Bijay Gurung

EmbeddingRetriever does not account for longer documents

3

**Describe the bug** The EmbeddingRetriever does not account for long sequences. More precisely, it passes on the documents to the underlying encoder model (sentence-transformer) which truncates the sequence before embedding....

Tutorial 09: Update to EmbeddingRetriever Training

14

## Overview With #2887, we replaced DPR with `EmbeddingRetriever` in Tutorial 06. Now, we might want to do the same for Tutorial 09 which covers training (or fine-tuning) a DPR...

**Describe the bug** Trying to train DPR with `num_positives` > 1 results in an error. This was discovered in #3084 **Error message** ``` Traceback (most recent call last): File "/home/user/train_dpr/train_dpr.py",...

Lazy Run/Evaluation of components based on downstream necessity

2

Came up in #6836 Wondering if there is a way to setup a component to only run if there is a downstream component needing its output. E.g. Possible usage: the...

Tutorial 09: Update to EmbeddingRetriever Training

19

## Overview With deepset-ai/haystack#2887, we replaced DPR with `EmbeddingRetriever` in Tutorial 06. Now, we might want to do the same for Tutorial 09 which covers training (or fine-tuning) a DPR...

feat: Add component SimilarDocumentsRetriever

3

### Related Issues Evolved from #5629 and #5666 ### Proposed Changes: Addition of a new component `SimilarDocumentsRetriever`. This component retrieves similar documents for each of the given documents for each...

Minor point but could we float up the `__version__` in the root `__init__` so that we can do `haystack.__version__` instead of `haystack.version.__version__`? I think that's more intuitive when users want...

**Note:** This issue seems to be a pipeline.run bug and is fixed by [the subgraphs branch](https://github.com/deepset-ai/haystack/tree/subgraphs). In a pipeline with dynamic llm-generated filters (based on query, similar to [this blog-post](https://haystack.deepset.ai/blog/extracting-metadata-filter))...

Surface underlying error on bedrock integration instead of assuming connection issue

1

In the bedrock components, [document_embeder.py#L219](https://github.com/deepset-ai/haystack-core-integrations/blob/fa8b9e480e5c5173206fd7ac71aba51d4cf4fd74/integrations/amazon_bedrock/src/haystack_integrations/components/embedders/amazon_bedrock/document_embedder.py#L219), [generator.py#L236](https://github.com/deepset-ai/haystack-core-integrations/blob/fa8b9e480e5c5173206fd7ac71aba51d4cf4fd74/integrations/amazon_bedrock/src/haystack_integrations/components/generators/amazon_bedrock/generator.py#L236), there is an exception catch on run which then produces a message about it being a connection error regardless of the underlying issue. ```python...