tonhowtf/omniget · error

nenhum .tex com \begin

Error message

nenhum .tex com \begin{{document}}

What it means

The tar had at least one .tex file, but pick_main could not determine a main document — none of the .tex texts contains \begin{document}, which any compilable LaTeX root document must have. extract_source therefore fails with this message.

Solutions

1. Extend pick_main to also recognize ConTeXt (\starttext) and plain TeX markers.
2. Check the tar contents — maybe the root .tex has an unexpected name or extension and wasn't collected.
3. Fall back to the largest .tex file as the main document when no \begin{document} is found.
4. Skip such papers if compilable LaTeX is required.

Example fix

// before
let main_key = pick_main(&texts).ok_or_else(|| anyhow!("nenhum .tex com \\begin{{document}}"))?;
// after
let main_key = pick_main(&texts)
    .or_else(|| texts.iter().max_by_key(|(_, t)| t.len()).map(|(k, _)| k.clone()))
    .ok_or_else(|| anyhow!("nenhum .tex com \\begin{{document}}"))?;

// heuristica propria antes da chamada:
fn parece_main(t: &str) -> bool {
    t.contains("\\begin{document}") || t.contains("\\starttext")
}

match extract_source(&bytes) {
    Err(e) if e.to_string().contains("begin{document}") => {
        // fallback: maior .tex como raiz
        eprintln!("sem raiz LaTeX clara; usando maior .tex como main");
    }
    other => other,
}

When it happens

Trigger: extract_source on a source bundle whose .tex files are all include/fragment files (no root), or files in non-LaTeX TeX dialects (plain TeX, ConTeXt) lacking \begin{document}.

Common situations: Plain TeX or ConTeXt submissions (different document markers); bundles where the root file uses a nonstandard class loaded via unusual preamble; fragmented multi-file papers whose root was excluded by the collector.

Related errors

• o source deste arXiv e um PDF, nao tem LaTeX
• source do arXiv nao parece LaTeX
• o tar do arXiv nao tem nenhum .tex
• download terminou mas nao achei o arquivo
• resposta do arXiv sem

AI-assisted analysis of tonhowtf/omniget@8600b91f42 (2026-09-12). Data as JSON: /api/errors/27bd3f455a9e39d1. Report an issue: GitHub.

Appendix: source

Thrown at src-tauri/omniget-core/src/core/tools/arxiv.rs:403

    for entry in archive.entries()? {
        let mut entry = entry?;
        let path = entry.path()?.to_string_lossy().to_string();
        let lower = path.to_lowercase();
        if !(lower.ends_with(".tex") || lower.ends_with(".bbl") || lower.ends_with(".ltx")) {
            continue;
        }
        let mut buf = Vec::new();
        use std::io::Read;
        entry.read_to_end(&mut buf)?;
        names.push(path.clone());
        texts.insert(path, String::from_utf8_lossy(&buf).to_string());
    }
    if texts.is_empty() {
        return Err(anyhow!("o tar do arXiv nao tem nenhum .tex"));
    }

    let main_key =
        pick_main(&texts).ok_or_else(|| anyhow!("nenhum .tex com \\begin{{document}}"))?;
    let main = texts.get(&main_key).cloned().unwrap_or_default();
    let main = resolve_inputs(&main, &texts, 0);
    names.sort();
    Ok(SourceBundle { main, files: names })
}

fn pick_main(texts: &HashMap<String, String>) -> Option<String> {
    let mut best: Option<(u32, usize, String)> = None;
    for (k, v) in texts {
        if !k.to_lowercase().ends_with(".tex") && !k.to_lowercase().ends_with(".ltx") {
            continue;
        }
        let mut score = 0u32;
        if v.contains("\\begin{document}") {
            score += 4;
        }
        if v.contains("\\documentclass") {
            score += 2;

View on GitHub (pinned to 8600b91f42)