oxml_pdf/
conformance.rs

1//! PDF/A profile selection, preflight, and deterministic metadata.
2
3use std::error::Error;
4use std::fmt;
5
6use oxml_layout::{Color, LayoutResult, Paint, PositionedElement};
7
8use crate::font;
9use crate::structure::valid_structure;
10
11pub(crate) const SRGB2014: &[u8] = include_bytes!("../assets/sRGB2014.icc");
12pub(crate) const FILE_IDENTIFIER: &[u8; 16] = b"rdocx-pdfa-00001";
13
14/// An archival conformance level supported by the PDF backend.
15#[derive(Debug, Clone, Copy, PartialEq, Eq)]
16pub enum PdfConformance {
17    /// ISO 19005-2, level B.
18    PdfA2b,
19    /// ISO 19005-3, level B.
20    PdfA3b,
21}
22
23impl PdfConformance {
24    pub(crate) const fn part(self) -> u8 {
25        match self {
26            Self::PdfA2b => 2,
27            Self::PdfA3b => 3,
28        }
29    }
30}
31
32/// Options for the explicit archival PDF renderer.
33#[derive(Debug, Clone, Copy, PartialEq, Eq)]
34pub struct PdfOptions {
35    /// The required archival profile.
36    pub profile: PdfConformance,
37}
38
39impl PdfOptions {
40    /// Select one supported archival profile.
41    pub const fn new(profile: PdfConformance) -> Self {
42        Self { profile }
43    }
44}
45
46/// A reason an input cannot be represented by the requested PDF/A profile.
47#[derive(Debug, Clone, PartialEq, Eq)]
48#[non_exhaustive]
49pub enum PdfError {
50    /// A shaped run refers to a font that cannot be embedded and subset.
51    MissingEmbeddedFont { font_id: u32 },
52    /// The logical structure tree or one of its marked-content references is invalid.
53    InvalidStructure,
54    /// A link requests an action that the archival path refuses.
55    ForbiddenAction { action: String },
56    /// A color component is non-finite or outside its declared range.
57    UnsupportedColor,
58    /// A paint kind has no archival PDF representation.
59    UnsupportedPaint { paint: &'static str },
60}
61
62impl fmt::Display for PdfError {
63    fn fmt(&self, formatter: &mut fmt::Formatter<'_>) -> fmt::Result {
64        match self {
65            Self::MissingEmbeddedFont { font_id } => {
66                write!(formatter, "font {font_id} cannot be embedded for PDF/A")
67            }
68            Self::InvalidStructure => formatter.write_str("invalid tagged PDF structure"),
69            Self::ForbiddenAction { action } => {
70                write!(formatter, "action is forbidden in PDF/A output: {action}")
71            }
72            Self::UnsupportedColor => formatter.write_str("unsupported PDF/A color state"),
73            Self::UnsupportedPaint { paint } => {
74                write!(formatter, "unsupported PDF/A paint: {paint}")
75            }
76        }
77    }
78}
79
80impl Error for PdfError {}
81
82pub(crate) fn preflight(layout: &LayoutResult) -> Result<(), PdfError> {
83    if let Some(structure) = &layout.structure
84        && !valid_structure(structure, &layout.pages)
85    {
86        return Err(PdfError::InvalidStructure);
87    }
88
89    let mut usage = font::collect_glyph_usage(layout);
90    let fonts = layout
91        .fonts
92        .iter()
93        .map(|font| (font.id, font))
94        .collect::<std::collections::HashMap<_, _>>();
95    for (font_id, font_usage) in &mut usage {
96        let Some(font_data) = fonts.get(font_id) else {
97            return Err(PdfError::MissingEmbeddedFont { font_id: font_id.0 });
98        };
99        if font::prepare_font(font_data, font_usage).is_none() {
100            return Err(PdfError::MissingEmbeddedFont { font_id: font_id.0 });
101        }
102    }
103
104    for page in &layout.pages {
105        if let Some(background) = &page.background {
106            validate_paint(background)?;
107        }
108        validate_elements(&page.elements)?;
109    }
110    Ok(())
111}
112
113fn validate_elements(elements: &[PositionedElement]) -> Result<(), PdfError> {
114    for element in elements {
115        match element {
116            PositionedElement::Text(run) => validate_color(run.color)?,
117            PositionedElement::MultilingualText(run) => validate_color(run.color)?,
118            PositionedElement::Line { color, .. } | PositionedElement::FilledRect { color, .. } => {
119                validate_color(*color)?
