xberg 1.1.1 → 1.1.3

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: d7be6ba0bd3df9bfdef3eb5586549a1fd63fafd8bd25d0910a842c850458ebb6
-  data.tar.gz: b4bad20fa6c0cd4058a85096ab475b4e5d9c4b1df1a53f3727b30ad54deae2b7
+  metadata.gz: 7dc92c33387f8c104dc35442acb6b524e50ec1d505778458efbf5d000b96d6fe
+  data.tar.gz: 71f22c054071c092d6ac931726185e3045e6dd1cc2a210b5da52c3fcd4a4156e
 SHA512:
-  metadata.gz: 36a2fe155923c2ee3928cd9dcd6d8fbd7a2f5361543df8f07ff6ab72cd9702eae131c6448176f56fd8a89ce8de34d2163779d37ff5cefd0a5640f298b3a1b353
-  data.tar.gz: 5f4a3ebdb856ff7702d0784be2153888c95ec3222101c5d83cda93e3a4ac0c75c830ccb2b7e999ae7c24d22d5a748c9407db99af0f260259cf8c4fe8c81c6cf2
+  metadata.gz: aa897709688e7f2f5d1cc515549daa00c4a7ae4c90e74d0f29a775888349a9266ccdf64d36cacadb3192881465fc55cbdf576531203bddef42722a314c3d6a7b
+  data.tar.gz: ba9b07d78f372cfbec36377e699e2cd79f512551e9a2d313bd31d639d4caff5cf80c31d4c5c6a59a10cdb6e92d924d673cdcbb6579e5e4bcbd5b186c7fd33ae7

data/ext/xberg_rb/native/Cargo.lock

@@ -1501,6 +1501,12 @@ version = "0.8.7"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "773648b94d0e5d620f64f280777445740e61fe701025087ec8b57f45c791888b"
 
+[[package]]
+name = "core_detect"
+version = "1.0.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "7f8f80099a98041a3d1622845c271458a2d73e688351bf3cb999266764b81d48"
+
 [[package]]
 name = "core_maths"
 version = "0.1.1"
@@ -2027,11 +2033,17 @@ dependencies = [
 
 [[package]]
 name = "encoding_rs"
-version = "0.8.35"
+version = "0.8.40"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "75030f3c4f45dafd7586dd6780965a8c7e8e285a5ecb86713e63a79c5b2766f3"
+checksum = "2a7a45518d2863d18aa47f4a0cf9faec2aa4304cc09df5e41299f276b3ad135e"
 dependencies = [
  "cfg-if 1.0.4",
+ "core_detect",
+ "multiversion",
+ "multiversion_no_op",
+ "rustversion",
+ "scopeguard",
+ "simdutf8",
 ]
 
 [[package]]
@@ -4545,6 +4557,34 @@ dependencies = [
  "version_check",
 ]
 
+[[package]]
+name = "multiversion"
+version = "0.8.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "7edb7f0ff51249dfda9ab96b5823695e15a052dc15074c9dbf3d118afaf2c201"
+dependencies = [
+ "multiversion-macros",
+ "target-features",
+]
+
+[[package]]
+name = "multiversion-macros"
+version = "0.8.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "b093064383341eb3271f42e381cb8f10a01459478446953953c75d24bd339fc0"
+dependencies = [
+ "proc-macro2",
+ "quote",
+ "syn 2.0.119",
+ "target-features",
+]
+
+[[package]]
+name = "multiversion_no_op"
+version = "1.0.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "743fb55ba31b18fb1ecef6bdc9aa2743314978ac084044301a7eee33fb99a20d"
+
 [[package]]
 name = "nalgebra"
 version = "0.35.0"
@@ -4674,9 +4714,9 @@ dependencies = [
 
 [[package]]
 name = "nom-exif"
-version = "3.7.0"
+version = "3.8.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "47a28872847c2602f03d141629a212d7a8c3b6bd8983f518862c51aa474594f8"
+checksum = "3217a90c724aa7c99f53b2fc7eec37b63e3e9aa1801ebe031dee75234d9c8339"
 dependencies = [
  "bytes",
  "chrono",
@@ -7458,6 +7498,12 @@ dependencies = [
  "xattr",
 ]
 
+[[package]]
+name = "target-features"
+version = "0.1.6"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "c1bbb9f3c5c463a01705937a24fdabc5047929ac764b2d5b9cf681c1f5041ed5"
+
 [[package]]
 name = "target-lexicon"
 version = "0.13.5"
@@ -9213,9 +9259,9 @@ dependencies = [
 
