[rust] add parser

.github/workflows/build_mac.yml

@@ -74,22 +74,6 @@ jobs:
       working-directory: build
     - name: Display version information
       run: ./build/ccextractor --version
-  cmake_ocr_hardsubx:
-    runs-on: macos-latest
-    steps:
-    - uses: actions/checkout@v4
-    - name: Install dependencies
-      run: brew install pkg-config autoconf automake libtool tesseract leptonica gpac ffmpeg
-    - name: cmake
-      run: |
-        mkdir build && cd build
-        cmake -DWITH_OCR=ON -DWITH_HARDSUBX=ON ../src
-    - name: build
-      run: |
-        make -j$(nproc)
-      working-directory: build
-    - name: Display version information
-      run: ./build/ccextractor --version
   build_rust:
     runs-on: macos-latest
     steps:

.github/workflows/format.yml

@@ -51,4 +51,4 @@ jobs:
         run: cargo fmt --all -- --check
       - name: clippy
         run: |
-          cargo clippy --all-features -- -D warnings
+          cargo clippy -- -D warnings

.github/workflows/test_rust.yml

@@ -0,0 +1,37 @@
+name: Unit Test Rust
+on:
+  push:
+    paths:
+      - ".github/workflows/test.yml"
+      - "src/rust/**"
+    tags-ignore:
+      - "*.*"
+  pull_request:
+    types: [opened, synchronize, reopened]
+    paths:
+      - ".github/workflows/test.yml"
+      - "src/rust/**"
+jobs:
+  test_rust:
+    runs-on: ubuntu-latest
+    defaults:
+      run:
+        working-directory: ./src/rust
+    steps:
+      - uses: actions/checkout@v4
+      - name: cache
+        uses: actions/cache@v3
+        with:
+          path: |
+            src/rust/.cargo/registry
+            src/rust/.cargo/git
+            src/rust/target
+          key: ${{ runner.os }}-cargo-${{ hashFiles('**/Cargo.lock') }}
+          restore-keys: ${{ runner.os }}-cargo-
+      - uses: actions-rs/toolchain@v1
+        with:
+          toolchain: stable
+          override: true
+      - name: Test main module
+        run: cargo test
+        working-directory: ./src/rust

.gitignore

@@ -17,8 +17,10 @@ CVS
 mac/ccextractor
 linux/ccextractor
 linux/depend
+windows/x86_64-pc-windows-msvc/**
 windows/Debug/**
 windows/Debug-OCR/**
+windows/release-with-debug/**
 windows/Release/**
 windows/Release-Full/**
 windows/Release-OCR/**
@@ -154,3 +156,4 @@ windows/ccx_rust.lib
 windows/*/debug/*
 windows/*/CACHEDIR.TAG
 windows/.rustc_info.json
+linux/configure~

docs/CHANGES.TXT

@@ -1,5 +1,6 @@
-0.95 (to be released)
+1.0 (to be released)
 -----------------
+- Breaking: Major argument flags revamp for CCExtractor (#1564 & #1619)
 - New: Create a Docker image to simplify the CCExtractor usage without any environmental hustle (#1611)
 - New: Add time units module in lib_ccxr (#1623)
 - New: Add bits and levenshtein module in lib_ccxr (#1627)

docs/FFMPEG.md

@@ -42,7 +42,7 @@ Note:If you installed ffmpeg on non-standard location, please change/update your
 
 ### On Windows:
 #### Set preprocessor flag `ENABLE_FFMPEG=1`
-1. In visual studio 2013 right click <Project> and select property.
+1. In visual studio 2022 right click <Project> and select property.
 2. In the left panel, select Configuration Properties, C/C++, Preprocessor.
 3. In the right panel, in the right-hand column of the Preprocessor Definitions property, open the drop-down menu and choose Edit.
 4. In the Preprocessor Definitions dialog box, add `ENABLE_FFMPEG=1`. Choose OK to save your changes.

docs/OCR.md

@@ -93,15 +93,15 @@ Download prebuild library of leptonica and tesseract from following link
 https://drive.google.com/file/d/0B2ou7ZfB-2nZOTRtc3hJMHBtUFk/view?usp=sharing  
 
 put the path of libs/include of leptonica and tesseract in library paths.  
-1. In visual studio 2013 right click <Project> and select property.
+1. In visual studio 2022 right click <Project> and select property.
 2. Select Configuration properties in left panel(column) of property.
 3. Select VC++ Directory.
 4. In the right pane, in the right-hand column of the VC++ Directory property, open the drop-down menu and choose Edit.
 5. Add path of Directory where you have kept uncompressed library of leptonica and tesseract.
 
 
 Set preprocessor flag ENABLE_OCR=1  
-1. In visual studio 2013 right click <Project> and select property.
+1. In visual studio 2022 right click <Project> and select property.
 2. In the left panel, select Configuration Properties, C/C++, Preprocessor.
 3. In the right panel, in the right-hand column of the Preprocessor Definitions property, open the drop-down menu and choose Edit.
 4. In the Preprocessor Definitions dialog box, add ENABLE_OCR=1. Choose OK to save your changes.

