KeyError: 'moduleConfig' Error Message

[kihumban]

I am getting KeyError: ‘moduleConfig’ error message when using weaviate with langchain. Weaviate is running in a docker container.

This is the code generating the error.

client = weaviate.connect_to_local()
embeddings = OpenAIEmbeddings()
db2 = WeaviateVectorStore.from_documents(docs_3, embeddings, client=client)

Here is my docker-compose.yml

This post on langchain github on the same issue recommended removing the DEFAULT_VECTORIZER_MODULE: ‘multi2vec-clip’ from the docker compose file. I did that, restarted the container but I am still getting the error.

Any suggestions?

[DudaNogueira]

hi @kihumban !!

Welcome to our community

We have a nice recipe for langchain here, that I crafted myself, check it out:

recipes/integrations/llm-frameworks/langchain/loading-data at main ·...

If you could share here the sull running example as well as the docker compose in text, I can try reproducing that.

The recipe I liked above has all you need to use all Langchain features with Weaviate.

Let me know if I can help you on that

Thanks!

[kihumban]

@DudaNogueira thank you for sharing the recipe. I will check it out. This forum does not allow me to attach any file other than images. It flag text for having more than 2 hyperlinks. What is the best way of sharing the code and the docker file?

[DudaNogueira]

You can paste the docker and code here too

[kihumban]

I am basically working on this Langchain Weaviate library example. Here is the code. The last expression is the one generating the error.
############
from langchain_community.document_loaders import TextLoader
from langchain_openai import OpenAIEmbeddings
from langchain_text_splitters import CharacterTextSplitter
import weaviate
from langchain_weaviate.vectorstores import WeaviateVectorStore
from langchain.chains import RetrievalQAWithSourcesChain
from langchain_openai import OpenAI

loader = TextLoader(“data/state_of_the_union.txt”)
documents = loader.load()
text_splitter = CharacterTextSplitter(chunk_size=1000, chunk_overlap=0)
docs_3 = text_splitter.split_documents(documents)
embeddings = OpenAIEmbeddings()

client = weaviate.connect_to_local()
db2 = WeaviateVectorStore.from_documents(docs_3, embeddings, client=client)
##############

######################
Initially I had the DEFAULT_VECTORIZER_MODULE set to ‘multi2vec-clip’. However I removed it based to advise on this thread. Apparently the issue is being cause by the conflict between declaring embeddings = OpenAIEmbeddings() while DEFAULT_VECTORIZER_MODULE set to ‘multi2vec-clip’. However, updating my docker file by setting DEFAULT_VECTORIZER_MODULE to ‘none’ did not resolve the issue.

[kihumban]

Here is the LangChain thread regarding the error

Error when using `WeaviateVectorStore.from_documents()`

[DudaNogueira]

Can you remove the DEFAULT_VECTORIZER_MODULE altogether?

Also, can you copy the docker compose here? Otherwise it’s really copy and paste from the image, as it gets all messed up

Have you looked into the recipe I pasted?

It should give you all the steps to start using Langchain with Weaviate the proper way.

Thanks!

[kihumban]

I removed the DEFAULT_VECTORIZER_MODULE and I am still getting the same error. However, I have worked through the the code in the recipe you shared and that is working ok. The only difference is that the recipe is using Embedded Weaviate rather than stand-alone server as was in my code. I am not sure if that has an impact. I will modify the recipe to try with client = weaviate.connect_to_local().

The forum platform does not allow me to copy the docker compose file. Every time I do that it flags the post with the following message.

[DudaNogueira]

Hi!

As long as you have a working weaviate client (connect to embedded/local/cloud) it should work accordingly.

You can paste the the docker compose here, like so:

---
services:
  weaviate:
    command:
    - --host
    - 0.0.0.0
    - --port
    - '8080'
    - --scheme
    - http
    image: cr.weaviate.io/semitechnologies/weaviate:1.27.0
    ports:
    - 8080:8080
    - 50051:50051
    volumes:
    - weaviate_data:/var/lib/weaviate
    restart: on-failure:0
    environment:
      QUERY_DEFAULTS_LIMIT: 25
      AUTHENTICATION_ANONYMOUS_ACCESS_ENABLED: 'true'
      PERSISTENCE_DATA_PATH: '/var/lib/weaviate'
      DEFAULT_VECTORIZER_MODULE: 'none'
      ENABLE_API_BASED_MODULES: 'true'
      CLUSTER_HOSTNAME: 'node1'
volumes:
  weaviate_data:
...