 [[package]]
 name = "xberg"
-version = "1.1.1"
+version = "1.1.3"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "8d88afb006b949d6de97555cb528437726bc4aa2c641c618d3b1c536828a9804"
+checksum = "b5ceb984af8471cb7d367eba18d687f2fe86f9dc2afa54214bc099b0e3aaa392"
 dependencies = [
  "aes-gcm",
  "ahash",
@@ -9350,9 +9396,9 @@ dependencies = [
 
 [[package]]
 name = "xberg-candle-ocr"
-version = "1.1.1"
+version = "1.1.3"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "73993303311e4441b037b0a1aed64ed40b4f230daf1b0990e68abfc54e625101"
+checksum = "ff0dd685c5fb93f4bd22f6867409410e5652ddc657cef59b8e64152a3c4e1d36"
 dependencies = [
  "base64 0.23.1",
  "candle-core",
@@ -9372,9 +9418,9 @@ dependencies = [
 
 [[package]]
 name = "xberg-gliner"
-version = "1.1.1"
+version = "1.1.3"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "53b8fb46063d23a6609ff2f189289458f756fa8f2e59575e3464753143b845d5"
+checksum = "cce382640b642185dab50b6d7029183cf12f8500c66887e8ccb76150d68a2fb2"
 dependencies = [
  "getrandom 0.2.17",
  "getrandom 0.3.4",
@@ -9388,9 +9434,9 @@ dependencies = [
 
 [[package]]
 name = "xberg-libheif"
-version = "1.1.1"
+version = "1.1.3"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "1b7da27f2e68e5e66c568cedc24892c220ba7b8362fc4b074e32d0f016cca937"
+checksum = "eb7aa698d6ea5f9ef36a0e7f32653b8d57c21e738beb964ba3646c1a4f8150bb"
 dependencies = [
  "cfg-if 1.0.4",
  "enumn",
@@ -9401,9 +9447,9 @@ dependencies = [
 
 [[package]]
 name = "xberg-libwpd"
-version = "1.1.1"
+version = "1.1.3"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "972715f175bc4180c255d76ebdfbbd935adca2a818f818151995feab9c0952e4"
+checksum = "1f01a67c7b74aa18d5850f50d2869888873a493741f188c66ed93b97ca91ebce"
 dependencies = [
  "cc",
  "flate2",
@@ -9416,9 +9462,9 @@ dependencies = [
 
 [[package]]
 name = "xberg-native-pdf"
-version = "1.1.1"
+version = "1.1.3"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "b768494ffc05dcd59a5003ead16e493affd36a741aa34f73f4ba3ef1b9045a7a"
+checksum = "3d2beb4d548a67f0c110a081ebdf9b61e58d4e82e02cdf53bea5a7e8cfcf3c64"
 dependencies = [
  "aes 0.9.3",
  "base64 0.23.1",
@@ -9463,9 +9509,9 @@ dependencies = [
 
 [[package]]
 name = "xberg-paddle-ocr"
-version = "1.1.1"
+version = "1.1.3"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "18ce27e08abe961f04239ed4dbaa789792d9f49bae755336242dedaaff97c6ff"
+checksum = "1cc1aaf00d7895a9d613c0f1e176faaccefc43d4ef9c0c3e5886df9381d3f771"
 dependencies = [
  "geo-clipper",
  "geo-types",
@@ -9480,7 +9526,7 @@ dependencies = [
 
 [[package]]
 name = "xberg-rb"
-version = "1.1.1"
+version = "1.1.3"
 dependencies = [
  "async-trait",
  "crawlberg",
@@ -9496,9 +9542,9 @@ dependencies = [
 
 [[package]]
 name = "xberg-tesseract"
-version = "1.1.1"
+version = "1.1.3"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "2ca506451bb0ef5c3636fb091ca02ee31eec420c42154b3dfebfafdc0986a5c4"
+checksum = "c259d98976ae7e5f06867e595bc41a3cda5335e9ea63cbae8d9028c94181fd7e"
 dependencies = [
  "cc",
  "cmake",

data/ext/xberg_rb/native/Cargo.toml

@@ -1,11 +1,11 @@
 # This file is auto-generated by alef — DO NOT EDIT.
-# alef:hash:1efb91aaf98fa801f534983b9e6216c6d04d660986f75509d467311c3a719c8e
+# alef:hash:6d69d48379ade66e534af835b520570ab6397625974dfe2112b3d5d87f3da54a
 # To regenerate: alef generate
 # To verify freshness: alef verify
 