mac/Makefile.am

@@ -249,6 +249,7 @@ GPAC_CPPFLAGS = $(shell pkg-config --cflags gpac)
 
 ccextractor_CPPFLAGS =-I../src/lib_ccx/ -I../src/thirdparty/libpng/ -I../src/thirdparty/zlib/ -I../src/lib_ccx/zvbi/ -I../src/thirdparty/lib_hash/ -I../src/thirdparty/protobuf-c/ -I../src/thirdparty -I../src/ -I../src/thirdparty/freetype/include/
 ccextractor_CPPFLAGS += $(GPAC_CPPFLAGS)
+ccextractor_CPPFLAGS += $(FFMPEG_CPPFLAGS)
 
 ccextractor_LDADD=-lm -lpthread -ldl
 
@@ -271,7 +272,7 @@ if HARDSUBX_IS_ENABLED
 ccextractor_CFLAGS += -DENABLE_HARDSUBX
 ccextractor_CPPFLAGS+= ${libavcodec_CFLAGS}
 ccextractor_CPPFLAGS+= ${libavformat_CFLAGS}
-ccextractor_CPPFLAGS+= ${libavutil_CFALGS}
+ccextractor_CPPFLAGS+= ${libavutil_CFLAGS}
 ccextractor_CPPFLAGS+= ${libswscale_CFLAGS}
 AV_LIB = ${libavcodec_LIBS}
 AV_LIB += ${libavformat_LIBS}

src/ccextractor.c

@@ -446,7 +446,11 @@ int main(int argc, char *argv[])
 	// If "ccextractor.cnf" is present, takes options from it.
 	// See docs/ccextractor.cnf.sample for more info.
 
+#ifndef DISABLE_RUST
+	int compile_ret = ccxr_parse_parameters(api_options, argc, argv);
+#else
 	int compile_ret = parse_parameters(api_options, argc, argv);
+#endif
 
 	if (compile_ret == EXIT_NO_INPUT_FILES)
 	{

src/lib_ccx/lib_ccx.h

@@ -161,6 +161,9 @@ extern void ccxr_init_basic_logger(struct ccx_s_options *opts);
 void print_end_msg(void);
 
 //params.c
+#ifndef DISABLE_RUST
+extern int ccxr_parse_parameters(struct ccx_s_options *opt, int argc, char *argv[]);
+#endif
 int parse_parameters (struct ccx_s_options *opt, int argc, char *argv[]);
 void print_usage (void);
 int atoi_hex (char *s);

src/lib_ccx/matroska.c

@@ -1362,8 +1362,8 @@ int matroska_loop(struct lib_ccx_ctx *ctx)
 {
 	if (ccx_options.write_format_rewritten)
 	{
-		mprint(MATROSKA_WARNING "You are using -out=<format>, but Matroska parser extract subtitles in a recorded format\n");
-		mprint("-out=<format> will be ignored\n");
+		mprint(MATROSKA_WARNING "You are using --out=<format>, but Matroska parser extract subtitles in a recorded format\n");
+		mprint("--out=<format> will be ignored\n");
 	}
 
 	// Don't need generated input file

src/lib_ccx/params.c

@@ -127,6 +127,13 @@ int parsedelay(struct ccx_s_options *opt, char *par)
 	return 0;
 }
 
+void set_binary_mode()
+{
+#ifdef WIN32
+	setmode(fileno(stdin), O_BINARY);
+#endif
+}
+
 int append_file_to_queue(struct ccx_s_options *opt, char *filename)
 {
 	if (filename[0] == '\0') // skip files with empty file name (ex : ./ccextractor "")
@@ -978,14 +985,14 @@ void print_usage(void)
 	mprint("  a .d extension. Each .png file will contain an image representing one caption\n");
 	mprint("  and named subNNNN.png, starting with sub0000.png.\n");
 	mprint("  For example, the command:\n");
-	mprint("      ccextractor -out=spupng input.mpg\n");
+	mprint("      ccextractor --out=spupng input.mpg\n");
 	mprint("  will create the files:\n");
 	mprint("      input.xml\n");
 	mprint("      input.d/sub0000.png\n");
 	mprint("      input.d/sub0001.png\n");
 	mprint("      ...\n");
 	mprint("  The command:\n");
-	mprint("      ccextractor -out=spupng -o /tmp/output --12 input.mpg\n");
+	mprint("      ccextractor --out=spupng -o /tmp/output --12 input.mpg\n");
 	mprint("  will create the files:\n");
 	mprint("      /tmp/output_1.xml\n");
 	mprint("      /tmp/output_1.d/sub0000.png\n");
@@ -1245,9 +1252,8 @@ int parse_parameters(struct ccx_s_options *opt, int argc, char *argv[])
 		}
 		if (strcmp(argv[i], "-") == 0 || strcmp(argv[i], "--stdin") == 0)
 		{
-#ifdef WIN32
-			setmode(fileno(stdin), O_BINARY);
-#endif
+			set_binary_mode();
+
 			opt->input_source = CCX_DS_STDIN;
 			if (!opt->live_stream)
 				opt->live_stream = -1;
@@ -2934,7 +2940,7 @@ int parse_parameters(struct ccx_s_options *opt, int argc, char *argv[])
 	}
 	if (opt->write_format == CCX_OF_SPUPNG && opt->cc_to_stdout)
 	{
-		print_error(opt->gui_mode_reports, "You cannot use -out=spupng with -stdout.\n");
+		print_error(opt->gui_mode_reports, "You cannot use --out=spupng with -stdout.\n");
 		return EXIT_INCOMPATIBLE_PARAMETERS;
 	}
 

src/lib_ccx/ts_tables.c

@@ -332,7 +332,7 @@ int parse_PMT(struct ccx_demuxer *ctx, unsigned char *buf, int len, struct progr
 #ifndef ENABLE_OCR
 					if (ccx_options.write_format != CCX_OF_SPUPNG)
 					{
-						mprint("DVB subtitles detected, OCR subsystem not present. Use -out=spupng for graphic output\n");
+						mprint("DVB subtitles detected, OCR subsystem not present. Use --out=spupng for graphic output\n");
 						continue;
 					}
 #endif