nigig-org/crates/apps/pdf/pdf-document/tests/transparency.rs

//! Transparency acceptance tests.
//!
//! The merge criteria from `REVIEWS/adr/0009-pdf-transparency.md`.
//!
//! The headline defect these guard against is not "transparency was
//! missing" — it is that `gs` **parsed as `CloseStroke`**, so every page
//! that set a graphics state gained a stroked path the document never
//! asked for, and lost its alpha, blend mode and soft mask silently.

use std::collections::HashMap;
use std::path::PathBuf;

use nigig_pdf_cos::PdfObj;
use nigig_pdf_document::PdfDocument;
use nigig_pdf_graphics::colorspace::resolve_page_color_spaces;
use nigig_pdf_graphics::content::{interpret_ops, parse_content_stream};
use nigig_pdf_graphics::recording::{RecordingDevice, RenderCommand};
use nigig_pdf_graphics::transparency::{
    resolve_page_ext_gstates, BlendMode, SoftMaskKind, TransparencyError,
};

fn corpus(relative: &str) -> Vec<u8> {
    let path = PathBuf::from(env!("CARGO_MANIFEST_DIR"))
        .join("../tests/corpus")
        .join(relative);
    std::fs::read(&path).unwrap_or_else(|e| panic!("missing {}: {e}", path.display()))
}

struct Rendered {
    commands: Vec<RenderCommand>,
    errors: Vec<TransparencyError>,
}

/// Interpret page 0 the way the render worker does.
fn render(relative: &str) -> Rendered {
    let bytes = corpus(relative);
    let mut doc = PdfDocument::parse(&bytes).expect("fixture should parse");
    let page = doc.page(0).expect("page 0");
    render_page(&page.content_data, &page.color_spaces, &page.ext_gstates)
}

fn render_page(
    content: &[u8],
    color_spaces: &HashMap<String, PdfObj>,
    ext_gstates: &HashMap<String, PdfObj>,
) -> Rendered {
    let mut device = RecordingDevice::new();
    for (name, space) in resolve_page_color_spaces(color_spaces) {
        if let Ok(space) = space {
            device.register_color_space(&name, space);
        }
    }
    let (states, mut errors) = resolve_page_ext_gstates(ext_gstates);
    for (name, state) in states {
        device.register_ext_gstate(&name, state);
    }
    let ops = parse_content_stream(content).expect("content should parse");
    interpret_ops(&ops, &mut device).expect("interpretation should not fail");
    errors.extend(device.transparency_errors().iter().cloned());
    Rendered {
        commands: device.into_commands(),
        errors,
    }
}

fn fill_alphas(cmds: &[RenderCommand]) -> Vec<f64> {
    cmds.iter()
        .filter_map(|c| match c {
            RenderCommand::SetFillAlpha(a) => Some(*a),
            _ => None,
        })
        .collect()
}

#[test]
fn gs_no_longer_injects_a_spurious_stroke() {
    let r = render("transparency/constant_alpha.pdf");
    assert!(
        !r.commands
            .iter()
            .any(|c| matches!(c, RenderCommand::Stroke)),
        "the page strokes nothing, but a stroke reached the device: {:?}",
        r.commands
    );
}

#[test]
fn constant_alpha_reaches_the_device() {
    let r = render("transparency/constant_alpha.pdf");
    let alphas = fill_alphas(&r.commands);
    assert!(
        alphas.contains(&0.5),
        "ca 0.5 never reached the device: {alphas:?}"
    );
    assert!(
        alphas.contains(&0.1),
        "the second ca never reached the device: {alphas:?}"
    );
    let stroke: Vec<f64> = r
        .commands
        .iter()
        .filter_map(|c| match c {
            RenderCommand::SetStrokeAlpha(a) => Some(*a),
            _ => None,
        })
        .collect();
    assert!(stroke.contains(&0.25), "CA 0.25 was lost: {stroke:?}");
}