 [package]
 name = "xberg-rb"
-version = "1.1.1"
+version = "1.1.3"
 edition = "2024"
 license = "MIT"
 description = "High-performance document intelligence library"
@@ -22,15 +22,19 @@ path = "../src/lib.rs"
 crate-type = ["cdylib"]
 
 [features]
-default = ["api", "api-types", "auto-rotate-types", "candle-ocr", "chunking-tokenizers", "classification", "diff", "embedding-presets", "heuristics", "html", "keywords-rake", "keywords-yake", "late-interaction", "late-interaction-presets", "layout-types", "markdown-footnotes", "ocr", "ocr-wasm", "office", "paddle-ocr-types", "pdf", "presets", "quality", "redaction", "reranker", "reranker-presets", "sparse-embedding-presets", "svg", "transcription-types", "tree-sitter", "url-config-types", "url-ingestion", "xml"]
+default = ["api", "api-types", "auto-rotate-types", "candle-ocr", "captioning", "chunking", "chunking-tokenizers", "classification", "diff", "embedding-presets", "embeddings", "formula-recognition", "heuristics", "html", "keywords-rake", "keywords-yake", "late-interaction", "late-interaction-presets", "layout-types", "liter-llm", "markdown-footnotes", "mcp", "mcp-http", "ner-onnx", "ocr", "ocr-wasm", "office", "paddle-ocr-types", "pdf", "presets", "quality", "redaction", "redaction-rehydrate", "reranker", "reranker-presets", "sparse-embedding-presets", "sparse-embeddings", "svg", "tokio-runtime", "transcription", "transcription-types", "tree-sitter", "url-config-types", "url-ingestion", "xml"]
 api = ["xberg/api"]
 api-types = ["xberg/api-types"]
 auto-rotate-types = ["xberg/auto-rotate-types"]
 candle-ocr = ["xberg/candle-ocr"]
+captioning = ["xberg/captioning"]
+chunking = ["xberg/chunking"]
 chunking-tokenizers = ["xberg/chunking-tokenizers"]
 classification = ["xberg/classification"]
 diff = ["xberg/diff"]
 embedding-presets = ["xberg/embedding-presets"]
+embeddings = ["xberg/embeddings"]
+formula-recognition = ["xberg/formula-recognition"]
 heic = ["xberg/heic"]
 heuristics = ["xberg/heuristics"]
 html = ["xberg/html"]
@@ -39,7 +43,11 @@ keywords-yake = ["xberg/keywords-yake"]
 late-interaction = ["xberg/late-interaction"]
 late-interaction-presets = ["xberg/late-interaction-presets"]
 layout-types = ["xberg/layout-types"]
+liter-llm = ["xberg/liter-llm"]
 markdown-footnotes = ["xberg/markdown-footnotes"]
+mcp = ["xberg/mcp"]
+mcp-http = ["xberg/mcp-http"]
+ner-onnx = ["xberg/ner-onnx"]
 ocr = ["xberg/ocr"]
 ocr-wasm = ["xberg/ocr-wasm"]
 office = ["xberg/office"]
@@ -48,10 +56,14 @@ pdf = ["xberg/pdf"]
 presets = ["xberg/presets"]
 quality = ["xberg/quality"]
 redaction = ["xberg/redaction"]
+redaction-rehydrate = ["xberg/redaction-rehydrate"]
 reranker = ["xberg/reranker"]
 reranker-presets = ["xberg/reranker-presets"]
 sparse-embedding-presets = ["xberg/sparse-embedding-presets"]
+sparse-embeddings = ["xberg/sparse-embeddings"]
 svg = ["xberg/svg"]
+tokio-runtime = ["xberg/tokio-runtime"]
+transcription = ["xberg/transcription"]
 transcription-types = ["xberg/transcription-types"]
 tree-sitter = ["xberg/tree-sitter"]
 url-config-types = ["xberg/url-config-types"]
@@ -70,13 +82,13 @@ tokio = { version = "1", features = ["rt-multi-thread"] }
 tracing = "0.1"
 
 [target.'cfg(all(target_os = "macos", target_arch = "x86_64"))'.dependencies]
-xberg = { version = "1.1.1", default-features = false, features = ["macos-intel-target", "simd-utf8"] }
+xberg = { version = "1.1.3", default-features = false, features = ["macos-intel-target", "simd-utf8"] }
 