120            }
121            PositionedElement::LinkAnnotation { url, .. } => {
122                let scheme = url
123                    .split_once(':')
124                    .map(|(scheme, _)| scheme.to_ascii_lowercase());
125                if scheme
126                    .as_deref()
127                    .is_some_and(|scheme| !matches!(scheme, "http" | "https" | "mailto" | "tel"))
128                {
129                    return Err(PdfError::ForbiddenAction {
130                        action: url.clone(),
131                    });
132                }
133            }
134            PositionedElement::Path(path) => {
135                if let Some(fill) = &path.fill {
136                    validate_paint(fill)?;
137                }
138                if let Some(stroke) = &path.stroke {
139                    validate_paint(&stroke.paint)?;
140                }
141            }
142            PositionedElement::Group(group) => {
143                if !unit_component(group.opacity) {
144                    return Err(PdfError::UnsupportedColor);
145                }
146                for effect in &group.effects {
147                    match effect {
148                        oxml_layout::Effect::OuterShadow { color, .. } => validate_color(*color)?,
149                        _ => return Err(PdfError::UnsupportedColor),
150                    }
151                }
152                validate_elements(&group.children)?;
153            }
154            PositionedElement::MarkedContent { children, .. } => validate_elements(children)?,
155            PositionedElement::Image { .. } => {}
156            _ => return Err(PdfError::UnsupportedColor),
157        }
158    }
159    Ok(())
160}
161
162fn validate_paint(paint: &Paint) -> Result<(), PdfError> {
163    match paint {
164        Paint::Solid(color) => validate_color(*color),
165        Paint::Linear { stops, .. } | Paint::Radial { stops, .. } => {
166            for stop in stops {
167                if !unit_component(stop.offset) {
168                    return Err(PdfError::UnsupportedColor);
169                }
170                validate_color(stop.color)?;
171            }
172            Ok(())
173        }
174        Paint::Tile { .. } => Err(PdfError::UnsupportedPaint { paint: "tile" }),
175    }
176}
177
178fn validate_color(color: Color) -> Result<(), PdfError> {
179    if [color.r, color.g, color.b, color.a]
180        .into_iter()
181        .all(unit_component)
182    {
183        Ok(())
184    } else {
185        Err(PdfError::UnsupportedColor)
186    }
187}
188
189fn unit_component(component: f64) -> bool {
190    component.is_finite() && (0.0..=1.0).contains(&component)
191}
192
193pub(crate) fn archival_xmp(
194    layout: &LayoutResult,
195    profile: PdfConformance,
196    declares_pdfua: bool,
197) -> Vec<u8> {
198    let metadata = layout.metadata.as_ref();
199    let title = xml_escape(
200        metadata
201            .and_then(|value| value.title.as_deref())
202            .unwrap_or("Untitled document"),
203    );
204    let author = metadata
205        .and_then(|value| value.author.as_deref())
206        .map(xml_escape);
207    let subject = metadata
208        .and_then(|value| value.subject.as_deref())
209        .map(xml_escape);
210    let keywords = metadata
211        .and_then(|value| value.keywords.as_deref())
212        .map(xml_escape);
