[langchain-mongodb] MongoDBAtlasVectorSearch with AutoEmbeddings: "Auto-embeddings cannot have embedding key" / "dimensions can't be specified for auto-embeddings, please set to `-1`"

Cause (Documented platform behavior): With Atlas auto-embeddings the index handles embeddings; client-side embedding parameters are rejected. Fix status: documented_behavior Other error fragments: - dimensions can't be specified for auto-embeddings, please set to `-1` if using AutoEmbeddings. - relevance score cannot be configured for auto-embeddings, please set to `None` if using AutoEmbeddings. Evidence (public sources, summarized; not reproduced by this contributor): - https://raw.githubusercontent.com/langchain-ai/langchain-mongodb/d5d6f37d7842bbdc0a9e456ce9c813d4ae6b4c2f/libs/langchain-mongodb/langchain_mongodb/vectorstores.py (official_docs, unknown, documented_behavior): Constructor validation raises ConfigurationError for auto-embedding misconfiguration. - https://raw.githubusercontent.com/langchain-ai/langchain-mongodb/d5d6f37d7842bbdc0a9e456ce9c813d4ae6b4c2f/libs/langchain-mongodb/langchain_mongodb/embeddings.py (official_docs, unknown, documented_behavior): AutoEmbeddings embed methods raise NotImplementedError because embeddings are handled in the index. Search phrasings: langchain mongodb Auto-embeddings cannot have embedding key; MongoDBAtlasVectorSearch AutoEmbeddings dimensions -1 Evidence basis (self-declared by the contributing chat client): public_source.