deepset-ai / haystack-cookbook Goto Github PK

one error that was easy to fix is this:

change mistralai/Mistral-7B-Instruct-v0.1 to mistralai/Mistral-7B-Instruct-v0.2

but not sure whats up with this error?

ImportError Traceback (most recent call last)
in <cell line: 5>()
3 from haystack.components.preprocessors import DocumentSplitter
4 from haystack.components.rankers import TransformersSimilarityRanker
----> 5 from haystack.components.generators import GPTGenerator
6 from haystack.components.builders.prompt_builder import PromptBuilder
7 from haystack import Pipeline

ImportError: cannot import name 'GPTGenerator' from 'haystack.components.generators' (/usr/local/lib/python3.10/dist-packages/haystack/components/generators/init.py)

Create an cookbook showcasing the new Ragas integration. The cookbook should evaluate a RAG pipeline with various metrics:
RagasEvaluator Docs

when running, after this:

query = "Should I write documentation for my plugin?"
results = querying.run({"retriever": {"queries": [query], "top_k": 3},
"prompt_builder": {"query": query},
"llm":{"generation_kwargs": {"max_new_tokens": 350}}})

got this:

ValueError Traceback (most recent call last)
in <cell line: 2>()
1 query = "Should I write documentation for my plugin?"
----> 2 results = querying.run({"retriever": {"queries": [query], "top_k": 3},
3 "prompt_builder": {"query": query},
4 "llm":{"generation_kwargs": {"max_new_tokens": 350}}})

1 frames
/usr/local/lib/python3.10/dist-packages/haystack/core/pipeline/pipeline.py in _validate_input(self, data)
602 for socket_name, socket in instance.haystack_input._sockets_dict.items():
603 if socket.senders == [] and socket.is_mandatory and socket_name not in component_inputs:
--> 604 raise ValueError(f"Missing input for component {component_name}: {socket_name}")
605 for input_name in component_inputs.keys():
606 if input_name not in instance.haystack_input._sockets_dict:

ValueError: Missing input for component retriever: query

Todo:

• Rename Multiplexer as BranchJoiner (Multiplexer has been deprecated with 2.2.0 and will be removed in 2.4.0)
• (if exists) Multiplexer shouldn't be used to distribute output to multiple components; we can remove the multiplexer from the pipeline if that's the case. We can directly connect one output edge to inputs of multiple components

Create an cookbook showcasing the new UpTrqain integration. The cookbook should evaluate a RAG pipeline with various metrics:
UpTrainEvaluator Docs

Haystack 1.x had this prompt template https://prompthub.deepset.ai/?prompt=deepset%2Fquestion-answering-with-references instructing the LLM to provide references "[1]" to documents in the generated answer. It would be great to have a cookbook showing how to do that with Haystack 2.x.

The existing https://github.com/deepset-ai/haystack-cookbook/blob/main/notebooks/prompt_customization_for_Anthropic.ipynb is related but uses quotes from the documents. What I am looking for is a free text generated answer with references to documents not quotes.

Prompt could be the following:

from haystack.components.builders import PromptBuilder

template = """
Create a concise and informative answer (no more than 50 words) for a given question based solely on the given documents. 
You must only use information from the given documents. Use an unbiased and journalistic tone. Do not repeat text. 
Cite the documents using Document[number] notation. 
If multiple documents contain the answer, cite those documents like ‘as stated in Document[number], Document[number], etc.’. 
If the documents do not contain the answer to the question, say that ‘answering is not possible given the available information.’

Given the following information, answer the question.

{% for document in documents %}
    Document[{{loop.index}}]: {{ document.content }} \n
{% endfor %}

Question: {{question}}
Answer:
"""

prompt_builder = PromptBuilder(template=template)

Create an cookbook showcasing the new DeepEval integration. The cookbook should evaluate a RAG pipeline with various metrics:
DeepEval Docs

this:

transcription = whisper.run(audio_files="/content/podcast.mp3")

returns this error:

TypeError: LocalWhisperTranscriber.run() got an unexpected keyword argument 'audio_files'

Soon we're going to have enough of these that we should organize them a bit. Should it be by topic? Alphabetical? What would make it easiest to skim and find what you're looking for?