Merge pull request #1173 from dylanmtaylor/codex/file-formats-api

Add shared file format detection

Cargo.lock

@@ -1454,6 +1454,18 @@ dependencies = [
  "thiserror 2.0.21",
 ]
 
+[[package]]
+name = "craft-file-formats"
+version = "0.1.0"
+dependencies = [
+ "craft-eps",
+ "craft-idml",
+ "craft-svg",
+ "craft-zip",
+ "serde",
+ "serde_json",
+]
+
 [[package]]
 name = "craft-flags"
 version = "0.1.0"

Cargo.toml

@@ -42,6 +42,7 @@ craft-docx = { path = "crates/docx" }
 craft-dxf = { path = "crates/dxf" }
 craft-drawingml = { path = "crates/drawingml" }
 craft-eps = { path = "crates/eps" }
+craft-file-formats = { path = "crates/file-formats" }
 craft-doc = { path = "crates/doc" }
 craft-docpath = { path = "crates/docpath" }
 craft-dpx = { path = "crates/dpx" }

crates/file-formats/Cargo.toml

@@ -0,0 +1,19 @@
+[package]
+name = "craft-file-formats"
+description = "Shared bounded file signature and extension resolution for Craft applications"
+version.workspace = true
+edition.workspace = true
+license.workspace = true
+repository.workspace = true
+rust-version.workspace = true
+
+[dependencies]
+craft-eps = { workspace = true }
+craft-idml = { workspace = true }
+craft-svg = { workspace = true }
+craft-zip = { workspace = true }
+serde = { workspace = true }
+serde_json = { workspace = true }
+
+[lints]
+workspace = true

