rustypaper 0.2.1__tar.gz → 0.2.2__tar.gz

{rustypaper-0.2.1 → rustypaper-0.2.2}/Cargo.lock

@@ -567,7 +567,7 @@ dependencies = [
 
 [[package]]
 name = "rustypaper"
-version = "0.2.1"
+version = "0.2.2"
 dependencies = [
  "anyhow",
  "clap",
@@ -580,7 +580,7 @@ dependencies = [
 
 [[package]]
 name = "rustypaper-py"
-version = "0.2.1"
+version = "0.2.2"
 dependencies = [
  "pyo3",
  "rustypaper",

{rustypaper-0.2.1 → rustypaper-0.2.2}/Cargo.toml

@@ -3,14 +3,14 @@ resolver = "2"
 members = ["rustypaper", "rustypaper-py"]
 
 [workspace.package]
-version = "0.2.1"
+version = "0.2.2"
 edition = "2021"
 rust-version = "1.94"
 license = "MIT OR Apache-2.0"
 repository = "https://github.com/pgarrett-scripps/rustypaper"
 
 [workspace.dependencies]
-rustypaper = { path = "rustypaper", version = "0.2.1" }
+rustypaper = { path = "rustypaper", version = "0.2.2" }
 # The pure-Rust PDF backend, from the registry. Deliberately not a path dependency: a path
 # only resolves where the sibling checkout happens to exist, which is this machine and not a
 # CI runner, and `cargo fmt --all` walks into it whether it is there or not. To develop the

{rustypaper-0.2.1 → rustypaper-0.2.2}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: rustypaper
-Version: 0.2.1
+Version: 0.2.2
 Classifier: Programming Language :: Rust
 Classifier: Topic :: Text Processing :: Markup
 Classifier: Topic :: Scientific/Engineering
@@ -28,15 +28,38 @@ crates that exist are generic text extractors with no notion of a paper. This ai
 ## Install
 
 ```sh
-cargo install rustypaper          # the command-line tool
+pip install rustypaper
+rustypaper convert paper.pdf
+python -m rustypaper convert paper.pdf
 ```
 
-```toml
-rustypaper = "0.2"            # the library
-```
+The Python installation includes the same command line interface as the Rust executable.
+All commands and options are shared, including batch conversion, figure extraction, and
+diagnostics. No Rust compiler is needed when installing a supported platform wheel.
+
+Standalone executables are available on the
+[GitHub Releases page](https://github.com/pgarrett-scripps/rustypaper/releases).
+Download the archive for your platform, verify it against `SHA256SUMS`, extract it,
+and put `rustypaper` (or `rustypaper.exe`) on your `PATH`. These executables do not need Python.
+
+| Platform | Targets |
+|---|---|
+| Linux with glibc 2.35 or newer | x86_64, aarch64 |
+| macOS | Intel, Apple Silicon |
+| Windows | x86_64 |
+
+Alternatively, use [cargo-binstall](https://github.com/cargo-bins/cargo-binstall)
+to download a release executable, or Cargo to compile it locally:
 
 ```sh
-pip install rustypaper        # the Python bindings
+cargo binstall rustypaper
+cargo install rustypaper
+```
+
+For the Rust library:
+
+```toml
+rustypaper = "0.2"
 ```
 
 **Nothing else to install.** PDFs are read with

{rustypaper-0.2.1/rustypaper → rustypaper-0.2.2}/README.md

@@ -15,15 +15,38 @@ crates that exist are generic text extractors with no notion of a paper. This ai
 ## Install
 
 ```sh
-cargo install rustypaper          # the command-line tool
+pip install rustypaper
+rustypaper convert paper.pdf
+python -m rustypaper convert paper.pdf
 ```
 
-```toml
-rustypaper = "0.2"            # the library
-```
+The Python installation includes the same command line interface as the Rust executable.
+All commands and options are shared, including batch conversion, figure extraction, and
+diagnostics. No Rust compiler is needed when installing a supported platform wheel.
+
+Standalone executables are available on the
+[GitHub Releases page](https://github.com/pgarrett-scripps/rustypaper/releases).
+Download the archive for your platform, verify it against `SHA256SUMS`, extract it,
+and put `rustypaper` (or `rustypaper.exe`) on your `PATH`. These executables do not need Python.
+
+| Platform | Targets |
+|---|---|
+| Linux with glibc 2.35 or newer | x86_64, aarch64 |
+| macOS | Intel, Apple Silicon |
+| Windows | x86_64 |
+
+Alternatively, use [cargo-binstall](https://github.com/cargo-bins/cargo-binstall)
+to download a release executable, or Cargo to compile it locally:
 
 ```sh
-pip install rustypaper        # the Python bindings
+cargo binstall rustypaper
+cargo install rustypaper
+```
+
+For the Rust library:
+
+```toml
+rustypaper = "0.2"
 ```
 
