ai code snippets 2

android Android

web Web

27-Jan-24 (Created: 27-Jan-24) | More in 'Generative Ai'

ai code snippets 2

satya - 1/24/2024, 10:43:53 PM

How do I load spacy stuff


pip install spacy
python -m spacy download en_core_web_sm

# Then

import spacy
# Load the installed model
nlp = spacy.load("en_core_web_sm")

# Now you can process text using the model
doc = nlp("This is a sentence.")

satya - 1/24/2024, 10:46:49 PM

Vscode frustration

venv sucks as pylance is not recognizing new modules although program works
Remved venv and pycache
Created a new work space with no venv
For some reason vscode insists that I have a venv and disables the entire work space from running a command line terminal
Started all over and have to explicitly "create" a work space instead of a folder to get the "basic" python work.
No venv at this point
Until I figure out pylance this is quite useless

satya - 1/25/2024, 8:39:57 AM

Workspaces, venv, and folders

Each folder is common to have its own .venv
Each folder corresponds to each project
Vscode should allow where to place the .venv
Shared virtual environments are possible but within the workspace
within the workspace means, any root folder in that work space
It is a common practice to have the workspace file outside of the root folders of that workspace

satya - 1/25/2024, 8:42:11 AM

How to debug the Pylance mess


Basic settings
*******************
"python.analysis.typeCheckingMode": "basic" 

"python.analysis.diagnosticSeverityOverrides": {
    "reportMissingImports": "error",
    "reportMissingModuleSource": "error",
    // other overrides
}

python: language server
*************************
Use "default" [or pylance]

In settings .json
*************************
{
  "python.analysis.extraPaths": [
    "/path/to/extra/library1",
    "/path/to/extra/library2"
  ]
}

More on the paths
***************************
{
  "python.analysis.extraPaths": [
    "./external_libs/lib1",
    "/absolute/path/to/lib2"
  ]
}

The "." is a root of the work space
Likely the ./.venv/libs/lib1
root folder of the project

Project/Folder level settings
**************************
.vscode/settings.json

satya - 1/25/2024, 4:47:30 PM

Streamlit conversational apps

satya - 1/25/2024, 4:48:01 PM

streamlit and HF

satya - 1/25/2024, 4:48:30 PM

Github Sample code: Biju

satya - 1/25/2024, 4:48:53 PM

Github Faisal

satya - 1/25/2024, 5:53:39 PM

Here is what happens with Streamlit

Just a python lib in your command line apps
pip install it, just another lib (of course like all other python stuff the dependencies galore)
You are ready to go
If you run streamlit on a command line it creates a webapp on your local box
It uses the python file you specify as its home page
That .py file has write statements in it like hello world
that is your first webapp!

satya - 1/25/2024, 5:55:26 PM

How long it took me: 1 hour for a basic page

Went to their web page
Read getting started
wrote what I want in English
Gave it to chatgpt
I copied the code
A web form with its output back to the page is the requirements
About an hour

satya - 1/25/2024, 5:57:03 PM

Here is English instructions


Filename: streamlit-ui-requirements.txt
A file to guide ChatGPT based development

**********************************
Requirement 1: for ChatGPT: Create a webpage with a from submission
**********************************
What I want:

1. Create a function
2. it should prompt the user on command line with a prompt
3. Gather the prompt from user
4. If the prompt is empty exit the program
5. if the prompt is "q" or "quit" exactly exit the program
6. call a function to process the prompt. Call this function "processPrompt"
7. Put all this in a while loop
8. Call the function "executePromptLoop"

Assume:
1. To write code modularly

satya - 1/25/2024, 5:59:44 PM

Here is the code for that web page


import streamlit as st

#
#**************************************************
# Key elements
# 1. A form with a submit
# 2. Process text input
# 3. Write back what is entered so far
# 4. Rememebers with each web refresh the state
#**************************************************
#
def writeIntro():
    # Streamlit page layout
    st.title("Demo page")

    st.write("hello there")
    st.write("hello there")
    st.write("hello there")


def initialize_state():
    writeIntro()
    
    """Initialize the state variable."""
    if 'totalResponseText' not in st.session_state:
        st.session_state.totalResponseText = []

def writeOutput():
    # Display the accumulated text
    st.write("Accumulated Text:")
    for line in st.session_state.totalResponseText:
        st.write(line)


def process_input(input_text):
    """Append the input text to the totalResponseText state variable."""
    st.session_state.totalResponseText.append(input_text)

def main():
    """Main function to render the Streamlit page."""
    initialize_state()

    # Create a form
    with st.form(key='input_form'):
        text_input = st.text_input("Enter your text")
        submit_button = st.form_submit_button("Submit")

    # Process the form submission
    if submit_button and text_input:
        process_input(text_input)

    writeOutput()

#Kick it off
main()

satya - 1/27/2024, 11:43:11 AM

nltk docs

satya - 1/27/2024, 11:45:20 AM

nltk corpus api

satya - 1/27/2024, 11:47:05 AM

where is nltk shakespeare corpus documented?

Search for: where is nltk shakespeare corpus documented?

satya - 1/27/2024, 11:52:58 AM

Closest I can find is here

satya - 1/27/2024, 11:54:28 AM

wiki documentation on github

satya - 1/27/2024, 11:54:36 AM