• created in 2023.06

Perhaps this is a more straightforward alternative to Zotero and CiteSpace. All the features are streamlined and easy to grasp.

The issue with Zotero is that it's challenging to add custom fields and it doesn't leverage the power of AI well. CiteSpace is complicated to operate; many of its algorithms are like black boxes.

While writing my PhD research proposal, I developed this program to address the above pain points and to offer a more customized solution.

The program can download papers based on their titles, obtain the paper's metadata and main content, conduct keyword co-occurrence analysis, and also use AI to add custom fields and pose questions to the papers.

Github: https://github.com/xiahan4956/Paper_manager

main.py

ask_paper.py

What are the features?

• Download: Download paper meta data(doi,abstarct,publish_year,if factor,citations) by titles.
• Parse: Add any columns by AI (openai,cluade) parsing.
• Analysis: Use community detection and centrality to analysis keywords.These algorithms are similar as the citepsace!
• Detail asking: If you have Claude api,you could ask any paper.
• Export: Export data to Excel.(Use navicat)

What's the difference between this and Zotero?

• You could auto download paper any data by titles. Zotero needs doi.
• You could add any columns by AI (openai,cluade) parsing. Zotero didn't support.If you use AI to parse papers,you understand what I mean.
• You could find keywords more detail information,such as the center betweenness of keywords,However Zotero didn't support.

What's the difference between this and Citespace?

• The 'keywords analysis' uses AI to extarct keywords. But the way of Citespace operates as a black box.
• Use community detection and centrality to analysis keywords.These algorithms are same as the citepsace.

How to use

1. You need to have openai api key or claude api key to parse paper.I recommend you to use the Claude api key, which is free and supports long text parsing.
2. You need to install Navicat to operate the database.
3. You need to install Chrome to download papers.
4. Clone this repository
5. Execute pip install -r requirements.txt to install Python packages.
6. Use Navicat to open 'data/paper.db' and create a table. Then copy paper titles to this table. The title column name must be 'title'.
7. Change .env.template to .env, and set PAPER_TABLE and MODEL, AI API_KEY
8. Execute main.py
9. If you have Claude API key,you could use ask_paper.py to ask any paper. Just put the paper in paper.txt