Split leptonica out of tesseract-sys

This is done because `cargo` can only specify one library to link to per
crate:
https://doc.rust-lang.org/cargo/reference/build-scripts.html#the-links-manifest-key

I've had to change the image used, becuase Tesseract had started reading
the apostrophe as a double quote. I could have changed the tests to
adapt, but thought it better not to test an incorrect OCR.

#7

Author: ccouzens

Cargo.toml

@@ -9,6 +9,7 @@ license = "MIT"
 
 [dependencies]
 libc = "0.1"
+leptonica-sys = "0.1.0"
 
 [dependencies.tesseract-sys]
 version = "0.2"

img.png

@@ -1,6 +1,4 @@
-Rust is a systems programming language
-that runs blazingly fast, prevents almost all
-crashes*, and eliminates data races.
-
-Show me more!
-
+Hundreds of companies around the world are using Rust in production today for fast, low-
+resource, cross-platform solutions. Software you know and love, like Firefox, Dropbox, and
+Cloudflare, uses Rust. From startups to large corporations, from embedded devices to
+scalable web services, Rust is a great fit.

img.txt

@@ -1,82 +1,83 @@
-extern crate tesseract_sys;
+extern crate leptonica_sys;
 extern crate libc;
+extern crate tesseract_sys;
 
-use tesseract_sys::*;
+use leptonica_sys::pixRead;
+use std::ffi::CStr;
 use std::ffi::CString;
 use std::ptr;
 use std::str;
-use std::ffi::CStr;
-
+use tesseract_sys::{
+    TessBaseAPI, TessBaseAPICreate, TessBaseAPIDelete, TessBaseAPIGetUTF8Text, TessBaseAPIInit3,
+    TessBaseAPIRecognize, TessBaseAPISetImage2, TessBaseAPISetVariable,
+};
 
 pub struct Tesseract {
-	raw: *mut TessBaseAPI
+    raw: *mut TessBaseAPI,
 }
 
 impl Drop for Tesseract {
-	fn drop(&mut self) {
-		println!("Ave Imperator! Nos morituri te salutamus.");
-		unsafe { TessBaseAPIDelete(self.raw) }
-	}
+    fn drop(&mut self) {
+        println!("Ave Imperator! Nos morituri te salutamus.");
+        unsafe { TessBaseAPIDelete(self.raw) }
+    }
 }
 
 fn cs(string: &str) -> CString {
-	// do not call as_ptr yet, since the data will be freed before we return
-	CString::new(string).unwrap()
+    // do not call as_ptr yet, since the data will be freed before we return
+    CString::new(string).unwrap()
 }
 
 impl Tesseract {
-	pub fn new() -> Tesseract {
-		Tesseract {
-			raw: unsafe { TessBaseAPICreate() }
-		}
-	}
-	pub fn set_lang(&self, language: &str) -> i32 {
-		let cs_language = cs(language);
-		unsafe { TessBaseAPIInit3(self.raw, ptr::null(), cs_language.as_ptr()) }
-	}
-	pub fn set_image(&self, filename: &str) {
-		let cs_filename = cs(filename);
-		unsafe {
-			let img = pixRead(cs_filename.as_ptr());
-			TessBaseAPISetImage2(self.raw, img);
-		}
-	}
-	pub fn set_variable(&self, name: &str, value: &str) -> i32 {
-		let cs_name = cs(name);
-		let cs_value = cs(value);
-		unsafe { TessBaseAPISetVariable(self.raw, cs_name.as_ptr(), cs_value.as_ptr()) }
-	}
-	pub fn recognize(&self) -> i32 {
-		unsafe {
-			TessBaseAPIRecognize(self.raw, ptr::null_mut())
-		}
-	}
-	pub fn get_text(&self) -> &str {
-		unsafe {
-			str::from_utf8(CStr::from_ptr(TessBaseAPIGetUTF8Text(self.raw)).to_bytes()).unwrap()
-		}
-	}
+    pub fn new() -> Tesseract {
+        Tesseract {
+            raw: unsafe { TessBaseAPICreate() },
+        }
+    }
+    pub fn set_lang(&self, language: &str) -> i32 {
+        let cs_language = cs(language);
+        unsafe { TessBaseAPIInit3(self.raw, ptr::null(), cs_language.as_ptr()) }
+    }
+    pub fn set_image(&self, filename: &str) {
+        let cs_filename = cs(filename);
+        unsafe {
+            let img = pixRead(cs_filename.as_ptr());
+            TessBaseAPISetImage2(self.raw, img);
+        }
+    }
+    pub fn set_variable(&self, name: &str, value: &str) -> i32 {
+        let cs_name = cs(name);
+        let cs_value = cs(value);
+        unsafe { TessBaseAPISetVariable(self.raw, cs_name.as_ptr(), cs_value.as_ptr()) }
+    }
+    pub fn recognize(&self) -> i32 {
+        unsafe { TessBaseAPIRecognize(self.raw, ptr::null_mut()) }
+    }
+    pub fn get_text(&self) -> &str {
+        unsafe {
+            str::from_utf8(CStr::from_ptr(TessBaseAPIGetUTF8Text(self.raw)).to_bytes()).unwrap()
+        }
+    }
 }
 
