feat(wasi): Wasmtime integration test target (8/8 ✅)

Makefile

@@ -530,6 +530,26 @@ wasm-clean: ## Remove WASM build artefacts (pkg/)
 	$(call warn,Removing $(WASM_CRATE)/pkg/ ...)
 	@rm -rf $(WASM_CRATE)/pkg/
 
+# ══════════════════════════════════════════════════════════════════════════════
+## WASI / RISC-V Integration Tests  (Docker-based)
+# ══════════════════════════════════════════════════════════════════════════════
+
+WASI_SCRIPT := tests/wasm-runtimes/wasm-test.sh
+
+wasi-build: ## Build all WASM runtime + RISC-V Docker test images
+	$(call log,Building WASM/RISC-V integration test images…)
+	@$(WASI_SCRIPT) build all
+
+wasi-test: ## Run WASM/RISC-V integration tests across all runtimes
+	$(call log,Running WASM/RISC-V integration tests…)
+	@$(WASI_SCRIPT) test all
+
+wasi-status: ## Show Docker image / container status for WASI tests
+	@$(WASI_SCRIPT) status
+
+wasi-clean: ## Remove all WASI test Docker images and artefacts
+	@$(WASI_SCRIPT) clean
+
 # ══════════════════════════════════════════════════════════════════════════════
 ## Clean
 # ══════════════════════════════════════════════════════════════════════════════

crates/edgeparse-cli/Cargo.toml

@@ -15,12 +15,16 @@ documentation = "https://docs.rs/edgeparse-cli"
 name = "edgeparse"
 path = "src/main.rs"
 
+[features]
+default = ["native"]
+native = ["rayon", "edgeparse-core/native"]
+
 [dependencies]
-edgeparse-core = { path = "../edgeparse-core", version = "0.2.0" }
+edgeparse-core = { path = "../edgeparse-core", version = "0.2.0", default-features = false }
 clap = { workspace = true }
 serde = { workspace = true }
 serde_json = { workspace = true }
 anyhow = { workspace = true }
 log = { workspace = true }
 env_logger = { workspace = true }
-rayon = { workspace = true }
+rayon = { workspace = true, optional = true }

crates/edgeparse-cli/src/main.rs

@@ -6,6 +6,7 @@ use std::sync::atomic::{AtomicBool, Ordering};
 
 use clap::Parser;
 use edgeparse_core::api::config::OutputFormat;
+#[cfg(feature = "native")]
 use rayon::prelude::*;
 
 /// EdgeParse: High-performance PDF-to-structured-data extraction
@@ -124,10 +125,12 @@ fn main() {
     // Build processing config
     let config = build_config(&cli);
 
-    // Process each input file in parallel
+    // Process each input file (parallel when native feature is enabled)
     let has_errors = AtomicBool::new(false);
-    cli.input.par_iter().for_each(|input_path| {
-        match edgeparse_core::convert(input_path, &config) {
+
+    let process_file = |input_path: &PathBuf| {
+        let result = convert_file(input_path, &config);
+        match result {
             Ok(doc) => {
                 log::info!(
                     "Processed {} ({} pages)",
@@ -144,7 +147,13 @@ fn main() {
                 has_errors.store(true, Ordering::Relaxed);
             }
         }
-    });
+    };
+
+    #[cfg(feature = "native")]
+    cli.input.par_iter().for_each(process_file);
+
+    #[cfg(not(feature = "native"))]
+    cli.input.iter().for_each(process_file);
 
     if has_errors.load(Ordering::Relaxed) {
         process::exit(1);
@@ -202,6 +211,30 @@ fn write_outputs(
     Ok(())
 }
 
+/// Convert a PDF file using the appropriate backend.
+///
+/// On native builds, uses `edgeparse_core::convert()` which supports raster
+/// table OCR via external tools. On WASI/non-native builds, reads the file
+/// into memory and uses `convert_bytes()` instead (no external tool support).
+fn convert_file(
+    input_path: &std::path::Path,
+    config: &edgeparse_core::api::config::ProcessingConfig,
+) -> Result<edgeparse_core::models::document::PdfDocument, edgeparse_core::EdgePdfError> {
+    #[cfg(feature = "native")]
+    {
+        edgeparse_core::convert(input_path, config)
+    }
+    #[cfg(not(feature = "native"))]
+    {
+        let data = std::fs::read(input_path)?;
+        let file_name = input_path
+            .file_name()
+            .and_then(|n| n.to_str())
+            .unwrap_or("unknown.pdf");
+        edgeparse_core::convert_bytes(&data, file_name, config)
+    }
+}
+
 fn build_config(cli: &Cli) -> edgeparse_core::api::config::ProcessingConfig {
     use edgeparse_core::api::config::*;
     use edgeparse_core::api::filter::FilterConfig;