[kihumban]

I have tried pasting the docker compose here but it does not work. I keep getting the error message I indicated above.

[DudaNogueira]

Oh, I mean, the content of the docker compose file.

Feel free to reach out to me on our slack. I would love to help you there too!

[kihumban]

I run the langchain-simple-pdf-multitenant.ipynb recipe using local stand-alone server instead of embedded and run into the same error. The only section of the code that changed was the initialization of the client

client = weaviate.connect_to_local(
headers={
“X-OpenAi-Api-Key”: os.environ.get(“OPENAI_API_KEY”), # Replace with your Cohere key
}
)

Here is my docker compose:

services:
  weaviate:
    command:
    - --host
    - 0.0.0.0
    - --port
    - '8080'
    - --scheme
    - http
    image: cr.weaviate.io/semitechnologies/weaviate:1.26.4
    ports:
    - 8080:8080
    - 50051:50051
    volumes:
    - weaviate_data:/var/lib/weaviate
    restart: on-failure:0
    environment:
      CLIP_INFERENCE_API: 'http://multi2vec-clip:8080'
      QUERY_DEFAULTS_LIMIT: 25
      AUTHENTICATION_ANONYMOUS_ACCESS_ENABLED: 'true'
      PERSISTENCE_DATA_PATH: '/var/lib/weaviate'
      ENABLE_API_BASED_MODULES: 'true'
      ENABLE_MODULES: 'multi2vec-clip,text2vec-openai,generative-openai'
      CLUSTER_HOSTNAME: 'node1'
  multi2vec-clip:
    image: cr.weaviate.io/semitechnologies/multi2vec-clip:sentence-transformers-clip-ViT-B-32-multilingual-v1
    environment:
      ENABLE_CUDA: '0'
volumes:
  weaviate_data:
...

[kihumban]

Here is the expression and the associated error dump:

db = WeaviateVectorStore.from_documents(docs, embeddings, client=client, index_name=“WikipediaLangChainMT”, tenant=“brazil”)

---------------------------------------------------------------------------
KeyError                                  Traceback (most recent call last)
Cell In[11], line 1
----> 1 db = WeaviateVectorStore.from_documents(docs, embeddings, client=client, index_name="WikipediaLangChainMT", tenant="brazil")

File /Library/Frameworks/Python.framework/Versions/3.12/lib/python3.12/site-packages/langchain_core/vectorstores/base.py:835, in VectorStore.from_documents(cls, documents, embedding, **kwargs)
    833 texts = [d.page_content for d in documents]
    834 metadatas = [d.metadata for d in documents]
--> 835 return cls.from_texts(texts, embedding, metadatas=metadatas, **kwargs)

File /Library/Frameworks/Python.framework/Versions/3.12/lib/python3.12/site-packages/langchain_weaviate/vectorstores.py:487, in WeaviateVectorStore.from_texts(cls, texts, embedding, metadatas, tenant, client, index_name, text_key, relevance_score_fn, **kwargs)
    475 attributes = list(metadatas[0].keys()) if metadatas else None
    477 weaviate_vector_store = cls(
    478     client,
    479     index_name,
   (...)
    484     use_multi_tenancy=tenant is not None,
    485 )
--> 487 weaviate_vector_store.add_texts(texts, metadatas, tenant=tenant, **kwargs)
    489 return weaviate_vector_store

File /Library/Frameworks/Python.framework/Versions/3.12/lib/python3.12/site-packages/langchain_weaviate/vectorstores.py:167, in WeaviateVectorStore.add_texts(self, texts, metadatas, tenant, **kwargs)
    164 if self._embedding:
    165     embeddings = self._embedding.embed_documents(list(texts))
--> 167 with self._client.batch.dynamic() as batch:
    168     for i, text in enumerate(texts):
    169         data_properties = {self._text_key: text}

File /Library/Frameworks/Python.framework/Versions/3.12/lib/python3.12/site-packages/weaviate/collections/batch/client.py:179, in _BatchClientWrapper.dynamic(self, consistency_level)
    177 self._batch_mode: _BatchMode = _DynamicBatching()
    178 self._consistency_level = consistency_level
--> 179 return self.__create_batch_and_reset()

