Minimal job

The smallest possible BiblioFetch job: one DOI, every other setting defaulted. Use this as the starting point for any new project.

The job file

examples/minimal-job.toml:

[folder]
target = "papers"

[doi]
list = [
    "arxiv:1008.3477",
]

Two sections only: where PDFs go ([folder]) and what to fetch ([doi]). Everything under [fetch] falls back to built-in defaults — the standard unpaywall / arxiv / direct cascade.

Relative target values resolve against the TOML file's directory (not the caller's cwd), so running from anywhere drops PDFs next to the job.

Running from the shell

bibliofetch run examples/minimal-job.toml

A fresh examples/papers/ directory appears on first run containing the fetched PDF and a parallel .metadata/ folder of per-entry TOML records.

Running from Julia

Loading a job parses and validates the TOML without performing any network I/O. BiblioFetch.run then executes it and returns a FetchJobResult.

using BiblioFetch

job_path = joinpath(@__DIR__, "minimal-job.toml")
job = BiblioFetch.load_job(job_path)

FetchJob("examples", "/home/runner/work/BiblioFetch.jl/BiblioFetch.jl/docs/build/examples/papers", nothing, "/home/runner/work/BiblioFetch.jl/BiblioFetch.jl/docs/build/examples/papers/.metadata/run.log", nothing, nothing, 1, false, [:unpaywall, :arxiv, :direct], false, :lenient, :pending, false, false, 1, 50, FetchEntry[FetchEntry("arxiv:1008.3477", "", "arxiv:1008.3477", 0, "", :pending, :none, nothing, AttemptLog[], nothing, nothing)], Tuple{String, String, String}[], String[])

job.refs is the parsed reference list — confirm it before hitting the network.

length(job.refs), first(job.refs).key

(1, "arxiv:1008.3477")

Uncomment the next line to actually fetch. The doc build skips it so CI doesn't hit the real APS archive.

# result = BiblioFetch.run(job_path)

What's next

Once minimal-job.toml succeeds, the next two examples show how to grow the job: citation-graph expansion and publisher-TDM sources.