crates/file-formats/src/lib.rs

@@ -0,0 +1,269 @@
+//! Bounded file kind detection and extension resolution shared by Craft applications.
+//!
+//! This crate identifies a format; it does not parse, validate, or decode the file.
+//! Callers remain responsible for choosing which identified formats their product supports.
+#![deny(clippy::unwrap_used, clippy::expect_used, clippy::panic, clippy::unimplemented, clippy::todo, clippy::unreachable)]
+#![forbid(unsafe_code)]
+
+/// A supported file family, independent of the application model that consumes it.
+#[derive(Debug, Clone, Copy, PartialEq, Eq)]
+pub enum Format {
+    Pdf,
+    Eps,
+    PostScript,
+    Illustrator,
+    Dxf,
+    Svg,
+    /// Gzip-compressed SVG. Kept distinct so writers do not place plain SVG under `.svgz`.
+    Svgz,
+    Png,
+    Jpeg,
+    Bmp,
+    Gif,
+    Tiff,
+    WebP,
+    Heif,
+    Ase,
+    Psd,
+    Affinity,
+    Ora,
+    Kra,
+    Xcf,
+    Pdn,
+    Raw,
+    Pcraft,
+    Idml,
+    DesignCraft,
+    /// Legacy DesignCraft documents serialized as one JSON object.
+    DesignCraftJson,
+    Indd,
+    Docx,
+    /// Legacy OLE Compound File Word document; identified by its extension only.
+    LegacyDoc,
+    Rtf,
+    Xlsx,
+    TaggedText,
+    Pptx,
+    PlainText,
+    Csv,
+    Tsv,
+    /// Legacy binary Excel workbooks, identified by extension only.
+    LegacyExcel,
+    Gcode,
+    Studio,
+    StencilCraft,
+}
+
+/// Resolve a name and file bytes to a format. Strong signatures take precedence over
+/// extension hints. Only format-specific bounded probes and extension tables live here.
+pub fn detect(name: &str, bytes: &[u8]) -> Option<Format> {
+    if craft_idml::is_idml(bytes) {
+        return Some(Format::Idml);
+    }
+    if craft_zip::stored_mimetype(bytes) == Some(b"application/vnd.designcraft+zip") || designcraft_zip_has_document(bytes) {
+        return Some(Format::DesignCraft);
+    }
+    if is_designcraft_json(bytes) {
+        return Some(Format::DesignCraftJson);
+    }
+    if bytes.starts_with(&INDD_SIGNATURE) {
+        return Some(Format::Indd);
+    }
+    if let Some(found) = craft_eps::sniff(bytes) {
+        return Some(match (found.kind, found.illustrator) {
+            (_, true) => Format::Illustrator,
+            (craft_eps::Kind::Pdf, _) => Format::Pdf,
+            (craft_eps::Kind::Eps, _) => Format::Eps,
+            (craft_eps::Kind::PostScript, _) => Format::PostScript,
+        });
+    }
+    if bytes.starts_with(b"8BPS") {
+        return Some(Format::Psd);
+    }
+    if bytes.starts_with(b"ASEF") {
+        return Some(Format::Ase);
+    }
+    if bytes.starts_with(b"PDN3") {
+        return Some(Format::Pdn);
+    }
+    if bytes.starts_with(b"gimp xcf ") || gzipped_xcf(name, bytes) {
+        return Some(Format::Xcf);
+    }
+    if bytes.starts_with(b"\0\xffKA") {
+        return Some(Format::Affinity);
+    }
+    if let Some(raster) = raster_signature(bytes) {
+        return Some(raster);
+    }
+    if bytes.starts_with(b"silhouette04;") || bytes.starts_with(b"silhouette05;") {
+        return Some(Format::Studio);
+    }
+    let tagged = bytes.strip_prefix(b"\xef\xbb\xbf").unwrap_or(bytes).trim_ascii_start();
+    if tagged.starts_with(b"<ASCII-") || tagged.starts_with(b"<UNICODE-") || tagged.starts_with(b"<ANSI-") {
+        return Some(Format::TaggedText);
+    }
+    if bytes.trim_ascii_start().starts_with(b"{\\rtf") {
+        return Some(Format::Rtf);
+    }
+    if craft_svg::sniff::is_svgz(bytes)
+        && let Ok(text) = craft_svg::svgz::text(bytes, 64 * 1024 * 1024)
+        && craft_svg::sniff::is_svg(text.as_bytes())
+    {
+        return Some(Format::Svgz);
+    }
+    if craft_svg::sniff::is_svg(bytes) {
+        return Some(Format::Svg);
+    }
+    if let Some(ext) = extension(name)
+        && let Some(format) = from_extension(ext)
+    {
+        return Some(format);
+    }
+    let head = bytes.get(..4096).unwrap_or(bytes);
+    if dxf_ascii_prefix(head) || head.starts_with(b"AutoCAD Binary DXF\r\n\x1a\0") {
+        return Some(Format::Dxf);
+    }
+    None
+}
+
+/// DesignCraft's older ZIPs predate their leading `mimetype`; they are identified by the
+/// `document.json` member. The directory is bounded before it is indexed and no entry is inflated.
+fn designcraft_zip_has_document(bytes: &[u8]) -> bool {
+    if !bytes.starts_with(b"PK") {
+        return false;
+    }
+    craft_zip::Archive::open(bytes, craft_zip::Limits { max_entries: 100_000, max_total_bytes: 0 })
+        .is_ok_and(|archive| archive.contains("document.json"))
+}
+
+/// Read only a bounded prefix, stopping as soon as the model's required top-level keys have
+/// appeared. This distinguishes legacy DesignCraft JSON from arbitrary JSON without decoding the
+/// document or walking its arrays and assets.
+fn is_designcraft_json(bytes: &[u8]) -> bool {
+    const LIMIT: usize = 1 << 20;
+    const REQUIRED: u16 = (1 << 10) - 1;
+    const KEYS: [&str; 10] = ["title", "settings", "spreads", "parents", "layers", "stories", "styles", "swatches", "sections", "next_id"];
+
+    struct Probe;