#[test]
fn an_extgstate_only_applies_the_entries_it_names() {
    // /GS1 sets ca but not CA; it must not reset the stroke alpha.
    let r = render("transparency/constant_alpha.pdf");
    let stroke_alphas: Vec<f64> = r
        .commands
        .iter()
        .filter_map(|c| match c {
            RenderCommand::SetStrokeAlpha(a) => Some(*a),
            _ => None,
        })
        .collect();
    assert_eq!(
        stroke_alphas.len(),
        1,
        "only /GS0 names CA, so only one stroke alpha may be emitted: {stroke_alphas:?}"
    );
}

#[test]
fn blend_modes_reach_the_device() {
    let r = render("transparency/blend_modes.pdf");
    let modes: Vec<BlendMode> = r
        .commands
        .iter()
        .filter_map(|c| match c {
            RenderCommand::SetBlendMode(m) => Some(*m),
            _ => None,
        })
        .collect();
    for expected in [
        BlendMode::Multiply,
        BlendMode::Screen,
        BlendMode::Overlay,
        BlendMode::Darken,
        BlendMode::Lighten,
        BlendMode::ColorDodge,
        BlendMode::ColorBurn,
        BlendMode::HardLight,
        BlendMode::SoftLight,
        BlendMode::Difference,
        BlendMode::Exclusion,
        BlendMode::Luminosity,
    ] {
        assert!(
            modes.contains(&expected),
            "{expected:?} never reached the device: {modes:?}"
        );
    }
    assert!(r.errors.is_empty(), "unexpected errors: {:?}", r.errors);
}

#[test]
fn an_unknown_blend_mode_falls_back_to_normal_and_is_reported() {
    let r = render("transparency/unknown_blend_mode.pdf");
    let modes: Vec<BlendMode> = r
        .commands
        .iter()
        .filter_map(|c| match c {
            RenderCommand::SetBlendMode(m) => Some(*m),
            _ => None,
        })
        .collect();
    assert_eq!(
        modes,
        vec![BlendMode::Normal],
        "§11.3.5.2 requires an unknown mode to render as Normal"
    );
    assert!(
        r.errors
            .iter()
            .any(|e| matches!(e, TransparencyError::UnknownBlendMode(n) if n == "Frobnicate")),
        "the unknown mode must be named: {:?}",
        r.errors
    );
}

#[test]
fn a_soft_mask_reaches_the_device_with_its_transfer_function() {
    let r = render("transparency/soft_mask.pdf");
    let mask = r
        .commands
        .iter()
        .find_map(|c| match c {
            RenderCommand::SetSoftMask(m) => Some(m.clone()),
            _ => None,
        })
        .expect("the soft mask never reached the device");
    assert_eq!(mask.kind, SoftMaskKind::Luminosity);
    assert!(mask.group.is_some(), "/G must be carried through");
    assert_eq!(mask.backdrop, Some(vec![0.0]), "/BC must be carried");
    // /TR squares its input, so a 0.5 mask value becomes 0.25. This is the
    // ADR 0006 function evaluator being used, which is why 0006 was the
    // prerequisite for this ADR.
    assert!(
        (mask.apply_transfer(0.5) - 0.25).abs() < 1e-9,
        "the transfer function was not applied: {}",
        mask.apply_transfer(0.5)
    );
}

#[test]
fn smask_none_clears_the_mask_rather_than_being_ignored() {
    let r = render("transparency/smask_none.pdf");
    let installed = r
        .commands
        .iter()
        .position(|c| matches!(c, RenderCommand::SetSoftMask(_)));
    let cleared = r
        .commands
        .iter()
        .position(|c| matches!(c, RenderCommand::ClearSoftMask));
    let installed = installed.expect("the first gs installs a mask");
    let cleared = cleared.expect("/SMask /None must emit a clear");
    assert!(
        cleared > installed,
        "the clear must come after the mask it clears"
    );
}

#[test]
fn a_missing_ext_gstate_is_reported_not_ignored() {
    let r = render("transparency/missing_gstate.pdf");
    assert!(
        r.errors
            .iter()
            .any(|e| matches!(e, TransparencyError::UnresolvedExtGState(n) if n == "Nope")),
        "an unresolvable /ExtGState must be named: {:?}",
        r.errors
    );
    // And the page still renders its fill rather than being abandoned.
    assert!(
        r.commands
            .iter()
            .any(|c| matches!(c, RenderCommand::FillWinding)),
        "one bad gs must not cost the page its content"
    );
}

