From 79c2003ac9c665bfd6a0dbc00c256d42cc3ad1e3 Mon Sep 17 00:00:00 2001
From: "Dylan M. Taylor" <dylan@dylanmtaylor.com>
Date: Fri, 13 Nov 2026 00:29:48 -0500
Subject: [PATCH] Add shared file format detection

---
 Cargo.lock                     |  7 +++
 Cargo.toml                     |  1 +
 crates/file-formats/Cargo.toml | 14 +++++
 crates/file-formats/src/lib.rs | 93 ++++++++++++++++++++++++++++++++++
 4 files changed, 115 insertions(+)
 create mode 100644 crates/file-formats/Cargo.toml
 create mode 100644 crates/file-formats/src/lib.rs

diff --git a/Cargo.lock b/Cargo.lock
index cf0e0bd47..5e004a7ce 100644
--- a/Cargo.lock
+++ b/Cargo.lock
@@ -1454,6 +1454,13 @@ dependencies = [
  "thiserror 2.0.21",
 ]
 
+[[package]]
+name = "craft-file-formats"
+version = "0.1.0"
+dependencies = [
+ "craft-eps",
+]
+
 [[package]]
 name = "craft-flags"
 version = "0.1.0"
diff --git a/Cargo.toml b/Cargo.toml
index 629bab64c..4af6f586a 100644
--- a/Cargo.toml
+++ b/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" }
diff --git a/crates/file-formats/Cargo.toml b/crates/file-formats/Cargo.toml
new file mode 100644
index 000000000..0a5be5e76
--- /dev/null
+++ b/crates/file-formats/Cargo.toml
@@ -0,0 +1,14 @@
+[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 }
+
+[lints]
+workspace = true
diff --git a/crates/file-formats/src/lib.rs b/crates/file-formats/src/lib.rs
new file mode 100644
index 000000000..b51fc399c
--- /dev/null
+++ b/crates/file-formats/src/lib.rs
@@ -0,0 +1,93 @@
+//! 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,
+    Raster,
+    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 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 raster_signature(bytes) {
+        return Some(Format::Raster);
+    }
+    if bytes.starts_with(b"silhouette04;") || bytes.starts_with(b"silhouette05;") {
+        return Some(Format::Studio);
+    }
+    if let Some(ext) = extension(name) {
+        if let Some(format) = from_extension(ext) {
+            return Some(format);
+        }
+    }
+    let head = bytes.get(..4096).unwrap_or(bytes);
+    let xml = head.trim_ascii_start();
+    if xml.starts_with(b"<svg") || (xml.starts_with(b"<?xml") && xml.windows(4).any(|w| w == b"<svg")) {
+        return Some(Format::Svg);
+    }
+    if dxf_ascii_prefix(head) || head.starts_with(b"AutoCAD Binary DXF\r\n\x1a\0") {
+        return Some(Format::Dxf);
+    }
+    None
+}
+
+/// 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" | "svgz" => Some(Format::Svg),
+        "png" | "jpg" | "jpeg" | "bmp" | "gif" | "tif" | "tiff" | "webp" | "heic" | "heif" => Some(Format::Raster),
+        "studio" | "studio3" => Some(Format::Studio),
+        "stencilcraft" => Some(Format::StencilCraft),
+        _ => None,
+    }
+}
+
+fn extension(name: &str) -> Option<&str> {
+    Some(name.rsplit(['/', '\\']).next()?.rsplit_once('.')?.1)
+}
+
+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]) -> bool {
+    let head = bytes.get(..16).unwrap_or(bytes);
+    head.starts_with(b"\x89PNG\r\n\x1a\n")
+        || head.starts_with(b"\xff\xd8\xff")
+        || head.starts_with(b"BM")
+        || head.starts_with(b"GIF87a")
+        || head.starts_with(b"GIF89a")
+        || head.starts_with(b"II*\0")
+        || head.starts_with(b"MM\0*")
+        || head.starts_with(b"II+\0")
+        || head.starts_with(b"MM\0+")
+        || (head.starts_with(b"RIFF") && head.get(8..12) == Some(b"WEBP"))
+}