File /Library/Frameworks/Python.framework/Versions/3.12/lib/python3.12/site-packages/weaviate/collections/batch/client.py:135, in _BatchClientWrapper.__create_batch_and_reset(self)
    133 def __create_batch_and_reset(self) -> _ContextManagerWrapper[_BatchClient]:
    134     if self._vectorizer_batching is None or not self._vectorizer_batching:
--> 135         configs = self.__config.list_all(simple=True)
    137         vectorizer_batching = False
    138         for config in configs.values():

File /Library/Frameworks/Python.framework/Versions/3.12/lib/python3.12/site-packages/weaviate/collections/collections/sync.py:286, in _Collections.list_all(self, simple)
    265 def list_all(
    266     self, simple: bool = True
    267 ) -> Union[Dict[str, CollectionConfig], Dict[str, CollectionConfigSimple]]:
    268     """List the configurations of the all the collections currently in the Weaviate instance.
    269 
    270     Arguments:
   (...)
    284             If Weaviate reports a non-OK status.
    285     """
--> 286     return self.__loop.run_until_complete(self.__collections.list_all, simple)

File /Library/Frameworks/Python.framework/Versions/3.12/lib/python3.12/site-packages/weaviate/event_loop.py:40, in _EventLoop.run_until_complete(self, f, *args, **kwargs)
     38     raise WeaviateClosedClientError()
     39 fut = asyncio.run_coroutine_threadsafe(f(*args, **kwargs), self.loop)
---> 40 return fut.result()

File /Library/Frameworks/Python.framework/Versions/3.12/lib/python3.12/concurrent/futures/_base.py:456, in Future.result(self, timeout)
    454     raise CancelledError()
    455 elif self._state == FINISHED:
--> 456     return self.__get_result()
    457 else:
    458     raise TimeoutError()

File /Library/Frameworks/Python.framework/Versions/3.12/lib/python3.12/concurrent/futures/_base.py:401, in Future.__get_result(self)
    399 if self._exception:
    400     try:
--> 401         raise self._exception
    402     finally:
    403         # Break a reference cycle with the exception in self._exception
    404         self = None

File /Library/Frameworks/Python.framework/Versions/3.12/lib/python3.12/site-packages/weaviate/collections/collections/async_.py:304, in _CollectionsAsync.list_all(self, simple)
    285 """List the configurations of the all the collections currently in the Weaviate instance.
    286 
    287 Arguments:
   (...)
    301         If Weaviate reports a non-OK status.
    302 """
    303 _validate_input([_ValidateArgument(expected=[bool], name="simple", value=simple)])
--> 304 return await self._get_all(simple=simple)

File /Library/Frameworks/Python.framework/Versions/3.12/lib/python3.12/site-packages/weaviate/collections/collections/base.py:76, in _CollectionsBase._get_all(self, simple)
     74 assert res is not None
     75 if simple:
---> 76     return _collection_configs_simple_from_json(res)
     77 return _collection_configs_from_json(res)

File /Library/Frameworks/Python.framework/Versions/3.12/lib/python3.12/site-packages/weaviate/collections/classes/config_methods.py:327, in _collection_configs_simple_from_json(schema)
    323 def _collection_configs_simple_from_json(
    324     schema: Dict[str, Any]
    325 ) -> Dict[str, _CollectionConfigSimple]:
    326     return {
--> 327         schema["class"]: _collection_config_simple_from_json(schema) for schema in schema["classes"]
    328     }

File /Library/Frameworks/Python.framework/Versions/3.12/lib/python3.12/site-packages/weaviate/collections/classes/config_methods.py:250, in _collection_config_simple_from_json(schema)
    242 def _collection_config_simple_from_json(schema: Dict[str, Any]) -> _CollectionConfigSimple:
    243     return _CollectionConfigSimple(
    244         name=schema["class"],
    245         description=schema.get("description"),
    246         generative_config=__get_generative_config(schema),
    247         properties=_properties_from_config(schema) if schema.get("properties") is not None else [],
    248         references=_references_from_config(schema) if schema.get("properties") is not None else [],
    249         reranker_config=__get_rerank_config(schema),
--> 250         vectorizer_config=__get_vectorizer_config(schema),
    251         vectorizer=__get_vectorizer(schema),
    252         vector_config=__get_vector_config(schema, simple=True),
    253     )