213    let creator = xml_escape(
214        metadata
215            .and_then(|value| value.creator.as_deref())
216            .unwrap_or("rdocx-pdf"),
217    );
218    let mut properties = format!(
219        "<dc:title><rdf:Alt><rdf:li xml:lang=\"x-default\">{title}</rdf:li></rdf:Alt></dc:title>\n"
220    );
221    if let Some(author) = author {
222        properties.push_str(&format!(
223            "<dc:creator><rdf:Seq><rdf:li>{author}</rdf:li></rdf:Seq></dc:creator>\n"
224        ));
225    }
226    if let Some(subject) = subject {
227        properties.push_str(&format!(
228            "<dc:description><rdf:Alt><rdf:li xml:lang=\"x-default\">{subject}</rdf:li></rdf:Alt></dc:description>\n"
229        ));
230    }
231    if let Some(keywords) = keywords {
232        properties.push_str(&format!("<pdf:Keywords>{keywords}</pdf:Keywords>\n"));
233    }
234    properties.push_str(&format!(
235        "<xmp:CreatorTool>{creator}</xmp:CreatorTool>\n<pdf:Producer>rdocx-pdf</pdf:Producer>\n"
236    ));
237    properties.push_str(&format!(
238        "<pdfaid:part>{}</pdfaid:part>\n<pdfaid:conformance>B</pdfaid:conformance>\n",
239        profile.part()
240    ));
241    if declares_pdfua {
242        properties.push_str("<pdfuaid:part>1</pdfuaid:part>\n");
243    }
244    properties.push_str(
245        "<xmpMM:DocumentID>uuid:72646f63-782d-7064-6661-2d3030303031</xmpMM:DocumentID>\n<xmpMM:InstanceID>uuid:72646f63-782d-7064-6661-2d3030303031</xmpMM:InstanceID>",
246    );
247    let extension = if declares_pdfua {
248        r#"
249<rdf:Description rdf:about="" xmlns:pdfaExtension="http://www.aiim.org/pdfa/ns/extension/" xmlns:pdfaSchema="http://www.aiim.org/pdfa/ns/schema#" xmlns:pdfaProperty="http://www.aiim.org/pdfa/ns/property#">
250<pdfaExtension:schemas><rdf:Bag><rdf:li rdf:parseType="Resource">
251<pdfaSchema:schema>PDF/UA identification schema</pdfaSchema:schema>
252<pdfaSchema:namespaceURI>http://www.aiim.org/pdfua/ns/id/</pdfaSchema:namespaceURI>
253<pdfaSchema:prefix>pdfuaid</pdfaSchema:prefix>
254<pdfaSchema:property><rdf:Seq><rdf:li rdf:parseType="Resource">
255<pdfaProperty:name>part</pdfaProperty:name>
256<pdfaProperty:valueType>Integer</pdfaProperty:valueType>
257<pdfaProperty:category>internal</pdfaProperty:category>
258<pdfaProperty:description>PDF/UA version identifier</pdfaProperty:description>
259</rdf:li></rdf:Seq></pdfaSchema:property>
260</rdf:li></rdf:Bag></pdfaExtension:schemas>
261</rdf:Description>"#
262    } else {
263        ""
264    };
265    format!(
266        r#"<?xpacket begin="" id="W5M0MpCehiHzreSzNTczkc9d"?>
267<x:xmpmeta xmlns:x="adobe:ns:meta/">
268<rdf:RDF xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#">
269<rdf:Description rdf:about="" xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:xmp="http://ns.adobe.com/xap/1.0/" xmlns:pdf="http://ns.adobe.com/pdf/1.3/" xmlns:xmpMM="http://ns.adobe.com/xap/1.0/mm/" xmlns:pdfaid="http://www.aiim.org/pdfa/ns/id/" xmlns:pdfuaid="http://www.aiim.org/pdfua/ns/id/">
270{properties}
271</rdf:Description>
272{extension}
273</rdf:RDF>
274</x:xmpmeta>
275<?xpacket end="w"?>"#
276    )
277    .into_bytes()
278}
279
280fn xml_escape(value: &str) -> String {
281    value
282        .replace('&', "&amp;")
283        .replace('<', "&lt;")
284        .replace('>', "&gt;")
285        .replace('"', "&quot;")
286        .replace('\'', "&apos;")
287}
oxml_pdf/conformance.rs

oxml_pdf/
conformance.rs