 [target.'cfg(not(any(target_os = "windows", all(target_os = "macos", target_arch = "x86_64"))))'.dependencies]
-xberg = { version = "1.1.1", features = ["full", "pdf", "ocr", "paddle-ocr", "paddle-ocr-types", "layout-detection", "layout-types", "embeddings", "embedding-presets", "reranker", "reranker-presets", "sparse-embeddings", "sparse-embedding-presets", "late-interaction", "late-interaction-presets", "chunking", "keywords-yake", "keywords-rake", "language-detection", "html", "tree-sitter", "office", "email", "archives", "stopwords", "auto-rotate", "auto-rotate-types", "tokio-runtime", "api", "api-types", "mcp", "liter-llm", "quality", "svg", "transcription", "transcription-types", "classification", "captioning", "ner-onnx", "ner-llm", "diff", "markdown-footnotes", "redaction", "enrichment", "heuristics", "presets", "structured", "url-ingestion"] }
+xberg = { version = "1.1.3", features = ["full", "pdf", "ocr", "paddle-ocr", "paddle-ocr-types", "layout-detection", "layout-types", "embeddings", "embedding-presets", "reranker", "reranker-presets", "sparse-embeddings", "sparse-embedding-presets", "late-interaction", "late-interaction-presets", "chunking", "keywords-yake", "keywords-rake", "language-detection", "html", "tree-sitter", "office", "email", "archives", "stopwords", "auto-rotate", "auto-rotate-types", "tokio-runtime", "api", "api-types", "mcp", "liter-llm", "quality", "svg", "transcription", "transcription-types", "classification", "captioning", "ner-onnx", "ner-llm", "diff", "markdown-footnotes", "redaction", "enrichment", "heuristics", "presets", "structured", "url-ingestion"] }
 
 [target.'cfg(target_os = "windows")'.dependencies]
-xberg = { version = "1.1.1", default-features = false, features = ["windows-gnu-target", "simd-utf8"] }
+xberg = { version = "1.1.3", default-features = false, features = ["windows-gnu-target", "simd-utf8"] }
 
 # This crate deliberately does not use `[lints]` / `workspace = true`: its C-ABI /
 # PyO3 / napi / ext-php-rs / NIF boundary requires `unsafe` code, and the workspace's

data/ext/xberg_rb/src/lib.rs

@@ -1,5 +1,5 @@
 // This file is auto-generated by alef. DO NOT EDIT.
-// alef:hash:2633d345d440ea6f70d180e16fe231f2acf5cd4edb2ac17c3d97ac84ddd59763
+// alef:hash:2708601d3a9f5bb14721ff6d8e7d75689d9ac6f2bb7902b241a407e9a9951ff6
 // Re-generate with: alef generate
 #![allow(dead_code, unused_imports, unused_variables)]
 #![allow(
@@ -14755,6 +14755,8 @@ pub struct GridCell {
     col_span: u32,
     is_header: bool,
     bbox: Option<BoundingBox>,
+    heading_level: Option<u8>,
+    style_name: Option<String>,
 }
 
 unsafe impl IntoValueFromNative for GridCell {}
@@ -14803,6 +14805,8 @@ impl Default for GridCell {
             col_span: 1,
             is_header: false,
             bbox: None,
+            heading_level: None,
+            style_name: None,
         }
     }
 }
@@ -14877,6 +14881,24 @@ impl GridCell {
                 })?),
                 None => None,
             },
+            heading_level: match kwargs.get(ruby.to_symbol("heading_level")) {
+                Some(v) => Some(u8::try_convert(v).map_err(|e| {
+                    magnus::Error::new(
+                        unsafe { magnus::Ruby::get_unchecked() }.exception_type_error(),
+                        format!("invalid value for `heading_level`: {}", e),
+                    )
+                })?),
+                None => None,
+            },
+            style_name: match kwargs.get(ruby.to_symbol("style_name")) {
+                Some(v) => Some(String::try_convert(v).map_err(|e| {
+                    magnus::Error::new(
+                        unsafe { magnus::Ruby::get_unchecked() }.exception_type_error(),
+                        format!("invalid value for `style_name`: {}", e),
+                    )
+                })?),
+                None => None,
+            },
         })
     }
 
@@ -14908,6 +14930,14 @@ impl GridCell {
         self.bbox.clone()
     }
 
+    fn heading_level(&self) -> Option<u8> {
+        self.heading_level
+    }
+
+    fn style_name(&self) -> Option<String> {
+        self.style_name.clone()
+    }
+
     #[allow(clippy::should_implement_trait)]
     fn to_s(&self) -> String {
         self.content.clone()
@@ -31696,6 +31726,7 @@ pub struct Table {
     page_number: u32,
     bounding_box: Option<BoundingBox>,
     table_id: Option<String>,
+    cell_styles: Vec<TableCellStyle>,
     columns: Option<Vec<String>>,
 }
 