tests/wasm-runtimes/.gitignore

@@ -0,0 +1,2 @@
+# Build artifacts — extracted binaries from Docker builds
+.build/

tests/wasm-runtimes/Dockerfile.build.riscv

@@ -0,0 +1,68 @@
+# ─── Reproducible RISC-V Cross-Compilation ──────────────────────────────────
+# Greg's AI coding buddy:
+#   Cross-compiles edgeparse for riscv64gc-unknown-linux-gnu.
+#   Produces TWO binaries:
+#     1. Dynamic-linked  → runs under QEMU user-mode (with sysroot)
+#     2. Static-linked   → runs on Spike/libriscv/RVVM/CKB-VM (no sysroot)
+#
+#   riscv64gc = RV64IMAFDC — the "general-purpose computing" profile
+#   that Linux distros target (Debian, Ubuntu, Fedora all ship riscv64gc).
+#
+# Usage:
+#   docker build -f tests/wasm-runtimes/Dockerfile.build.riscv \
+#                -t edgeparse-riscv-build .
+# ─────────────────────────────────────────────────────────────────────────────
+
+FROM rust:1-slim-bookworm AS builder
+
+# Cross-compilation toolchain for RISC-V 64-bit
+RUN apt-get update && apt-get install -y --no-install-recommends \
+    gcc-riscv64-linux-gnu \
+    libc6-dev-riscv64-cross \
+    pkg-config \
+    && rm -rf /var/lib/apt/lists/*
+
+RUN rustup target add riscv64gc-unknown-linux-gnu
+
+# Tell cargo which linker to use for the RISC-V target
+ENV CARGO_TARGET_RISCV64GC_UNKNOWN_LINUX_GNU_LINKER=riscv64-linux-gnu-gcc
+
+WORKDIR /build
+
+# ── Cache cargo registry: copy manifests first ──────────────────────────────
+COPY Cargo.toml Cargo.lock ./
+COPY crates/edgeparse-core/Cargo.toml   crates/edgeparse-core/
+COPY crates/edgeparse-cli/Cargo.toml    crates/edgeparse-cli/
+COPY crates/pdf-cos/Cargo.toml          crates/pdf-cos/
+
+# Strip workspace members that need native platform SDKs
+RUN sed -E -i \
+      '/"crates\/(edgeparse-python|edgeparse-node|edgeparse-wasm)"/d' \
+      Cargo.toml && \
+    mkdir -p crates/edgeparse-core/src crates/edgeparse-cli/src crates/pdf-cos/src && \
+    echo "fn main() {}" > crates/edgeparse-cli/src/main.rs && \
+    touch crates/edgeparse-core/src/lib.rs && \
+    touch crates/pdf-cos/src/lib.rs && \
+    cargo build --target riscv64gc-unknown-linux-gnu --release \
+        -p edgeparse-cli 2>/dev/null || true
+
+# ── Copy real source and build (dynamic) ────────────────────────────────────
+COPY crates/ crates/
+RUN find crates/ -name '*.rs' -exec touch {} + && \
+    cargo build --target riscv64gc-unknown-linux-gnu --release \
+        -p edgeparse-cli
+
+# Save dynamic binary before static build overwrites it
+RUN cp target/riscv64gc-unknown-linux-gnu/release/edgeparse /tmp/edgeparse-dynamic
+
+# ── Build statically-linked variant for Spike/libriscv/RVVM/CKB-VM ─────────
+# These VMs have minimal/no dynamic linker — static linking is essential.
+RUN RUSTFLAGS="-C target-feature=+crt-static" \
+    cargo build --target riscv64gc-unknown-linux-gnu --release \
+        -p edgeparse-cli && \
+    cp target/riscv64gc-unknown-linux-gnu/release/edgeparse /tmp/edgeparse-static
+
+# ── Output stage ────────────────────────────────────────────────────────────
+FROM scratch AS output
+COPY --from=builder /tmp/edgeparse-dynamic /out/edgeparse
+COPY --from=builder /tmp/edgeparse-static  /out/edgeparse-static