 pub fn ocr(filename: &str, language: &str) -> String {
-	let cube = Tesseract::new();
-	cube.set_lang(language);
-	cube.set_image(filename);
-	cube.recognize();
-	return cube.get_text().to_string()
+    let cube = Tesseract::new();
+    cube.set_lang(language);
+    cube.set_image(filename);
+    cube.recognize();
+    return cube.get_text().to_string();
 }
 
-
 #[test]
-fn blah(){
-	ocr("img.png", "eng");
+fn blah() {
+    ocr("img.png", "eng");
 }
 
 #[test]
 fn it_works() {
-	let cube = Tesseract::new();
-	cube.set_lang("eng");
-	cube.set_image("img.png");
-	cube.recognize();
-	println!("{:?}", cube.get_text());
+    let cube = Tesseract::new();
+    cube.set_lang("eng");
+    cube.set_image("img.png");
+    cube.recognize();
+    println!("{:?}", cube.get_text());
 }

src/lib.rs

@@ -4,8 +4,15 @@ version = "0.2.0"
 authors = ["Kevin Kwok <[email protected]>", "Chris Couzens <[email protected]>"]
 description = "Rust Bindings for Tesseract OCR"
 license = "MIT"
+edition = '2018'
+repository = "https://github.com/antimatter15/tesseract-rs"
+keywords = ["OCR", "tesseract"]
+categories = ["external-ffi-bindings", "multimedia::images"]
+links = "tesseract"
+build = "build.rs"
 
 [dependencies]
+leptonica-sys = "0.1.0"
 
 [build-dependencies]
-bindgen = "0.45.0"
+bindgen = "0.49.3"

tesseract-sys/Cargo.toml

@@ -0,0 +1,18 @@
+# tesseract-sys
+Rust bindings for [Tesseract](https://github.com/tesseract-ocr/tesseract)
+
+## Building
+
+This links to the C libraries [leptonica](https://github.com/danbloomberg/leptonica) and tesseract.
+
+On Fedora 30 the additional dependencies can be installed by running:
+
+```bash
+sudo dnf install leptonica-devel tesseract-devel clang
+```
+
+On Termux 2019 (Android, Android on Chromebooks) the additional dependencies can be installed by running:
+
+```bash
+pkg install libclang leptonica-dev tesseract-dev
+```

tesseract-sys/README.md

@@ -7,7 +7,6 @@ fn main() {
     // Tell cargo to tell rustc to link the system tesseract
     // and leptonica shared libraries.
     println!("cargo:rustc-link-lib=tesseract");
-    println!("cargo:rustc-link-lib=lept");
 
     // The bindgen::Builder is the main entry point
     // to bindgen, and lets you build up options for
@@ -17,8 +16,13 @@ fn main() {
         // bindings for.
         .header("wrapper.h")
         .whitelist_function("^Tess.*")
-        .whitelist_function("^pixRead.*")
-        .whitelist_function("pixFreeData")
+        .blacklist_type("Boxa")
+        .blacklist_type("Pix")
+        .blacklist_type("Pixa")
+        .blacklist_type("_IO_FILE")
+        .blacklist_type("_IO_codecvt")
+        .blacklist_type("_IO_marker")
+        .blacklist_type("_IO_wide_data")
         // Finish the builder and generate the bindings.
         .generate()
         // Unwrap the Result and panic on failure.
@@ -29,4 +33,4 @@ fn main() {
     bindings
         .write_to_file(out_path.join("bindings.rs"))
         .expect("Couldn't write bindings!");
-}
+}

tesseract-sys/build.rs

@@ -2,11 +2,14 @@
 #![allow(non_camel_case_types)]
 #![allow(non_snake_case)]
 
+use leptonica_sys::{Boxa, Pix, Pixa, _IO_FILE};
+
 include!(concat!(env!("OUT_DIR"), "/bindings.rs"));
 
 #[cfg(test)]
 mod tests {
     use super::*;
+    use leptonica_sys::{pixFreeData, pixRead};
     use std::ffi::CStr;
     use std::ffi::CString;
     use std::ptr;

tesseract-sys/src/lib.rs

@@ -1,2 +1 @@
-#include <leptonica/allheaders.h>
 #include <tesseract/capi.h>