@@ -31793,6 +31824,15 @@ impl Table {
                 })?),
                 None => None,
             },
+            cell_styles: match kwargs.get(ruby.to_symbol("cell_styles")) {
+                Some(v) => <Vec<TableCellStyle>>::try_convert(v).map_err(|e| {
+                    magnus::Error::new(
+                        unsafe { magnus::Ruby::get_unchecked() }.exception_type_error(),
+                        format!("invalid value for `cell_styles`: {}", e),
+                    )
+                })?,
+                None => Default::default(),
+            },
             columns: match kwargs.get(ruby.to_symbol("columns")) {
                 Some(v) => Some(<Vec<String>>::try_convert(v).map_err(|e| {
                     magnus::Error::new(
@@ -31825,6 +31865,10 @@ impl Table {
         self.table_id.clone()
     }
 
+    fn cell_styles(&self) -> Vec<TableCellStyle> {
+        self.cell_styles.clone()
+    }
+
     fn columns(&self) -> Option<Vec<String>> {
         self.columns.clone()
     }
@@ -31950,6 +31994,121 @@ impl TableCell {
     }
 }
 
+#[derive(Clone, Debug, serde::Serialize)]
+#[serde(default)]
+#[magnus::wrap(class = "Xberg::TableCellStyle")]
+pub struct TableCellStyle {
+    row: u32,
+    col: u32,
+    heading_level: Option<u8>,
+    style_name: Option<String>,
+}
+
+unsafe impl IntoValueFromNative for TableCellStyle {}
+
+impl magnus::TryConvert for TableCellStyle {
+    fn try_convert(val: magnus::Value) -> Result<Self, magnus::Error> {
+        if let Ok(r) = <&TableCellStyle as magnus::TryConvert>::try_convert(val) {
+            return Ok(r.clone());
+        }
+        let json_str: String = if let Ok(s) = <String as magnus::TryConvert>::try_convert(val) {
+            s
+        } else {
+            val.funcall::<_, _, String>("to_json", ()).map_err(|e| {
+                magnus::Error::new(
+                    unsafe { magnus::Ruby::get_unchecked() }.exception_type_error(),
+                    format!("no implicit conversion into TableCellStyle: {}", e),
+                )
+            })?
+        };
+        serde_json::from_str::<TableCellStyle>(&json_str).map_err(|e| {
+            magnus::Error::new(
+                unsafe { magnus::Ruby::get_unchecked() }.exception_type_error(),
+                format!("failed to deserialize TableCellStyle: {}", e),
+            )
+        })
+    }
+}
+
+unsafe impl TryConvertOwned for TableCellStyle {}
+impl<'de> serde::Deserialize<'de> for TableCellStyle {
+    fn deserialize<D>(deserializer: D) -> ::core::result::Result<Self, D::Error>
+    where
+        D: serde::Deserializer<'de>,
+    {
+        <xberg::TableCellStyle as serde::Deserialize>::deserialize(deserializer).map(Into::into)
+    }
+}
+
+impl Default for TableCellStyle {
+    fn default() -> Self {
+        xberg::TableCellStyle::default().into()
+    }
+}
+
+impl TableCellStyle {
+    fn new(args: &[magnus::Value]) -> Result<Self, magnus::Error> {
+        let ruby = unsafe { magnus::Ruby::get_unchecked() };
+        let args = magnus::scan_args::scan_args::<(), (Option<magnus::RHash>,), (), (), (), ()>(args)?;
+        let (kwargs_opt,) = args.optional;
+        let kwargs = kwargs_opt.unwrap_or_else(|| ruby.hash_new());
+        Ok(Self {
+            row: match kwargs.get(ruby.to_symbol("row")) {
+                Some(v) => u32::try_convert(v).map_err(|e| {
+                    magnus::Error::new(
+                        unsafe { magnus::Ruby::get_unchecked() }.exception_type_error(),
+                        format!("invalid value for `row`: {}", e),
+                    )
+                })?,
+                None => Default::default(),
+            },
+            col: match kwargs.get(ruby.to_symbol("col")) {
+                Some(v) => u32::try_convert(v).map_err(|e| {
+                    magnus::Error::new(
+                        unsafe { magnus::Ruby::get_unchecked() }.exception_type_error(),
+                        format!("invalid value for `col`: {}", e),
+                    )
+                })?,
+                None => Default::default(),
+            },
+            heading_level: match kwargs.get(ruby.to_symbol("heading_level")) {
+                Some(v) => Some(u8::try_convert(v).map_err(|e| {
+                    magnus::Error::new(
+                        unsafe { magnus::Ruby::get_unchecked() }.exception_type_error(),
+                        format!("invalid value for `heading_level`: {}", e),
+                    )
+                })?),
+                None => None,
+            },
+            style_name: match kwargs.get(ruby.to_symbol("style_name")) {
+                Some(v) => Some(String::try_convert(v).map_err(|e| {
+                    magnus::Error::new(
+                        unsafe { magnus::Ruby::get_unchecked() }.exception_type_error(),
+                        format!("invalid value for `style_name`: {}", e),
+                    )
+                })?),
+                None => None,
+            },
+        })
+    }
+
+    fn row(&self) -> u32 {
+        self.row
+    }
+
+    fn col(&self) -> u32 {
+        self.col
+    }
+
+    fn heading_level(&self) -> Option<u8> {
+        self.heading_level
+    }
+
+    fn style_name(&self) -> Option<String> {
+        self.style_name.clone()
+    }
+}
+
 #[cfg(feature = "diff")]
 #[derive(Clone, Debug, serde::Serialize, serde::Deserialize)]
 #[magnus::wrap(class = "Xberg::TableDiff")]
@@ -32141,7 +32300,7 @@ impl TableGrid {
 #[magnus::wrap(class = "Xberg::TesseractConfig")]
 pub struct TesseractConfig {
     language: Vec<String>,
-    psm: i32,
+    psm: Option<i32>,
     output_format: String,
     oem: i32,
     min_confidence: f64,
@@ -32222,13 +32381,13 @@ impl TesseractConfig {
                 None => vec!["eng".to_string()],
             },
             psm: match kwargs.get(ruby.to_symbol("psm")) {
-                Some(v) => i32::try_convert(v).map_err(|e| {
+                Some(v) => Some(i32::try_convert(v).map_err(|e| {
                     magnus::Error::new(
                         unsafe { magnus::Ruby::get_unchecked() }.exception_type_error(),
                         format!("invalid value for `psm`: {}", e),
                     )
-                })?,
-                None => 3,
+                })?),
+                None => None,
             },
             output_format: match kwargs.get(ruby.to_symbol("output_format")) {
                 Some(v) => String::try_convert(v).map_err(|e| {
@@ -32408,7 +32567,7 @@ impl TesseractConfig {
         self.language.clone()
     }
 