tests/wasm-runtimes/Dockerfile.build.wasix

@@ -0,0 +1,47 @@
+# ─── Reproducible WASIX Build ───────────────────────────────────────────────
+# Greg's AI coding buddy:
+#   Builds edgeparse as a WASIX binary (.wasm) using cargo-wasix.
+#   WASIX = WASI + POSIX superpowers (threads, sockets, fork/exec).
+#   Only runs on Wasmer — the one runtime to rule them all (for WASIX).
+#
+#   The target triple is wasm32-wasmer-wasi — a custom Wasmer target
+#   that extends wasm32-wasi with the full POSIX syscall surface.
+#
+# Usage:
+#   docker build -f tests/wasm-runtimes/Dockerfile.build.wasix \
+#                -t edgeparse-wasix-build .
+# ─────────────────────────────────────────────────────────────────────────────
+
+FROM rust:1-slim-bookworm AS builder
+
+# cargo-wasix installs its own rustup toolchain + wasm32-wasmer-wasi target
+RUN cargo install cargo-wasix
+
+WORKDIR /build
+
+# ── Cache cargo registry: copy manifests first ──────────────────────────────
+COPY Cargo.toml Cargo.lock ./
+COPY crates/edgeparse-core/Cargo.toml   crates/edgeparse-core/
+COPY crates/edgeparse-cli/Cargo.toml    crates/edgeparse-cli/
+COPY crates/pdf-cos/Cargo.toml          crates/pdf-cos/
+
+# Strip workspace members that don't compile for WASIX
+RUN sed -E -i \
+      '/"crates\/(edgeparse-python|edgeparse-node|edgeparse-wasm)"/d' \
+      Cargo.toml && \
+    mkdir -p crates/edgeparse-core/src crates/edgeparse-cli/src crates/pdf-cos/src && \
+    echo "fn main() {}" > crates/edgeparse-cli/src/main.rs && \
+    touch crates/edgeparse-core/src/lib.rs && \
+    touch crates/pdf-cos/src/lib.rs && \
+    cargo wasix build --release -p edgeparse-cli --no-default-features 2>/dev/null || true
+
+# ── Copy real source and build ──────────────────────────────────────────────
+COPY crates/ crates/
+RUN find crates/ -name '*.rs' -exec touch {} + && \
+    cargo wasix build --release -p edgeparse-cli --no-default-features
+
+# ── Output stage ────────────────────────────────────────────────────────────
+FROM scratch AS output
+COPY --from=builder \
+     /build/target/wasm32-wasmer-wasi/release/edgeparse.wasm \
+     /out/edgeparse-wasix.wasm

tests/wasm-runtimes/Dockerfile.build.wasm

@@ -0,0 +1,54 @@
+# ─── Reproducible WASM (WASI) Build ─────────────────────────────────────────
+# Greg's AI coding buddy:
+#   Builds edgeparse as a WASI Preview 1 binary (.wasm) that runs on
+#   any conformant runtime (wasmtime, wasmer, wasmedge, wamr).
+#
+#   The wasm32-wasip1 target compiles edgeparse-core without rayon/image/zip
+#   (those need native threads or JS glue). Pure PDF parsing still works —
+#   you just lose parallelism and image extraction.
+#
+# Usage:
+#   docker build -f tests/wasm-runtimes/Dockerfile.build.wasm \
+#                -t edgeparse-wasi-build .
+#   # Extract the binary:
+#   id=$(docker create edgeparse-wasi-build) && \
+#     docker cp "$id":/out/edgeparse.wasm tests/wasm-runtimes/.build/ && \
+#     docker rm "$id"
+# ─────────────────────────────────────────────────────────────────────────────
+
+FROM rust:1-slim-bookworm AS builder
+
+RUN rustup target add wasm32-wasip1
+
+WORKDIR /build
+
+# ── Cache cargo registry: copy manifests first ──────────────────────────────
+COPY Cargo.toml Cargo.lock ./
+COPY crates/edgeparse-core/Cargo.toml   crates/edgeparse-core/
+COPY crates/edgeparse-cli/Cargo.toml    crates/edgeparse-cli/
+COPY crates/pdf-cos/Cargo.toml          crates/pdf-cos/
+
+# Strip workspace members that don't compile for wasm32-wasip1
+# (pyo3, napi-rs, wasm-bindgen are browser-only / native-only)
+RUN sed -E -i \
+      '/"crates\/(edgeparse-python|edgeparse-node|edgeparse-wasm)"/d' \
+      Cargo.toml && \
+    mkdir -p crates/edgeparse-core/src crates/edgeparse-cli/src crates/pdf-cos/src && \
+    echo "fn main() {}" > crates/edgeparse-cli/src/main.rs && \
+    touch crates/edgeparse-core/src/lib.rs && \
+    touch crates/pdf-cos/src/lib.rs && \
+    cargo build --target wasm32-wasip1 --release \
+        -p edgeparse-cli --no-default-features 2>/dev/null || true
+    # ^^ dummy build warms the dep cache; may fail on empty libs — that's fine
+
+# ── Copy real source and build ──────────────────────────────────────────────
+COPY crates/ crates/
+RUN find crates/ -name '*.rs' -exec touch {} + && \
+    cargo build --target wasm32-wasip1 --release \
+        -p edgeparse-cli --no-default-features
+
+# ── Output stage — just the binary ──────────────────────────────────────────
+FROM scratch AS output
+COPY --from=builder \
+     /build/target/wasm32-wasip1/release/edgeparse.wasm \
+     /out/edgeparse.wasm

