Include main classes in canopy's main __init__

[igiloh-pinecone]

Problem

The import path for some of our classes was too convoluted, making our examples complicated.

Solution

1. I included all the major classes + data classes in canopy's main __init__.py
2. Changed the README, example notebook, and docstring examples accordingly

Type of Change

• Breaking change (fix or feature that would cause existing functionality to not work as expected)
• This change requires a documentation update

Test Plan

None

[review-notebook-app]

Check out this pull request on 

See visual diffs & provide feedback on Jupyter Notebooks.

---

Powered by ReviewNB

[acatav]

I'm not sure we want to import so many things in Canopy highest level because it means that you can't take small parts of it without importing many many stuff (that right now takes more than 4 seconds to import). I'm leaving it to you decision but just noting that putting it in highest level is one way door. Adding some of it later is always possible so you can consider putting non data models stuff outside the canopy init file

[acatav]

Suggested change

	z# Canopy Library
	# Canopy Library

[acatav]

I'm not sure if we want this things to be imported in highest level. It means that even if I just want the chunker from Canopy, I'm going to import many other stuff. Is there anyway around that?

[igiloh-pinecone]

1. I still think we should implement the Chunkers themselves in pinecone-text, and wrap in canopy. You need to distinct between the chunker logic, which operates on simple data types (List[str]) and the Chunker class, which is a sub-component of KnowledgeBase.
   As I see it, people should use canopy if they want the functionality of KnowledgeBase or ContextEngine. If someone just needs a Chunker \ Ranker \ Encoder etc - they should take it directly from pinecone-text. Even if we'll decide to merge them into a mono repo - they would still be two separate packages, with their own dependencies etc.
2. Regardless of all of the above, the import itself should be super light and dependency free. Even if someone wants to use only the KnowledgeBase without any LLM or ChatEngine - importing the ChatEngine should be quick and hassle free. Any special dependencies should happen on __init__(), not on import. So it should be ok to import everything at the root level.

[acatav]

I agree in long term. Right now for both points this is not what we have and we can't tell if ever we will have either. So I wonder if we should not do that until both points or at least the second are true

[miararoy]

I am unable to do from canopy import KnowledgeBase for some reason

[igiloh-pinecone]

Closing for now, will come back to this in the future