bogdankostic
bogdankostic
**Is your feature request related to a problem? Please describe.** LayoutLM is a transformer-based model that is able to take PDFs as input and perform different tasks on them. We...
**Describe the bug** `write_labels` currently crashes with a `BulkIndexError` if we try to write labels whose context is longer than 32766 bytes. I suspect that this is due to the...
**Describe the solution you'd like** Parsr has built-in heading detection. We should make use of it and add headline information of PDFs to the converted Documents. As far as I...
**Describe the solution you'd like** Markdown files tag headlines with the `#` character. Different levels of headings are marked by the number of `#` characters. We should use this information...
With #5027, we added support for Elasticsearch 8 in `ElasticsearchDocumentStore`. As a transition phase, `ElasticsearchDocumentStore` will still use Elasticsearch 7 as default for Haystack 1.19 and 1.20. Starting from Haystack...
### Related Issues - fixes #4598 ### Proposed Changes: This PR adds the custom loader `CustomPythonLoader` for `pydoc-markdown` and sets it as the loader for all documentation config files. `CustomPythonLoader`...
We provide a Helm chart for Haystack in our [haystack-helm repository](https://github.com/deepset-ai/haystack-helm). Let’s use this chart to deploy our _Explore the World_ demo with Kubernetes.
We just merged this [PR](https://github.com/deepset-ai/haystack/pull/3001) which added support for dot product and l2 distance for the `WeaviateDocumentStore`. We need, therefore, to update the _Choosing the Right Document Store_ section for...