#[test]
fn a_transparency_group_is_parsed_from_a_form_xobject() {
    use nigig_pdf_graphics::transparency::TransparencyGroup;

    let bytes = corpus("transparency/group.pdf");
    let mut doc = PdfDocument::parse(&bytes).expect("parses");
    let page = doc.page(0).expect("page 0");
    let form = page.xobjects.get("Fm0").expect("the form xobject");
    let obj = doc.resolve_ref(form.obj_ref).expect("resolves");
    let dict = obj.as_stream().map(|s| s.dict.clone()).expect("a stream");
    let group_obj = dict.get("Group").expect("/Group");
    let group_dict = group_obj.as_dict().expect("a dict");

    let mut resolve = |_: &PdfObj| -> Option<(PdfObj, Option<Vec<u8>>)> { None };
    let mut lookup = |_: &str| -> Option<PdfObj> { None };
    let group = TransparencyGroup::parse(group_dict, &mut resolve, &mut lookup).expect("a group");
    assert!(group.isolated, "/I true");
    assert!(group.knockout, "/K true");
    assert!(group.color_space.is_some(), "/CS DeviceRGB");
}

#[test]
fn page_ext_gstates_are_resolved_and_decoded_by_the_document_layer() {
    // The transfer function is an indirect object. If the document layer
    // did not dereference it, /TR would silently be dropped.
    let bytes = corpus("transparency/soft_mask.pdf");
    let mut doc = PdfDocument::parse(&bytes).expect("parses");
    let page = doc.page(0).expect("page 0");
    let gs = page
        .ext_gstates
        .get("GS0")
        .expect("/GS0 should be in the page resources");
    let smask = gs
        .as_dict()
        .and_then(|d| d.get("SMask"))
        .and_then(|s| s.as_dict())
        .expect("/SMask should be a resolved dict");
    assert!(
        smask.get("TR").and_then(|t| t.as_dict()).is_some(),
        "/TR must be dereferenced, not left as a reference"
    );
}

#[test]
fn every_transparency_fixture_parses_without_panicking() {
    for name in [
        "constant_alpha.pdf",
        "blend_modes.pdf",
        "soft_mask.pdf",
        "smask_none.pdf",
        "group.pdf",
        "unknown_blend_mode.pdf",
        "missing_gstate.pdf",
    ] {
        let bytes = corpus(&format!("transparency/{name}"));
        let mut doc = PdfDocument::parse(&bytes)
            .unwrap_or_else(|e| panic!("transparency/{name} should parse: {e}"));
        let page = doc
            .page(0)
            .unwrap_or_else(|e| panic!("transparency/{name} page 0: {e}"));
        let _ = render_page(&page.content_data, &page.color_spaces, &page.ext_gstates);
    }
}

/// `PdfPage::xobjects` was empty for every page of every document:
/// `extract_xobjects` called `doc.resolve()`, which *follows* the
/// reference, and then asked the resolved object for its `as_ref()` — which
/// is always `None`. Nothing downstream could resolve a `Do` operator.
///
/// This is the same defect that once destroyed annotation object
/// references, so it gets its own regression test rather than being left
/// implicit in the transparency-group test that found it.
#[test]
fn page_xobjects_carry_their_object_reference() {
    for (fixture, name, subtype) in [
        ("transparency/group.pdf", "Fm0", "Form"),
        ("images/xobject.pdf", "Im1", "Image"),
    ] {
        let bytes = corpus(fixture);
        let mut doc = PdfDocument::parse(&bytes).expect("parses");
        let page = doc.page(0).expect("page 0");
        let entry = page
            .xobjects
            .get(name)
            .unwrap_or_else(|| panic!("{fixture}: /{name} missing from the xobject map"));
        assert_eq!(entry.subtype, subtype, "{fixture}: wrong subtype");
        // The reference must actually resolve.
        assert!(
            doc.resolve_ref(entry.obj_ref).is_ok(),
            "{fixture}: /{name} has an unresolvable reference"
        );
    }
}