-    fn psm(&self) -> i32 {
+    fn psm(&self) -> Option<i32> {
         self.psm
     }
 
@@ -47907,11 +48066,13 @@ impl From<GridCell> for xberg::GridCell {
             col_span: val.col_span,
             is_header: val.is_header,
             bbox: val.bbox.map(Into::into),
+            heading_level: val.heading_level,
+            style_name: val.style_name,
         }
     }
 }
 
-#[allow(clippy::useless_conversion)]
+#[allow(clippy::redundant_closure, clippy::useless_conversion)]
 impl From<xberg::GridCell> for GridCell {
     fn from(val: xberg::GridCell) -> Self {
         Self {
@@ -47922,6 +48083,8 @@ impl From<xberg::GridCell> for GridCell {
             col_span: val.col_span,
             is_header: val.is_header,
             bbox: val.bbox.map(Into::into),
+            heading_level: val.heading_level,
+            style_name: val.style_name.map(|v| v.to_string()),
         }
     }
 }
@@ -51079,6 +51242,7 @@ impl From<Table> for xberg::Table {
             page_number: val.page_number,
             bounding_box: val.bounding_box.map(Into::into),
             table_id: val.table_id,
+            cell_styles: val.cell_styles.into_iter().map(Into::into).collect(),
             columns: val.columns.map(|v| v.into_iter().collect()),
             ..Default::default()
         }
@@ -51094,6 +51258,7 @@ impl From<xberg::Table> for Table {
             page_number: val.page_number,
             bounding_box: val.bounding_box.map(Into::into),
             table_id: val.table_id.map(|v| v.to_string()),
+            cell_styles: val.cell_styles.into_iter().map(Into::into).collect(),
             columns: val.columns.map(|v| v.into_iter().collect()),
         }
     }
@@ -51110,6 +51275,31 @@ impl From<xberg::TableCell> for TableCell {
     }
 }
 
