searchi/crates/searchi-extract-worker/tests/parent_roundtrip.rs

//! Exercise the real helper through the shared engine's process owner.
use searchi_core::{
    extraction_process::{run, ProcessConfig, ProcessError},
    extraction_protocol::{DocumentFormat, ExtractionOutcome, RequestHeader},
};
use std::{
    io::{Cursor, Write},
    path::PathBuf,
    sync::atomic::AtomicBool,
    time::Duration,
};

fn config() -> ProcessConfig {
    ProcessConfig {
        executable: PathBuf::from(env!("CARGO_BIN_EXE_searchi-extract-worker")),
        timeout: Duration::from_secs(10),
    }
}

#[test]
fn actual_parent_and_worker_return_original_docx_passage_and_digest() {
    let mut document = zip::ZipWriter::new(Cursor::new(Vec::new()));
    document
        .start_file(
            "word/document.xml",
            zip::write::SimpleFileOptions::default(),
        )
        .unwrap();
    document.write_all(br#"<w:document xmlns:w="http://schemas.openxmlformats.org/wordprocessingml/2006/main"><w:body><w:p><w:r><w:t>Cancel with sixty days written notice.</w:t></w:r></w:p></w:body></w:document>"#).unwrap();
    let input = document.finish().unwrap().into_inner();
    let request = RequestHeader::from_bytes(DocumentFormat::Docx, &input).unwrap();
    let expected_hash = request.input_sha256.clone();
    // Exercise short-lived native worker exit/task-inspection races too.
    for _ in 0..10 {
        let response = run(
            &config(),
            request.clone(),
            input.clone(),
            &AtomicBool::new(false),
        )
        .unwrap();
        assert_eq!(response.input_sha256, expected_hash);
        let ExtractionOutcome::Extracted { passages } = response.outcome else {
            panic!("Valid document rejected");
        };
        assert_eq!(passages.len(), 1);
        assert_eq!(passages[0].text, "Cancel with sixty days written notice.");
        assert_eq!(passages[0].source.kind, "paragraph");
        assert_eq!(passages[0].source.number, 1);
    }
}

#[test]
fn parser_rejection_and_cancellation_remain_distinct() {
    let input = b"invalid PDF".to_vec();
    let request = RequestHeader::from_bytes(DocumentFormat::Pdf, &input).unwrap();
    let response = run(
        &config(),
        request.clone(),
        input.clone(),
        &AtomicBool::new(false),
    )
    .unwrap();
    assert!(matches!(
        response.outcome,
        ExtractionOutcome::Rejected { .. }
    ));
    assert!(matches!(
        run(&config(), request, input, &AtomicBool::new(true)),
        Err(ProcessError::Cancelled)
    ));
}