PaperManager

Ingesting Papers

PaperManager supports three ways to add papers to your library.

---

PDF Upload

Drag a PDF onto the Library page or click the + button → PDF tab.

What Happens Automatically

The upload follows a multi-step pipeline:

flowchart TD
    A[📄 PDF uploaded] --> B[Extract text with Docling]
    B --> C{DOI / arXiv ID\nin text?}
    C -->|Yes| D[Semantic Scholar API]
    D -->|Not found| E[CrossRef API]
    C -->|No title found| F[Ollama LLM on first 3 000 chars]
    F -->|Ollama unavailable| G[Regex heuristics]
    D --> H[Confirmation modal]
    E --> H
    F --> H
    G --> H
    H -->|User confirms| I[Upload PDF to Google Drive]
    I --> J[Create Paper node in Neo4j]
    J --> K[Generate AI summary — Claude Opus]
    K --> L[Suggest topics — Claude Haiku]
    L --> M[Extract references]
    M --> N[Extract figures from PDF]
    N --> O[✅ Paper ready in library]

Step-by-Step: The Upload Modal

After choosing a PDF you walk through up to five optional steps (each can be toggled in Settings):

1. Metadata confirmation — review extracted title, year, authors, DOI, venue; fix any errors
2. Source — record how you found this paper (person, LinkedIn, Twitter, conference, newsletter…)
3. Summary prompt — edit the Claude prompt before the AI summary is generated
4. References review — approve which references to save as CITES links
5. Tags — review AI-suggested tags before saving

Upload Modal Options

Option	Default	Description
Source step	on	Record how you found the paper
Summary prompt step	on	Edit the AI summary instructions
Auto-save references	off	Skip review and save all references automatically
Tags step	on	Review AI-suggested tags before saving

Metadata Source Badge

After upload each paper card shows a colour-coded badge:

Colour	Source
🟢 Green	Semantic Scholar or CrossRef
🟡 Yellow	Ollama LLM extraction
🔴 Red	Heuristic guess

---

URL / DOI Ingest

Click + → URL / DOI tab. Paste any of:

Input format	Example
arXiv URL	https://arxiv.org/abs/1706.03762
arXiv ID	1706.03762 or arXiv:1706.03762
DOI URL	https://doi.org/10.1038/nature14539
Bare DOI	10.1038/nature14539
PubMed URL	https://pubmed.ncbi.nlm.nih.gov/12345678/
bioRxiv URL	https://www.biorxiv.org/content/10.1101/...
medRxiv URL	https://www.medrxiv.org/content/10.1101/...

What Happens

1. The URL/ID is parsed to detect the source type
2. Metadata is fetched from the appropriate API (arXiv Atom, Semantic Scholar, CrossRef, PubMed eUtils, bioRxiv)
3. If a DOI is found, Semantic Scholar is queried for richer data (citation count, affiliations, topics)
4. Paper is auto-tagged from-url
5. No PDF is stored — the paper is metadata-only

!!! note To get a PDF for a URL-ingested paper, open the Paper Detail page and click Upload PDF.

---

Bulk Import

Go to Bulk Import in the nav bar. Upload or paste a JSON file:

{
  "fetch_pdf": true,
  "project_id": "optional-project-uuid",
  "papers": [
    {"url": "https://arxiv.org/abs/1706.03762"},
    {"arxiv": "1810.04805"},
    {"doi": "10.1038/nature14539"},
    {"url": "https://pubmed.ncbi.nlm.nih.gov/30082513/"},
    {"title": "AlphaFold protein structure prediction"},
    {"title": "CRISPR-Cas9 genome editing", "fetch_pdf": false}
  ]
}

Each entry needs at least one of: url, arxiv, doi, title. Formats can be mixed freely.

A sample file is included at sample_bulk_import.json.

Resolution Order

For each paper entry the system tries, in order:

flowchart LR
    E1["url → URL resolver"] --> E2["arxiv → arXiv API"]
    E2 --> E3["doi → S2 → CrossRef"]
    E3 --> E4["title → S2 title search\n→ arXiv title search\n→ Ollama-improved arXiv"]

PDF Fetching

When fetch_pdf: true:

• arXiv papers — downloaded from arxiv.org/pdf/{id}
• Other papers — Unpaywall API checked for an open-access PDF URL

Progress Stream

Bulk import shows a live log stream as papers are processed. Papers that already exist by DOI are reported as “skipped”. All imported papers are auto-tagged bulk-import.

!!! tip “Project assignment” Set "project_id" in the JSON to automatically add all imported papers to a project.

---

Duplicate Detection

Before a paper is saved, a duplicate check runs against existing papers:

• Exact DOI match → blocked (the paper already exists)
• High title similarity → warning shown in the confirmation modal

If a DOI-matched stub already exists (from reference extraction), the stub is enriched with full metadata rather than creating a duplicate.