File /Library/Frameworks/Python.framework/Versions/3.12/lib/python3.12/site-packages/weaviate/collections/classes/config_methods.py:88, in __get_vectorizer_config(schema)
     86 def __get_vectorizer_config(schema: Dict[str, Any]) -> Optional[_VectorizerConfig]:
     87     if __is_vectorizer_present(schema) is not None and schema.get("vectorizer", "none") != "none":
---> 88         vec_config: Dict[str, Any] = schema["moduleConfig"].pop(schema["vectorizer"])
     89         try:
     90             vectorizer = Vectorizers(schema["vectorizer"])

KeyError: 'moduleConfig'

[kihumban]

Is this some kind of a bug? It seem it’s looking for a property called “moduleConfig” in the collection object. But when I look at the collection object configuration (below), there is no property “moduleConfig”

<weaviate.Collection config={
  "name": "WikipediaLangChainMT",
  "description": null,
  "generative_config": {
    "generative": "generative-openai",
    "model": {}
  },
  "inverted_index_config": {
    "bm25": {
      "b": 0.75,
      "k1": 1.2
    },
    "cleanup_interval_seconds": 60,
    "index_null_state": false,
    "index_property_length": false,
    "index_timestamps": false,
    "stopwords": {
      "preset": "en",
      "additions": null,
      "removals": null
    }
  },
  "multi_tenancy_config": {
    "enabled": true,
    "auto_tenant_creation": true,
    "auto_tenant_activation": true
  },
  "properties": [],
  "references": [],
  "replication_config": {
    "factor": 1,
    "async_enabled": false
  },
  "reranker_config": null,
  "sharding_config": null,
  "vector_index_config": {
    "quantizer": null,
    "cleanup_interval_seconds": 300,
    "distance_metric": "cosine",
    "dynamic_ef_min": 100,
    "dynamic_ef_max": 500,
    "dynamic_ef_factor": 8,
    "ef": -1,
    "ef_construction": 128,
    "flat_search_cutoff": 40000,
    "max_connections": 32,
    "skip": false,
    "vector_cache_max_objects": 1000000000000
  },
  "vector_index_type": "hnsw",
  "vectorizer_config": {
    "vectorizer": "text2vec-openai",
    "model": {
      "baseURL": "https://api.openai.com",
      "model": "ada"
    },
    "vectorize_collection_name": true
  },
  "vectorizer": "text2vec-openai",
  "vector_config": null
}>

[DudaNogueira]

Hi!

This example in our recipes uses text2vec-openai and generative-openai

And those modules are enabled:

ENABLE_MODULES: 'multi2vec-clip,text2vec-openai,generative-openai'

So it should work.

have you restarted your docker so the changes take effect?

you can call:

docker compose up -d

[kihumban]

I have all those modules enabled in my docker compose as previously indicated. I have restarted the docker several times, the outcome has been consistently the same - KeyError: ‘moduleConfig’.

Below is my docker compose at it currently is. I just run the program and got the same error.

services:
  weaviate:
    command:
    - --host
    - 0.0.0.0
    - --port
    - '8080'
    - --scheme
    - http
    image: cr.weaviate.io/semitechnologies/weaviate:1.26.4
    ports:
    - 8080:8080
    - 50051:50051
    volumes:
    - weaviate_data:/var/lib/weaviate
    restart: on-failure:0
    environment:
      CLIP_INFERENCE_API: 'http://multi2vec-clip:8080'
      QUERY_DEFAULTS_LIMIT: 25
      AUTHENTICATION_ANONYMOUS_ACCESS_ENABLED: 'true'
      PERSISTENCE_DATA_PATH: '/var/lib/weaviate'
      ENABLE_MODULES: 'multi2vec-clip,text2vec-openai,generative-openai'
      ENABLE_API_BASED_MODULES: 'true'
      CLUSTER_HOSTNAME: 'node1'
  multi2vec-clip:
    image: cr.weaviate.io/semitechnologies/multi2vec-clip:sentence-transformers-clip-ViT-B-32-multilingual-v1
    environment:
      ENABLE_CUDA: '0'
volumes:
  weaviate_data:
...

[DudaNogueira]

Hi! This is strange.

