Hugging Face to Notion

This workflow is ideal for:

• Researchers looking to automate the collection of recent academic papers from Hugging Face.
• Data Analysts who require structured data from academic resources for analysis or reporting.
• Content Managers needing to keep track of relevant literature and abstracts for content creation or updates.
• Developers interested in integrating academic research into their applications or services.
• Students who want to stay updated with the latest research in their fields of study.

This workflow addresses the challenge of manually tracking and collecting recent academic papers from Hugging Face. By automating the process, users can:

• Save time and effort in searching for relevant papers.
• Ensure they do not miss important publications by running the workflow on a scheduled basis.
• Maintain an organized database of papers and their abstracts in Notion, making it easier to access and reference them later.

1. Schedule Trigger: The workflow begins by triggering automatically every week on specified days at 8 AM.
2. Request Hugging Face Paper: It sends a request to the Hugging Face papers API to fetch papers published one day prior.
3. Extract Hugging Face Paper: The HTML content of the response is parsed to extract the URLs of the papers.
4. Split Out: The URLs are split into separate items for further processing.
5. Loop Over Items: Each extracted URL is processed in a loop.
6. Check Paper URL Existed: For each URL, the workflow checks if it already exists in the Notion database to avoid duplicates.
7. Request Hugging Face Paper Detail: If the paper does not exist, it fetches detailed information about the paper using the URL.
8. Extract Hugging Face Paper Abstract: The abstract and title of the paper are extracted from the detailed response.
9. OpenAI Analysis Abstract: The abstract is sent to OpenAI for analysis, extracting key details such as core introduction, keywords, data and results, technical details, and classification.
10. Store Abstract Notion: Finally, the extracted information is stored in a Notion database, creating a structured record of the paper.