feat: Add retrieval for images using multi-modal messages (#870)

.changeset/hip-cycles-cheer.md

packages/core/e2e/fixtures/embeddings/OpenAIEmbedding.ts

packages/core/src/Retriever.ts

packages/core/src/callbacks/CallbackManager.ts

packages/core/src/cloud/LlamaCloudRetriever.ts

packages/core/src/embeddings/ClipEmbedding.ts

packages/core/src/embeddings/GeminiEmbedding.ts

packages/core/src/embeddings/HuggingFaceEmbedding.ts

packages/core/src/embeddings/MistralAIEmbedding.ts

packages/core/src/embeddings/MultiModalEmbedding.ts

packages/core/src/embeddings/OpenAIEmbedding.ts

packages/core/src/embeddings/types.ts

packages/core/src/engines/chat/DefaultContextGenerator.ts

packages/core/src/indices/keyword/index.ts

packages/core/src/indices/summary/index.ts

packages/core/src/indices/vectorStore/index.ts

packages/core/src/llm/ollama.ts

packages/core/src/llm/types.ts

packages/core/src/llm/utils.ts

packages/core/src/postprocessors/rerankers/CohereRerank.ts