I have used your docker compose, and was able to run this:

client.collections.delete("Kihu")
collection = client.collections.create(
    "Kihu",
    vectorizer_config=wvc.config.Configure.Vectorizer.text2vec_openai(),
    generative_config=wvc.config.Configure.Generative.openai(),
)
collection.data.insert({"text": "This is a test"})
obj = collection.generate.fetch_objects(single_prompt="translate {text} to spanish", include_vector=True).objects[0]
print("Generation", obj.generated)
print("Generation", obj.vector)

Let me know if you can run this code.

Thanks!

[DudaNogueira]

Also, you can check what are the installed modules in a server with:

client.get_meta().get("modules")

[kihumban]

@DudaNogueira

client.get_meta().get("modules")

the modules are there. See below.

{'generative-anthropic': {'documentationHref': 'https://docs.anthropic.com/en/api/getting-started',
  'name': 'Generative Search - Anthropic'},
 'generative-anyscale': {'documentationHref': 'https://docs.anyscale.com/endpoints/overview',
  'name': 'Generative Search - Anyscale'},
 'generative-aws': {'documentationHref': 'https://docs.aws.amazon.com/bedrock/latest/APIReference/welcome.html',
  'name': 'Generative Search - AWS'},
 'generative-cohere': {'documentationHref': 'https://docs.cohere.com/reference/chat',
  'name': 'Generative Search - Cohere'},
 'generative-databricks': {'documentationHref': 'https://docs.databricks.com/en/machine-learning/foundation-models/api-reference.html#completion-task',
  'name': 'Generative Search - Databricks'},
 'generative-friendliai': {'documentationHref': 'https://docs.friendli.ai/openapi/create-chat-completions',
  'name': 'Generative Search - FriendliAI'},
 'generative-mistral': {'documentationHref': 'https://docs.mistral.ai/api/',
  'name': 'Generative Search - Mistral'},
 'generative-octoai': {'documentationHref': 'https://octo.ai/docs/text-gen-solution/getting-started',
  'name': 'Generative Search - OctoAI'},
 'generative-openai': {'documentationHref': 'https://platform.openai.com/docs/api-reference/completions',
  'name': 'Generative Search - OpenAI'},
 'generative-palm': {'documentationHref': 'https://cloud.google.com/vertex-ai/docs/generative-ai/chat/test-chat-prompts',
  'name': 'Generative Search - Google PaLM'},
 'multi2vec-palm': {'documentationHref': 'https://cloud.google.com/vertex-ai/generative-ai/docs/embeddings/get-multimodal-embeddings',
  'name': 'Google PaLM Multimodal Module'},
 'reranker-cohere': {'documentationHref': 'https://txt.cohere.com/rerank/',
  'name': 'Reranker - Cohere'},
 'reranker-voyageai': {'documentationHref': 'https://docs.voyageai.com/reference/reranker-api',
  'name': 'Reranker - VoyageAI'},
 'text2vec-aws': {'documentationHref': 'https://docs.aws.amazon.com/bedrock/latest/userguide/titan-embedding-models.html',
  'name': 'AWS Module'},
 'text2vec-cohere': {'documentationHref': 'https://docs.cohere.ai/embedding-wiki/',
  'name': 'Cohere Module'},
 'text2vec-databricks': {'documentationHref': 'https://docs.databricks.com/en/machine-learning/foundation-models/api-reference.html#embedding-task',
  'name': 'Databricks Foundation Models Module - Embeddings'},
 'text2vec-huggingface': {'documentationHref': 'https://huggingface.co/docs/api-inference/detailed_parameters#feature-extraction-task',
  'name': 'Hugging Face Module'},
 'text2vec-jinaai': {'documentationHref': 'https://jina.ai/embeddings/',
  'name': 'JinaAI Module'},
 'text2vec-octoai': {'documentationHref': 'https://octo.ai/docs/text-gen-solution/getting-started',
  'name': 'OctoAI Module'},
 'text2vec-openai': {'documentationHref': 'https://platform.openai.com/docs/guides/embeddings/what-are-embeddings',
  'name': 'OpenAI Module'},
 'text2vec-palm': {'documentationHref': 'https://cloud.google.com/vertex-ai/docs/generative-ai/embeddings/get-text-embeddings',
  'name': 'Google PaLM Module'},
 'text2vec-voyageai': {'documentationHref': 'https://docs.voyageai.com/docs/embeddings',
  'name': 'VoyageAI Module'}}

[DudaNogueira]

were you able to run that code?