tests/wasm-runtimes/Dockerfile.runner.base

@@ -0,0 +1,31 @@
+# ─── Shared Runner Base ─────────────────────────────────────────────────────
+# Greg's AI coding buddy:
+#   Common base layer for all WASM runtime test containers.
+#   Build this FIRST — subsequent runtime images inherit its cached layers.
+#
+#   docker build -f tests/wasm-runtimes/Dockerfile.runner.base \
+#                -t edgeparse-wasi-base .
+# ─────────────────────────────────────────────────────────────────────────────
+
+FROM ubuntu:24.04
+
+# Shared dependencies — curl for runtime installers, ca-certs for HTTPS,
+# xz-utils for compressed release tarballs, file for MIME sniffing
+RUN apt-get update && apt-get install -y --no-install-recommends \
+    curl \
+    ca-certificates \
+    xz-utils \
+    file \
+    && rm -rf /var/lib/apt/lists/*
+
+WORKDIR /test
+
+# Copy test fixtures and scripts
+COPY tests/fixtures/sample.pdf          /test/fixtures/
+COPY tests/wasm-runtimes/run-tests.sh   /test/
+RUN chmod +x /test/run-tests.sh
+
+# The .wasm binary is expected at /test/edgeparse.wasm
+# It gets copied in by each runtime Dockerfile or mounted at runtime.
+
+ENTRYPOINT ["/test/run-tests.sh"]

tests/wasm-runtimes/Dockerfile.runner.ckb-vm

@@ -0,0 +1,46 @@
+# ─── CKB-VM Runner ──────────────────────────────────────────────────────────
+# Greg's AI coding buddy:
+#   CKB-VM — Production blockchain VM from Nervos Network.
+#   rv64imc ISA, W^X memory protection, gas metering, JIT compilation.
+#   2.5x faster than Wasmer Singlepass. Deployed on CKB mainnet.
+#
+#   Uses ckb-debugger to execute RISC-V ELF binaries with
+#   Linux syscall emulation. Limited POSIX support — designed for
+#   deterministic computation, not general-purpose Linux apps.
+#
+#   EXPERIMENTAL: CKB-VM has limited syscall support. edgeparse's
+#   file I/O may not be fully supported. Tests may partially fail.
+#
+#   docker build -f tests/wasm-runtimes/Dockerfile.runner.ckb-vm \
+#                -t edgeparse-riscv-ckb-vm .
+# ─────────────────────────────────────────────────────────────────────────────
+
+FROM rust:1-slim-bookworm AS builder
+
+# Build ckb-standalone-debugger which can run RISC-V ELF binaries
+RUN cargo install ckb-debugger 2>/dev/null || \
+    (apt-get update && apt-get install -y --no-install-recommends git pkg-config libssl-dev && \
+     cargo install --git https://github.com/nervosnetwork/ckb-standalone-debugger ckb-debugger)
+
+FROM ubuntu:24.04
+
+RUN apt-get update && apt-get install -y --no-install-recommends \
+    file \
+    && rm -rf /var/lib/apt/lists/*
+
+COPY --from=builder /usr/local/cargo/bin/ckb-debugger /usr/local/bin/
+
+# Verify
+RUN ckb-debugger --version 2>&1 || echo "CKB debugger installed"
+
+WORKDIR /test
+
+COPY tests/fixtures/sample.pdf          /test/fixtures/
+COPY tests/wasm-runtimes/run-tests.sh   /test/
+RUN chmod +x /test/run-tests.sh
+
+# Copy pre-built RISC-V binary (statically linked)
+COPY tests/wasm-runtimes/.build/edgeparse-riscv64-static /test/edgeparse-riscv64
+
+ENTRYPOINT ["/test/run-tests.sh"]
+CMD ["ckb-vm"]