* Getting device_map working to support 8bit loading and multi device inference
* Update to take account the device specified by the user
* add release notes
* Add device_map support for ExtractiveReader
* Update test
* Update to model that doesn't have issues
* Update test
* Update pytest approx
* Update release notes
* Start supporting device map
* Update ExtractiveReader to use new ComponentDevice
* Update similarity ranker to follow extractive reader implementation
* Fixing pylint
* Make mypy mostly happy
* Add new unit test to test device_map
* Adding unit tests
* Some refactoring
* Add more tests
* Add more tests
* Add another unit test
* Update first_device property to return a ComponentDevice to be able to use the to methods
* Updating tests for test_device
* Update tests and now explicitly modify device_map in model_kwargs
* Update haystack/utils/hf.py
Co-authored-by: Madeesh Kannan <shadeMe@users.noreply.github.com>
* Make mypy happy
* mypy
* Remove unneeded optional flag
* Update ExtractiveReader with new logic
* Update ranker to follow new logic
* Removing unneeded code
* Make mypy happy
* fxi pylint
* Fix test
* Adding unit tests for device_map="auto"
* Add unit tests for ranker
* PR comments
* Make util method
* Adding unit tests
* Fix type annotation
* Fix pylint
* Fix test
---------
Co-authored-by: Madeesh Kannan <shadeMe@users.noreply.github.com>
* rename model parameter in the openai doc embedder
* fix tests for openai doc embedder
* rename model parameter in the openai text embedder
* fix tests for openai text embedder
* rename model parameter in the st doc embedder
* fix tests for st doc embedder
* rename model parameter in the st backend
* fix tests for st backend
* rename model parameter in the st text embedder
* fix tests for st text embedder
* fix docstring
* fix pipeline utils
* fix e2e
* reno
* fix the indexing pipeline _create_embedder function
* fix e2e eval rag pipeline
* pytest
* support single metadata dict in markdown2document
* reno
* unwrap list
* direct key access
* typing
* add example of indexing pipeline using Multiplexer
* reno
* replace metadata w meta in tests/examples
* do not touch already broken e2e tests
* Revert "do not touch already broken e2e tests"
This reverts commit 1f911920d98954b57daacfe8d8ed02fd77d136db.
* Improve rag and indexing pipelines
* Update examples
* Simplify user interface and code, improve embedder model
* Improve default vals for embedder
* resolve typing
* resolve typing 2
* Fix unit test
---------
Co-authored-by: Timo Möller <timo.moeller@deepset.ai>
* add example for pipeline loop
* add pydantic to CI
* Fix comment
---------
Co-authored-by: Stefano Fiorucci <44616784+anakin87@users.noreply.github.com>
* added hybrid search example
Added an example about hybrid search for faq pipeline on covid dataset
* formatted with back formatter
* renamed document
* fixed
* fixed typos
* added test
added test for hybrid search
* fixed withespaces
* removed test for hybrid search
* fixed pylint
* commented logging
* updated hybrid search example
* release notes
* Update hybrid_search_faq_pipeline.py-815df846dca7e872.yaml
* Update hybrid_search_faq_pipeline.py
* mention hybrid search example in release notes
* reduce installed dependencies in examples test workflow
* do not install cuda dependencies
* skip models if API key not set; delete document indices
* skip models if API key not set; delete document indices
* skip models if API key not set; delete document indices
* keep roberta-base model and inference extra
* pylint
* disable pylint no-logging-basicconfig rule
---------
Co-authored-by: Julian Risch <julian.risch@deepset.ai>
* Add lost in the middle ranker
* Add release note
* Julian's feedback: more precise version of truncate
* Better comments for the litm algorithm
* Sebastian PR feedback
* Add check for invalid values of word_count_threshold
* Remove _truncate as it is not needed any more
---------
Co-authored-by: Darja Fokina <daria.f93@gmail.com>
* added hybrid search example
Added an example about hybrid search for faq pipeline on covid dataset
* formatted with back formatter
* renamed document
* fixed
* fixed typos
* added test
added test for hybrid search
* fixed withespaces
* removed test for hybrid search
* fixed pylint
* commented logging
* Update Claude support with the latest models, new streaming API, context window sizes
* Use Github Anthropic SDK link for tokenizer, revert _init_tokenizer
* Change example key name to ANTHROPIC_API_KEY
* Extract ToolsManager, add it to Agent by the composition
* PR feedback Massi
---------
Co-authored-by: Massimiliano Pippi <mpippi@gmail.com>
Co-authored-by: Darja Fokina <daria.f93@gmail.com>
* Deprecate Seq2SeqGenerator
* changed the warning to include suggestion
* Added example and msg to API reference docs
* Added RAG deprecation
* renamed name to adapt to naming conven
* update docstrings
---------
Co-authored-by: Mayank Jobanputra <mayankjobanputra@gmail.com>
Co-authored-by: Darja Fokina <daria.f93@gmail.com>