+#[allow(clippy::needless_update)]
+impl From<TableCellStyle> for xberg::TableCellStyle {
+    fn from(val: TableCellStyle) -> Self {
+        Self {
+            row: val.row,
+            col: val.col,
+            heading_level: val.heading_level,
+            style_name: val.style_name,
+            ..Default::default()
+        }
+    }
+}
+
+#[allow(clippy::redundant_closure)]
+impl From<xberg::TableCellStyle> for TableCellStyle {
+    fn from(val: xberg::TableCellStyle) -> Self {
+        Self {
+            row: val.row,
+            col: val.col,
+            heading_level: val.heading_level,
+            style_name: val.style_name.map(|v| v.to_string()),
+        }
+    }
+}
+
 #[cfg(feature = "diff")]
 #[allow(clippy::useless_conversion)]
 impl From<xberg::TableDiff> for TableDiff {
@@ -56276,6 +56466,10 @@ fn ruby_init(ruby: &Ruby) -> Result<(), Error> {
 
     class.define_method("bbox", method!(GridCell::bbox, 0))?;
 
+    class.define_method("heading_level", method!(GridCell::heading_level, 0))?;
+
+    class.define_method("style_name", method!(GridCell::style_name, 0))?;
+
     class.define_method("to_s", method!(GridCell::to_s, 0))?;
 
     let class = module.define_class("HeaderMetadata", ruby.class_object())?;
@@ -58611,6 +58805,8 @@ fn ruby_init(ruby: &Ruby) -> Result<(), Error> {
 
     class.define_method("table_id", method!(Table::table_id, 0))?;
 
+    class.define_method("cell_styles", method!(Table::cell_styles, 0))?;
+
     class.define_method("columns", method!(Table::columns, 0))?;
 
     let class = module.define_class("TableCell", ruby.class_object())?;
@@ -58627,6 +58823,18 @@ fn ruby_init(ruby: &Ruby) -> Result<(), Error> {
 
     class.define_method("to_s", method!(TableCell::to_s, 0))?;
 
+    let class = module.define_class("TableCellStyle", ruby.class_object())?;
+
+    class.define_singleton_method("new", function!(TableCellStyle::new, -1))?;
+
+    class.define_method("row", method!(TableCellStyle::row, 0))?;
+
+    class.define_method("col", method!(TableCellStyle::col, 0))?;
+
+    class.define_method("heading_level", method!(TableCellStyle::heading_level, 0))?;
+
+    class.define_method("style_name", method!(TableCellStyle::style_name, 0))?;
+
     let class = module.define_class("TableDiff", ruby.class_object())?;
 
     #[cfg(feature = "diff")]

data/lib/xberg/native.rb

@@ -1,5 +1,5 @@
 # This file is auto-generated by alef — DO NOT EDIT.
-# alef:hash:95dd25c8cd2698d5ee247c92e046162237595aeaacf861c48f18911ce66ec4cf
+# alef:hash:2ed8d28561d80117ab2d77bb163218ecfac8cd3a943b8656907f3fbc2a23425c
 # To regenerate: alef generate
 # To verify freshness: alef verify
 # frozen_string_literal: true
@@ -5295,15 +5295,23 @@ module Xberg
   ## Try the classical OCR backend first. If the quality score is below
   ## `quality_threshold`, send the page to the VLM.
   ##
-  ## `quality_threshold` is in the `[0.0, 1.0]` range produced by
-  ## `text::quality::calculate_quality_score`. A value of `0.5` is a
-  ## reasonable starting point; calibrate with the Stage 0 benchmark harness.
+  ## `quality_threshold` is in the `[0.0, 1.0]` range, but it is **not** the same
+  ## quantity reported on `score` (GH#1584).
+  ## The accept decision blends text-shape quality with confidence, weighted 0.7/0.3
+  ## (`extractors::pdf::ocr::pipeline_stage_score`) -- when the backend's confidence is on
+  ## a known scale it contributes only 30% of the compared score, so a page can clear this
+  ## threshold on clean-looking text even while its own `PageOcrConfidence.score` reads
+  ## below it. Do not calibrate this value by reading `PageOcrConfidence.score` off a
+  ## sample page and expecting an equal `quality_threshold` to reproduce the same
+  ## accept/reject outcome. A value of `0.5` is a reasonable starting point; calibrate with
+  ## the Stage 0 benchmark harness.
   VlmFallbackPolicyOnLowQuality = Data.define(:quality_threshold) do
     include VlmFallbackPolicy
     extend T::Sig
 
