tesseract_ocr_static/
types.rs

1use core::ffi::CStr;
2use core::ptr::NonNull;
3use std::os::raw::c_char;
4
5macro_rules! define_enum {
6    ($enum: ident
7     $doc: literal
8     $(($variant: ident $value: ident $variant_doc: literal $($deprecated: literal)*))+) => {
9        #[doc = $doc]
10        #[derive(Debug, Clone, Copy, PartialEq, Eq, PartialOrd, Ord, Hash)]
11        pub enum $enum {
12            $(
13                #[doc = $variant_doc]
14                $(#[deprecated = $deprecated])*
15                $variant = c::$value as isize,
16            )+
17        }
18
19        impl $enum {
20            #[allow(unused)]
21            #[allow(deprecated)]
22            pub(crate) fn from_raw(value: u32) -> Self {
23                match value {
24                    $(c::$value => Self::$variant,)+
25                    _ => panic!(concat!("Invalid ", stringify!($enum), " value")),
26                }
27            }
28        }
29    };
30}
31
32define_enum! {
33    LayoutLevel "Element of layout hierarchy"
34    (Block TessPageIteratorLevel_RIL_BLOCK "Block of text/image/separator line.")
35    (Paragraph TessPageIteratorLevel_RIL_PARA "Paragraph within a block.")
36    (TextLine TessPageIteratorLevel_RIL_TEXTLINE "Line within a paragraph.")
37    (Word TessPageIteratorLevel_RIL_WORD "Word within a textline.")
38    (Symbol TessPageIteratorLevel_RIL_SYMBOL "Symbol/character within a word.")
39}
40
41define_enum! {
42    OcrEngineMode "OCR engine mode"
43    (TesseractOnly TessOcrEngineMode_OEM_TESSERACT_ONLY "Run Tesseract only." "Use LstmOnly instead.")
44    (LstmOnly TessOcrEngineMode_OEM_LSTM_ONLY "Run just the LSTM line recognizer.")
45    (TesseractLstmCombined TessOcrEngineMode_OEM_TESSERACT_LSTM_COMBINED
46     "Run LSTM, fall back to Tesseract when things get difficult." "Use LstmOnly instead.")
47    (Default TessOcrEngineMode_OEM_DEFAULT "Choose mode automatically based on the configuration.")
48}
49
50define_enum! {
51    PageSegmentationMode "Page layout analysis modes."
52    (OsdOnly TessPageSegMode_PSM_OSD_ONLY "Orientation and script detection only.")
53    (AutoOsd TessPageSegMode_PSM_AUTO_OSD "Automatic page segmentation with orientation and script detection (OSD).")
54    (AutoOnly TessPageSegMode_PSM_AUTO_ONLY "Automatic page segmentation, but no OSD, or OCR.")
55    (Auto TessPageSegMode_PSM_AUTO "Fully automatic page segmentation, but no OSD.")
56    (SingleColumn TessPageSegMode_PSM_SINGLE_COLUMN "Assume a single column of text of variable sizes.")
57    (SingleBlockVertText TessPageSegMode_PSM_SINGLE_BLOCK_VERT_TEXT "Assume a single uniform block of vertically aligned text.")
58    (SingleBlock TessPageSegMode_PSM_SINGLE_BLOCK "Assume a single uniform block of text (default).")
59    (SingleLine TessPageSegMode_PSM_SINGLE_LINE "Treat the image as a single text line.")
60    (SingleWord TessPageSegMode_PSM_SINGLE_WORD "Treat the image as a single word.")
61    (CircleWord TessPageSegMode_PSM_CIRCLE_WORD "Treat the image as a single word in a circle.")
62    (SingleChar TessPageSegMode_PSM_SINGLE_CHAR "Treat the image as a single character.")
63    (SparseText TessPageSegMode_PSM_SPARSE_TEXT "Find as much text as possible in no particular order.")
64    (SparseTextOsd TessPageSegMode_PSM_SPARSE_TEXT_OSD "Sparse text with orientation and script det.")