 **Nothing else to install.** PDFs are read with

{rustypaper-0.2.1 → rustypaper-0.2.2}/pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "rustypaper"
-version = "0.2.1"
+version = "0.2.2"
 description = "Structure-aware conversion of born-digital scientific PDFs to Markdown, Typst and JSON."
 readme = "README.md"
 requires-python = ">=3.9"
@@ -15,6 +15,9 @@ classifiers = [
 [project.urls]
 Repository = "https://github.com/pgarrett-scripps/rustypaper"
 
+[project.scripts]
+rustypaper = "rustypaper.cli:main"
+
 # Built with maturin from the workspace root:
 #   maturin build -m rustypaper-py/Cargo.toml
 # `scripts/build.sh` is the no-maturin path: cargo produces an importable .so directly.

{rustypaper-0.2.1 → rustypaper-0.2.2}/rustypaper/Cargo.toml

@@ -18,8 +18,7 @@ name = "rustypaper"
 path = "src/main.rs"
 
 [dependencies]
-# Used by the binary only, but cargo has no way to say so for a bin inside a library crate.
-# They are small and pure Rust, so the cost to a library consumer is compile time, not weight.
+# The native executable and Python entry point share the command line implementation.
 anyhow.workspace = true
 clap.workspace = true
 

{rustypaper-0.2.1 → rustypaper-0.2.2/rustypaper}/README.md

@@ -15,15 +15,38 @@ crates that exist are generic text extractors with no notion of a paper. This ai
 ## Install
 
 ```sh
-cargo install rustypaper          # the command-line tool
+pip install rustypaper
+rustypaper convert paper.pdf
+python -m rustypaper convert paper.pdf
 ```
 
-```toml
-rustypaper = "0.2"            # the library
-```
+The Python installation includes the same command line interface as the Rust executable.
+All commands and options are shared, including batch conversion, figure extraction, and
+diagnostics. No Rust compiler is needed when installing a supported platform wheel.
+
+Standalone executables are available on the
+[GitHub Releases page](https://github.com/pgarrett-scripps/rustypaper/releases).
+Download the archive for your platform, verify it against `SHA256SUMS`, extract it,
+and put `rustypaper` (or `rustypaper.exe`) on your `PATH`. These executables do not need Python.
+
+| Platform | Targets |
+|---|---|
+| Linux with glibc 2.35 or newer | x86_64, aarch64 |
+| macOS | Intel, Apple Silicon |
+| Windows | x86_64 |
+
+Alternatively, use [cargo-binstall](https://github.com/cargo-bins/cargo-binstall)
+to download a release executable, or Cargo to compile it locally:
 
 ```sh
-pip install rustypaper        # the Python bindings
+cargo binstall rustypaper
+cargo install rustypaper
+```
+
+For the Rust library:
+
+```toml
+rustypaper = "0.2"
 ```
 
 **Nothing else to install.** PDFs are read with

rustypaper-0.2.2/rustypaper/__main__.py

@@ -0,0 +1,3 @@
+from .cli import main
+
+raise SystemExit(main())

rustypaper-0.2.2/rustypaper/cli.py

@@ -0,0 +1,10 @@
+"""Command line entry point backed by the shared Rust implementation."""
+
+import sys
+
+from ._rustypaper import _run_cli
+
+
+def main():
+    """Return the native CLI's exit status without exiting the Python host."""
+    return _run_cli(["rustypaper", *sys.argv[1:]])