-    # Minimum acceptable quality score from the classical backend.
-    # Pages scoring below this are retried with VLM.
+    # Minimum acceptable quality score from the classical backend. Pages scoring below
+    # this are retried with VLM -- see this variant's doc comment for what "scoring"
+    # means here.
     sig { returns(Float) }
     # rubocop:disable Lint/UselessMethodDefinition
     def quality_threshold = super

data/lib/xberg/version.rb

@@ -1,10 +1,10 @@
 # This file is auto-generated by alef — DO NOT EDIT.
-# alef:hash:71dc9990e02ee56c24327e6bf2b5874c3d7c8a177543612ea54d00c7891b0a16
+# alef:hash:73c70c86b0b71cc6dde88c9c0ee7245420504146c2e2122be9b741ddce361099
 # To regenerate: alef generate
 # To verify freshness: alef verify
 # frozen_string_literal: true
 
 module Xberg
   ## The version string for this package.
-  VERSION = "1.1.1"
+  VERSION = "1.1.3"
 end

data/sig/types.rbs

@@ -1,5 +1,5 @@
 # This file is auto-generated by alef — DO NOT EDIT.
-# alef:hash:f3e7d9702356d5d894258ff5f2a1b2558cabf46397da7c92dbf36f6d9a11f43e
+# alef:hash:983fd32130659037e7cd0d012a6583376687bf78f5f876bf66af9db448c8e3f0
 # To regenerate: alef generate
 # To verify freshness: alef verify
 
@@ -1144,8 +1144,10 @@ def initialize: (block_type: BlockType, ?level: Integer, inline_content: Array[I
         attr_reader col_span: Integer
         attr_reader is_header: bool
         attr_reader bbox: BoundingBox?
+        attr_reader heading_level: Integer?
+        attr_reader style_name: String?
 
-    def initialize: (content: String, row: Integer, col: Integer, row_span: Integer, col_span: Integer, is_header: bool, ?bbox: BoundingBox) -> void
+    def initialize: (content: String, row: Integer, col: Integer, row_span: Integer, col_span: Integer, is_header: bool, ?bbox: BoundingBox, ?heading_level: Integer, ?style_name: String) -> void
     end
 
     class HeaderMetadata
@@ -2452,9 +2454,10 @@ attr_reader cells: Array[Array[String]]
         attr_reader page_number: Integer
         attr_reader bounding_box: BoundingBox?
         attr_reader table_id: String?
+    attr_reader cell_styles: Array[TableCellStyle]
     attr_reader columns: Array[String]?
 
-def initialize: (?cells: Array[Array[String]], ?markdown: String, ?page_number: Integer, ?bounding_box: BoundingBox, ?table_id: String, ?columns: Array[String]) -> void
+def initialize: (?cells: Array[Array[String]], ?markdown: String, ?page_number: Integer, ?bounding_box: BoundingBox, ?table_id: String, ?cell_styles: Array[TableCellStyle], ?columns: Array[String]) -> void
     end
 
     class TableCell
@@ -2466,6 +2469,15 @@ def initialize: (?cells: Array[Array[String]], ?markdown: String, ?page_number:
     def initialize: (?content: String, ?row_span: Integer, ?col_span: Integer, ?is_header: bool) -> void
     end
 
+    class TableCellStyle
+        attr_reader row: Integer
+        attr_reader col: Integer
+        attr_reader heading_level: Integer?
+        attr_reader style_name: String?
+
+    def initialize: (?row: Integer, ?col: Integer, ?heading_level: Integer, ?style_name: String) -> void
+    end
+
     class TableDiff
         attr_reader from_index: Integer
         attr_reader to_index: Integer
@@ -2484,7 +2496,7 @@ def initialize: (?rows: Integer, ?cols: Integer, ?cells: Array[GridCell]) -> voi
 
     class TesseractConfig
     attr_reader language: Array[String]
-        attr_reader psm: Integer
+        attr_reader psm: Integer?
         attr_reader output_format: String
         attr_reader oem: Integer
         attr_reader min_confidence: Float

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: xberg
 version: !ruby/object:Gem::Version
-  version: 1.1.1
+  version: 1.1.3
 platform: ruby
 authors:
 - Na'aman Hirschfeld <naaman@xberg.io>
 autorequire:
 bindir: bin
 cert_chain: []
-date: 2026-09-07 00:00:00.000000000 Z
+date: 2026-09-09 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: rb_sys