KeyError: 'blob_data' when running Pipeline.run() using costume collection settings in WeaviateDocumentStore

[Poaceae1999]

Describe the bug
When running the Pipeline().run method, a KeyError: 'blob_data' is raised. This error originates from the WeaviateDocumentStore class in the haystack_integrations package.

Error message

#   File "C:\RAG\issue_1.py", line 68, in <module>
#     # router for different pipeline
#              ^^^^^^^^^^^^^^^^^^^^^^^
#   File "c:\Users\ttim3\AppData\Local\Programs\Python\Python312\Lib\site-packages\haystack\core\pipeline\pipeline.py", line 771, in run
#     res = comp.run(**last_inputs[name])
#           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
#   File "c:\Users\ttim3\AppData\Local\Programs\Python\Python312\Lib\site-packages\haystack_integrations\components\retrievers\weaviate\embedding_retriever.py", line 74, in run
#     documents = self._document_store._embedding_retrieval(
#                 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
#   File "c:\Users\ttim3\AppData\Local\Programs\Python\Python312\Lib\site-packages\haystack_integrations\document_stores\weaviate\document_store.py", line 470, in _embedding_retrieval
#     return [self._to_document(doc) for doc in result["data"]["Get"][collection_name]]
#             ^^^^^^^^^^^^^^^^^^^^^^
#   File "c:\Users\ttim3\AppData\Local\Programs\Python\Python312\Lib\site-packages\haystack_integrations\document_stores\weaviate\document_store.py", line 232, in _to_document
#     data.pop("blob_data")
# KeyError: 'blob_data'

Expected behavior
Get query result.

Additional context
The script initializes a WeaviateDocumentStore, sets up documents with embeddings, and adds them to the document store. It then sets up a query pipeline with a text embedder and a WeaviateEmbeddingRetriever. When running a query through this pipeline, the error occurs.

It seems that the WeaviateDocumentStore is trying to remove the "blob_data" key from a dictionary, but this key does not exist in the dictionary. Due to the custom collection settings not containing the property "blob_data".
I did not see a requirement to add "blob_data" in the Weaviate or Haystack documentation.

To Reproduce

from haystack import Document
from haystack.components.embedders import SentenceTransformersDocumentEmbedder
from haystack import Pipeline
from haystack_integrations.document_stores.weaviate.document_store import (
    WeaviateDocumentStore,
)
from haystack.components.embedders import (
    SentenceTransformersTextEmbedder,
    SentenceTransformersDocumentEmbedder,
)
from haystack_integrations.components.retrievers.weaviate.embedding_retriever import (
    WeaviateEmbeddingRetriever,
)

# initialize weaviate----
document_store = WeaviateDocumentStore(
    url="http://localhost:8080",
    collection_settings={
        "class": "Article",
        "properties": [
            {
                "name": "title",
                "dataType": ["text"],
            },
            {
                "name": "abstract",
                "dataType": ["text"],
            },
        ],
        "vectorizer": "none",
    },
)
# set up documents----
documents = [
    Document(content="This is first", meta={"title": "hello", "abstract": "hello"}),
    Document(content="This is second", meta={"name": "second"}),
]

text_embedder = SentenceTransformersTextEmbedder()
document_embedder = SentenceTransformersDocumentEmbedder()
document_embedder.warm_up()
documents_with_embeddings = document_embedder.run(documents)["documents"]
document_store.write_documents(documents_with_embeddings)

# set up query pipeline----
query_pipeline = Pipeline()
query_pipeline.add_component("text_embedder", SentenceTransformersTextEmbedder())
query_pipeline.add_component(
    "retriever", WeaviateEmbeddingRetriever(document_store=document_store)
)
query_pipeline.connect("text_embedder.embedding", "retriever.query_embedding")

# query, retrieve and print result----
query = "Who lives in Berlin?"
result = query_pipeline.run({"text_embedder": {"text": query}})

System:

• OS: Windows 11 23H2
• GPU/CPU: RTX 4080 SUPER/Intel(R) Core(TM) i7-14700K
• Haystack version (commit or version number): v2.0.0-beta.8
• DocumentStore: WeaviateDocumentStore
• Reader: None
• Retriever: WeaviateEmbeddingRetriever

[silvanocerza]

This has been solved with #463.