rustypaper-0.2.1/rustypaper/src/main.rs → rustypaper-0.2.2/rustypaper/src/cli.rs

@@ -1,4 +1,8 @@
+//! Shared command line implementation for the native executable and Python package.
+
+use crate as rustypaper;
 use std::collections::BTreeMap;
+use std::ffi::OsString;
 use std::io::{BufWriter, Write};
 use std::path::PathBuf;
 use std::time::Instant;
@@ -52,7 +56,7 @@ impl From<Caveman> for rustypaper::compress::Level {
 enum Format {
     /// GitHub-flavoured Markdown.
     Md,
-    /// The document model, which is the real output; every other format renders this.
+    /// The document model. Every other format renders this.
     Json,
     /// Typst, which models figures, tables and references natively.
     Typst,
@@ -95,7 +99,7 @@ enum Command {
         #[arg(long, default_value_t = 150.0)]
         figure_dpi: f32,
         /// Strip grammatical scaffolding from the prose, for feeding to a model that charges by
-        /// the token. `light` drops articles, copulas and stock phrases; `hard` also drops
+        /// the token. `light` drops articles, copulas and stock phrases. `hard` also drops
         /// prepositions, pronouns and filler. Maths, tables and citations are left alone.
         #[arg(long, value_enum, num_args = 0..=1, default_missing_value = "light")]
         caveman: Option<Caveman>,
@@ -112,10 +116,35 @@ enum Command {
     },
 }
 
-fn main() -> Result<()> {
+/// Run a command, including its program name, without terminating the host process.
+pub fn run(args: impl IntoIterator<Item = OsString>) -> i32 {
+    let cli = match Cli::try_parse_from(args) {
+        Ok(cli) => cli,
+        Err(error) => {
+            let code = error.exit_code();
+            if let Err(error) = error.print() {
+                return if error.kind() == std::io::ErrorKind::BrokenPipe {
+                    0
+                } else {
+                    1
+                };
+            }
+            return code;
+        }
+    };
+    match execute(cli.command) {
+        Ok(()) => 0,
+        Err(error) => {
+            let _ = writeln!(std::io::stderr().lock(), "Error: {error:?}");
+            1
+        }
+    }
+}
+
+fn execute(command: Command) -> Result<()> {
     let mut out = BufWriter::new(std::io::stdout().lock());
 
-    let result = match Cli::parse().command {
+    let result = match command {
         Command::Dump { pdf, page, pretty } => dump(&mut out, pdf, page, pretty),
         Command::Convert {
             pdf,
@@ -146,7 +175,7 @@ fn is_broken_pipe(error: &anyhow::Error) -> bool {
         // A write that fails on a closed pipe can arrive as either type, and which one depends
         // on who was writing. `serde_json` wraps the io error rather than returning it, so
         // looking only for `io::Error` misses every failure from `--pretty`, which serialises
-        // straight to the stream — `dump --pretty | head` exited 1 for exactly that reason.
+        // straight to the stream. `dump --pretty | head` exited 1 for exactly that reason.
         if let Some(io) = e.downcast_ref::<std::io::Error>() {
             return io.kind() == std::io::ErrorKind::BrokenPipe;
         }

{rustypaper-0.2.1 → rustypaper-0.2.2}/rustypaper/src/lib.rs

@@ -13,6 +13,7 @@
 //! The `Document` is the real output; Markdown is one rendering of it.
 
 pub mod backend;
+pub mod cli;
 pub mod compress;
 pub mod doc;
 pub mod emit;

rustypaper-0.2.2/rustypaper/src/main.rs

@@ -0,0 +1,3 @@
+fn main() -> std::process::ExitCode {
+    std::process::ExitCode::from(rustypaper::cli::run(std::env::args_os()) as u8)
+}

{rustypaper-0.2.1 → rustypaper-0.2.2}/rustypaper-py/src/lib.rs

@@ -12,6 +12,7 @@
 use pyo3::create_exception;
 use pyo3::exceptions::{PyIOError, PyValueError};
 use pyo3::prelude::*;
+use std::ffi::OsString;
 
 create_exception!(
     _rustypaper,
@@ -152,6 +153,7 @@ fn extract_json(py: Python<'_>, path: &str) -> PyResult<String> {
 
 #[pymodule]
 fn _rustypaper(module: &Bound<'_, PyModule>) -> PyResult<()> {
+    module.add_function(wrap_pyfunction!(_run_cli, module)?)?;
     module.add("__version__", env!("CARGO_PKG_VERSION"))?;
     module.add("ScannedDocument", module.py().get_type::<ScannedDocument>())?;
     module.add_function(wrap_pyfunction!(to_markdown, module)?)?;
@@ -162,3 +164,9 @@ fn _rustypaper(module: &Bound<'_, PyModule>) -> PyResult<()> {
     module.add_function(wrap_pyfunction!(extract_json, module)?)?;
     Ok(())
 }
+
+/// Run the same CLI as the native executable, returning its exit status.
+#[pyfunction]
+fn _run_cli(py: Python<'_>, args: Vec<OsString>) -> i32 {
+    py.detach(|| rustypaper::cli::run(args))
+}