65    (RawLine TessPageSegMode_PSM_RAW_LINE "Treat the image as a single text line, bypassing hacks that are Tesseract-specific.")
66}
67
68#[allow(clippy::derivable_impls)]
69impl Default for PageSegmentationMode {
70    fn default() -> Self {
71        Self::SingleBlock
72    }
73}
74
75define_enum! {
76    BlockType "Block type."
77    (Unknown TessPolyBlockType_PT_UNKNOWN "Type is not yet known.")
78    (FlowingText TessPolyBlockType_PT_FLOWING_TEXT "Text that lives inside a column.")
79    (HeadingText TessPolyBlockType_PT_HEADING_TEXT "Text that spans more than one column.")
80    (PulloutText TessPolyBlockType_PT_PULLOUT_TEXT "Text that is in a cross-column pull-out region.")
81    (Equation TessPolyBlockType_PT_EQUATION "Partition belonging to an equation region.")
82    (InlineEquation TessPolyBlockType_PT_INLINE_EQUATION "Partition has inline equation.")
83    (Table TessPolyBlockType_PT_TABLE "Partition belonging to a table region.")
84    (VerticalText TessPolyBlockType_PT_VERTICAL_TEXT "Text-line runs vertically.")
85    (CaptionText TessPolyBlockType_PT_CAPTION_TEXT "Text that belongs to an image.")
86    (FlowingImage TessPolyBlockType_PT_FLOWING_IMAGE "Image that lives inside a column.")
87    (HeadingImage TessPolyBlockType_PT_HEADING_IMAGE "Image that spans more than one column.")
88    (PulloutImage TessPolyBlockType_PT_PULLOUT_IMAGE "Image that is in a cross-column pull-out region.")
89    (HorzLine TessPolyBlockType_PT_HORZ_LINE "Horizontal line.")
90    (VertLine TessPolyBlockType_PT_VERT_LINE "Vertical line.")
91    (Noise TessPolyBlockType_PT_NOISE "Lies outside of any column.")
92}
93
94define_enum! {
95    Orientation "Text position on the page."
96    (Up TessOrientation_ORIENTATION_PAGE_UP "Up")
97    (Right TessOrientation_ORIENTATION_PAGE_RIGHT "Right")
98    (Down TessOrientation_ORIENTATION_PAGE_DOWN "Down")
99    (Left TessOrientation_ORIENTATION_PAGE_LEFT "Left")
100}
101
102define_enum! {
103    WritingDirection "Text writing direction."
104    (LeftToRight TessWritingDirection_WRITING_DIRECTION_LEFT_TO_RIGHT "Left-to-right")
105    (RightToLeft TessWritingDirection_WRITING_DIRECTION_RIGHT_TO_LEFT "Right-to-left")
106    (TopToBottom TessWritingDirection_WRITING_DIRECTION_TOP_TO_BOTTOM "Top-to-bottom")
107}
108
109define_enum! {
110    TextlineOrder "The order of the lines of text."
111    (LeftToRight TessTextlineOrder_TEXTLINE_ORDER_LEFT_TO_RIGHT "Left-to-right")
112    (RightToLeft TessTextlineOrder_TEXTLINE_ORDER_RIGHT_TO_LEFT "Right-to-left")
113    (TopToBottom TessTextlineOrder_TEXTLINE_ORDER_TOP_TO_BOTTOM "Top-to-bottom")
114}
115
116define_enum! {
117    ParagraphJustification "Paragraph alignment."
118    (Unknown TessParagraphJustification_JUSTIFICATION_UNKNOWN "The alignment is unclear.")
119    (Left TessParagraphJustification_JUSTIFICATION_LEFT
120     "Each line, except possibly the first, is flush to the same left tab stop.")
121    (Center TessParagraphJustification_JUSTIFICATION_CENTER
122     "The text lines of the paragraph are centered about a line going down through their middle of the text lines.")
123    (Right TessParagraphJustification_JUSTIFICATION_RIGHT
124     "Each line, except possibly the first, is flush to the same right tab stop.")
