
Messages Placeholder replaces itself with the variable it is passed with (If the var exists, else remains empty)
here in our case, it is chat_history
Langchain allows to send in messages as tuple
prompt works exactly as a f string in the python, so the f string gets populated with the value it has been fed with
This prompt will be used to ask llm to generate a search query for contextual data chunk

# Line: 139
Flag to check if var already exists in session state
Yes -> no need to intialize

# Line:
___________________________________________________

# Importing Rules for Py File:
- Group imports by category:
Group imports into sections based on their source (standard library, third-party libraries, local imports).
Use blank lines to separate these sections for clarity.

-Use absolute imports:
Prefer absolute imports over relative imports to ensure clarity and prevent potential issues with module resolution.

-Import one module per line:
Import each module on its own line to improve readability and maintainability.

-Sort imports alphabetically:
Sort imports alphabetically within each section to make them easier to find.

# In this version:

- Each function has a docstring that explains its purpose, parameters, and return value.
- Comments are used sparingly and only for clarifying complex or non-obvious code sections.
- Variable and function names are descriptive, following PEP8 naming conventions.
- The code structure is clear, with consistent indentation and spacing.



___________________________________________________
# Linting Error on Outer Scope
- The "redefining name from outer scope" linting error typically occurs when a variable or function with the same name is declared in an inner scope, 
which shadows a variable or function with the same name in the outer scope. This can lead to confusion and unintended behavior, especially in languages 
where scoping rules are strict, like Python.

# Linting Error:
- The Pylint error "expected 2 blank lines, found 1" indicates that Pylint expects to see two blank lines between function definitions or classes.

# Linting Error:
- The Pylint error "Missing function or method docstring" indicates that a function or method is missing a docstring. Docstrings are important for 
documenting the purpose, arguments, and return values of functions or methods, which helps improve code readability and maintainability.

# Linting Error for naming Constants:
The linting error "Constant name 'message_type' doesn't conform to UPPER_CASE naming style" suggests that the variable message_type is being treated as a constant, 
but its naming convention doesn't match the convention typically used for constants, which is usually UPPER_CASE_WITH_UNDERSCORES.