+    impl<'de> serde::de::Visitor<'de> for Probe {
+        type Value = bool;
+
+        fn expecting(&self, formatter: &mut std::fmt::Formatter<'_>) -> std::fmt::Result {
+            formatter.write_str("a DesignCraft document object")
+        }
+
+        fn visit_map<A: serde::de::MapAccess<'de>>(self, mut map: A) -> Result<Self::Value, A::Error> {
+            let mut found = 0u16;
+            while let Some(key) = map.next_key::<String>()? {
+                if let Some(index) = KEYS.iter().position(|candidate| *candidate == key) {
+                    found |= 1 << index;
+                }
+                map.next_value::<serde::de::IgnoredAny>()?;
+                if found == REQUIRED {
+                    return Ok(true);
+                }
+            }
+            Ok(false)
+        }
+    }
+
+    use serde::de::DeserializeSeed;
+    struct Seed;
+    impl<'de> DeserializeSeed<'de> for Seed {
+        type Value = bool;
+        fn deserialize<D: serde::Deserializer<'de>>(self, deserializer: D) -> Result<Self::Value, D::Error> {
+            deserializer.deserialize_map(Probe)
+        }
+    }
+
+    let prefix = bytes.get(..bytes.len().min(LIMIT)).unwrap_or(bytes);
+    let mut deserializer = serde_json::Deserializer::from_slice(prefix);
+    Seed.deserialize(&mut deserializer).unwrap_or(false)
+}
+
+/// Resolve a file extension to a family, case-insensitively.
+pub fn from_extension(ext: &str) -> Option<Format> {
+    match ext.trim_start_matches('.').to_ascii_lowercase().as_str() {
+        "pdf" => Some(Format::Pdf),
+        "eps" => Some(Format::Eps),
+        "ps" => Some(Format::PostScript),
+        "ai" => Some(Format::Illustrator),
+        "dxf" => Some(Format::Dxf),
+        "svg" => Some(Format::Svg),
+        "svgz" => Some(Format::Svgz),
+        "png" => Some(Format::Png),
+        "jpg" | "jpeg" => Some(Format::Jpeg),
+        "bmp" => Some(Format::Bmp),
+        "gif" => Some(Format::Gif),
+        "tif" | "tiff" => Some(Format::Tiff),
+        "webp" => Some(Format::WebP),
+        "heic" | "heif" => Some(Format::Heif),
+        "ase" => Some(Format::Ase),
+        "psd" | "psb" => Some(Format::Psd),
+        "afdesign" | "afphoto" | "afpub" => Some(Format::Affinity),
+        "ora" => Some(Format::Ora),
+        "kra" => Some(Format::Kra),
+        "xcf" => Some(Format::Xcf),
+        "pdn" => Some(Format::Pdn),
+        "dng" | "3fr" | "ari" | "arw" | "bay" | "cap" | "cr2" | "cr3" | "dcs" | "dcr" | "drf" | "eip" | "erf" | "fff" | "gpr" | "iiq" | "k25"
+        | "kdc" | "mdc" | "mef" | "mos" | "mrw" | "nef" | "nrw" | "obm" | "orf" | "pef" | "ptx" | "pxn" | "r3d" | "raf" | "raw" | "rwl" | "rw2"
+        | "rwz" | "sr2" | "srf" | "sti" | "x3f" => Some(Format::Raw),
+        "pcraft" => Some(Format::Pcraft),
+        "idml" => Some(Format::Idml),
+        "designcraft" | "dcdoc" => Some(Format::DesignCraft),
+        "indd" | "indt" => Some(Format::Indd),
+        "docx" => Some(Format::Docx),
+        "doc" => Some(Format::LegacyDoc),
+        "rtf" => Some(Format::Rtf),
+        "xlsx" => Some(Format::Xlsx),
+        "pptx" => Some(Format::Pptx),
+        "txt" | "text" | "md" => Some(Format::PlainText),
+        "csv" => Some(Format::Csv),
+        "tsv" | "tab" => Some(Format::Tsv),
+        "xls" | "xlsm" => Some(Format::LegacyExcel),
+        "studio" | "studio3" => Some(Format::Studio),
+        "stencilcraft" => Some(Format::StencilCraft),
+        "gcode" | "nc" => Some(Format::Gcode),
+        _ => None,
+    }
+}
+
+const INDD_SIGNATURE: [u8; 16] = [0x06, 0x06, 0xED, 0xF5, 0xD8, 0x1D, 0x46, 0xE5, 0xBD, 0x31, 0xEF, 0xE7, 0xFE, 0x74, 0xB7, 0x1D];
+
+fn extension(name: &str) -> Option<&str> {
+    Some(name.rsplit(['/', '\\']).next()?.rsplit_once('.')?.1)
+}
+
+fn gzipped_xcf(name: &str, bytes: &[u8]) -> bool {
+    let lower = name.rsplit(['/', '\\']).next().unwrap_or(name).to_ascii_lowercase();
+    bytes.starts_with(&[0x1f, 0x8b]) && (lower.ends_with(".xcf.gz") || lower.ends_with(".xcfgz"))
+}
+
+fn dxf_ascii_prefix(bytes: &[u8]) -> bool {
+    let text = String::from_utf8_lossy(bytes);
+    let mut lines = text.lines().map(str::trim);
+    matches!((lines.next(), lines.next(), lines.next(), lines.next()), (Some("0"), Some("SECTION"), Some("2"), Some(section)) if matches!(section, "HEADER" | "CLASSES" | "TABLES" | "BLOCKS" | "ENTITIES" | "OBJECTS"))
+}
+
+fn raster_signature(bytes: &[u8]) -> Option<Format> {
+    let head = bytes.get(..16).unwrap_or(bytes);
+    if head.starts_with(b"\x89PNG\r\n\x1a\n") {
+        return Some(Format::Png);
+    }
+    if head.starts_with(b"\xff\xd8\xff") {
+        return Some(Format::Jpeg);
+    }
+    if head.starts_with(b"BM") {
+        return Some(Format::Bmp);
+    }
+    if head.starts_with(b"GIF87a") || head.starts_with(b"GIF89a") {
+        return Some(Format::Gif);
+    }
+    if head.starts_with(b"II*\0") || head.starts_with(b"MM\0*") || head.starts_with(b"II+\0") || head.starts_with(b"MM\0+") {
+        return Some(Format::Tiff);
+    }
+    if head.starts_with(b"RIFF") && head.get(8..12) == Some(b"WEBP") {
+        return Some(Format::WebP);
+    }
+    None
+}