problem getting embeddings to store in chroma_db

flat_vs_chroma.py

# llm is a valid OpenAI|Ollama object
# Settings.embed_model is a valid "nomic-embed-text" embedding model producing 768-d vectors

# This code works and 'default_vector_store.json' contains the embeddings 
documents = SimpleDirectoryReader(
	"data/DIR_WORD_DOCs",
	filename_as_id=False,
	file_extractor={'.docx':DocxReader()}
).load_data(num_workers=num_workers)

default_pt_paragraph_seperator = '\n\n'
text_splitter = SentenceSplitter(
	chunk_size=512,
	chunk_overlap=20,
	paragraph_separator=default_pt_paragraph_seperator
)
Settings.text_splitter = text_splitter

title_extractor  = TitleExtractor( nodes=5,llm=llm, num_workers=num_workers )
qa_extractor     = QuestionsAnsweredExtractor( questions=3, llm=llm, num_workers=num_workers,)

vector_index = VectorStoreIndex.from_documents(
	documents,
	transformations=[text_splitter, qa_extractor, title_extractor],
	embed_model = Settings.embed_model,
)
vector_index.storage_context.persist(persist_dir='data/RAG_Storage_2')

# If I run querys against this I get top_k nodes that have rational similarity scores like 0.7023 or similar


# The chromadb example below DOES NOT work. 

db = chromadb.PersistentClient(path="data/chroma_db")
chroma_collection = db.get_or_create_collection("my_collection")
vector_store    = ChromaVectorStore(chroma_collection=chroma_collection)
storage_context = StorageContext.from_defaults(vector_store=vector_store)

documents = SimpleDirectoryReader(
	"../data/DIR_WORD_DOCs",
	filename_as_id=False,
	file_extractor={'.docx':DocxReader()}
).load_data(num_workers=num_workers)

default_pt_paragraph_seperator = '\n\n'
text_splitter = SentenceSplitter(
	chunk_size=512, chunk_overlap=20,
	paragraph_separator=default_pt_paragraph_seperator
)
Settings.text_splitter = text_splitter

title_extractor  = TitleExtractor(nodes=5,llm=llm, num_workers=num_workers )
qa_extractor = QuestionsAnsweredExtractor(questions=3,llm=llm, num_workers=num_workers)

vector_index = VectorStoreIndex.from_documents(
	documents,
	transformations=[text_splitter, qa_extractor, title_extractor],
	storage_context=storage_context,
	embed_model = Settings.embed_model,
)

# I can see that it has stored 80 Nodes into the chromadb (which is the same number of nodes as the first example) so it appears
# that all the processing work is happending correctly.
# When I run queries against the loaded chromadb data, all similarity scores are 0.000000

# its as if there are no embeddings accociated with the nodes stored in the chromadb ???

  
# when I use just the vanialla vector_index.storage_context.persist(persist_dir='data/RAG_Storage_2') from the first example, 
# my similarity scores on searches I know should work return reasonable values in the range of 0.65 to 0.80 and I get valid
# search results.

# for queries against the chromadb I always get some vairiant of "There's nothing in the context matching the query'