Implement Paper Learner

[alirezamshi]

Description

The PaperLearner extracts knowledge from research papers (PDFs) and adds it to the wiki. Currently has placeholder implementation.

Location: src/knowledge/learners/paper_learner.py

Goal

Implement the full paper learning pipeline that:

1. Loads PDF (local or from URL like arXiv)
2. Parses sections and extracts knowledge
3. Creates wiki pages from extracted knowledge

What to Extract

• Abstract and introduction → Overview/concept pages
• Methodology sections → Workflow pages
• Key findings and conclusions → Principle pages
• Formulas and algorithms → Implementation pages
• References → Related pages links

Input Format

{"path": "./paper.pdf"}  # Local file
# or
{"url": "https://arxiv.org/pdf/..."}  # Remote URL

Implementation Steps

1. Load PDF (local or download from URL)
2. Extract text using PyPDF2 or pdfplumber
3. Identify sections (Abstract, Methods, Results, etc.)
4. Extract formulas using OCR if needed
5. Create structured chunks per section
6. Convert chunks to wiki pages
7. Index into knowledge graph

Deliverable

1. Full implementation of PaperLearner.learn() method
2. PDF parsing and section extraction
3. Wiki page generation from paper content

Test

1. Call expert.learn(Source.Paper("./sample_paper.pdf"))
2. Verify wiki pages created with paper sections
3. Verify formulas and algorithms extracted correctly

References

• src/knowledge/learners/paper_learner.py - Current placeholder
• src/knowledge/learners/base.py - Base class and KnowledgeChunk
• src/knowledge/wiki_structure/extension_of_product.md - Wiki page structure

Dependencies

• wiki structure in Structure of Wiki #2
• wiki tools in Wiki Indexer and Retriever #4 and Wiki Editor Tool for Agents #5