Add Adanos Market Sentiment integration - #553
Conversation
|
@alexander-schneider is attempting to deploy a commit to the deepset Team on Vercel. A member of the Team first needs to authorize it. |
kacperlukawski
left a comment
kacperlukawski
left a comment
There was a problem hiding this comment.
Thank you, @alexander-schneider Have you created this integration based on an actual need coming from a user? I would like to understand whether it is something truly useful for Haystack developers.
| ```python | ||
| from haystack import Pipeline | ||
| from haystack_integrations.components.tools.adanos import AdanosMarketSentiment | ||
|
|
||
| pipeline = Pipeline() | ||
| pipeline.add_component("sentiment", AdanosMarketSentiment()) | ||
|
|
||
| result = pipeline.run( | ||
| { | ||
| "sentiment": { | ||
| "operation": "trending", | ||
| "asset_type": "stock", | ||
| "source": "reddit", | ||
| "limit": 5, | ||
| } | ||
| } | ||
| ) | ||
| ``` |
There was a problem hiding this comment.
What would be a real-life scenario where this component would be useful? A single-component pipeline does not clearly justify that.
There was a problem hiding this comment.
Agreed — the single-component pipeline only demonstrated invocation, not the Haystack-specific value. I replaced it with a concrete investment-research Agent workflow that combines two tools:
- an
InMemoryBM25Retrieverover private research notes, and AdanosMarketSentimentfor external sentiment evidence.
The Agent is instructed to keep internal and external evidence distinct, identify conflicting signals, and produce a research assessment rather than trading advice. I also verified that the documented Agent and both ComponentTool definitions construct successfully against the current Haystack and adanos-haystack packages.
|
Thanks for asking. This was not commissioned by a single Haystack user, but it was informed by broader feedback from Adanos users who want market-sentiment data available in agentic investment-research workflows. We believe financial research is moving toward agent-based systems, where proprietary documents, company information, and external market signals are combined dynamically. That is why we started making Adanos data available early as a Haystack-compatible tool. A concrete use case is an investment-research agent that retrieves filings or internal research through Haystack, calls Adanos for stock sentiment and attention signals across Reddit, X / FinTwit, news, and Polymarket, and then produces a combined research or risk summary. I agree that the current single-component pipeline does not demonstrate this clearly. I can replace it with a realistic agent or multi-component research example. |
e075eed to
4783c3f
Compare
2 participants
Summary
Integration package
The public package repository is https://github.com/adanos-software/adanos-haystack. It provides a serializable
AdanosMarketSentimentcomponent with synchronous and asynchronous execution and supports HaystackComponentTool.Supported operations are sentiment, trending assets, market sentiment, asset comparison, search, and dataset statistics. Stock sources include Reddit, X / FinTwit, financial news, and Polymarket; crypto currently uses Reddit.
Validation