125}
126
127/// Rectangle.
128///
129/// Coordinate system has the origin at left-top.
130#[derive(Debug, Clone, PartialEq, Eq)]
131pub struct Rectangle {
132    pub left: u32,
133    pub top: u32,
134    pub width: u32,
135    pub height: u32,
136}
137
138/// Line.
139///
140/// Coordinate system has the origin at left-top.
141#[derive(Debug, Clone, PartialEq, Eq)]
142pub struct Line {
143    pub x1: u32,
144    pub y1: u32,
145    pub x2: u32,
146    pub y2: u32,
147}
148
149/// Orientation.
150#[derive(Debug, Clone)]
151pub struct OrientationParams {
152    /// Text position on the page.
153    pub orientation: Orientation,
154    /// How text is aligned.
155    pub writing_direction: WritingDirection,
156    /// How lines are aligned.
157    pub textline_order: TextlineOrder,
158    /// Counter-clockwise rotation angle in radians; the range is _[π/4; π/4]_.
159    ///
160    /// Rotation the block by this angle makes the text upright.
161    pub deskew_angle: f32,
162}
163
164/// Paragraph information.
165#[derive(Debug, Clone)]
166pub struct ParagraphInfo {
167    /// Paragraph alignment.
168    pub justification: ParagraphJustification,
169    /// Whether we believe this is a member of an ordered or unordered list or not.
170    pub is_list_item: bool,
171    /// Whether the first line of the paragraph is aligned with the other
172    /// lines of the paragraph even though subsequent paragraphs have first
173    /// line indents.  This typically indicates that this is the continuation
174    /// of a previous paragraph or that it is the very first paragraph in
175    /// the chapter.
176    pub is_crown: bool,
177    /// First line indentation with respect to `justification`.
178    ///
179    /// Can be negative.
180    pub first_line_indent: i32,
181}
182
183/// Font attributes.
184#[derive(Debug, Clone)]
185pub struct FontAttrs {
186    pub is_bold: bool,
187    pub is_italic: bool,
188    pub is_underlined: bool,
189    pub is_monospace: bool,
190    pub is_serif: bool,
191    pub is_smallcaps: bool,
192    pub point_size: u32,
193    pub font_id: i32,
194}
195
196/// Tesseract-specific string.
197pub struct Text {
198    pub(crate) ptr: NonNull<c_char>,
199}
200
201impl Text {
202    /// Returns text as C string.
203    pub fn as_c_str(&self) -> &CStr {
204        unsafe { CStr::from_ptr(self.ptr.as_ptr()) }
205    }
206}
207
208impl Drop for Text {
209    fn drop(&mut self) {
210        unsafe { c::TessDeleteText(self.ptr.as_ptr()) };
211    }
212}
213
214impl AsRef<CStr> for Text {
215    fn as_ref(&self) -> &CStr {
216        self.as_c_str()
217    }
218}
219
220impl core::fmt::Debug for Text {
221    fn fmt(&self, f: &mut core::fmt::Formatter) -> core::fmt::Result {
222        core::fmt::Debug::fmt(self.as_c_str(), f)
223    }
224}
225
226/// Tesseract-specific UTF-8 string.
227pub struct Utf8Text(pub(crate) Text);
228
229impl Utf8Text {
230    /// Returns text as UTF-8 string.
231    pub fn as_str(&self) -> &str {
232        unsafe { core::str::from_utf8_unchecked(self.0.as_c_str().to_bytes()) }
233    }
234}
235
236impl AsRef<str> for Utf8Text {
237    fn as_ref(&self) -> &str {
238        self.as_str()
239    }
240}
241
242impl core::fmt::Debug for Utf8Text {
243    fn fmt(&self, f: &mut core::fmt::Formatter) -> core::fmt::Result {
244        core::fmt::Debug::fmt(self.as_str(), f)
245    }
246}
tesseract_ocr_static/types.rs

tesseract_ocr_static/
types.rs
