xberg 1.1.2 → 1.1.4

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
Files changed (7) hide show
  1. checksums.yaml +4 -4
  2. data/ext/xberg_rb/native/Cargo.lock +109 -203
  3. data/ext/xberg_rb/native/Cargo.toml +18 -6
  4. data/ext/xberg_rb/src/lib.rs +984 -591
  5. data/lib/xberg/version.rb +2 -2
  6. data/sig/types.rbs +15 -3
  7. metadata +2 -2
data/ext/xberg_rb/native/Cargo.toml CHANGED
@@ -1,11 +1,11 @@
1
1
  # This file is auto-generated by alef — DO NOT EDIT.
2
- # alef:hash:07aa5bc5a192a77758e42f213f201383ef913b7d438265bf93de7863c4bfbfaf
2
+ # alef:hash:1c66f3f7f6055f927f78a26bd6200308161a6ceac62336347103827902fed583
3
3
  # To regenerate: alef generate
4
4
  # To verify freshness: alef verify
5
5
 
6
6
  [package]
7
7
  name = "xberg-rb"
8
- version = "1.1.2"
8
+ version = "1.1.4"
9
9
  edition = "2024"
10
10
  license = "MIT"
11
11
  description = "High-performance document intelligence library"
@@ -22,15 +22,19 @@ path = "../src/lib.rs"
22
22
  crate-type = ["cdylib"]
23
23
 
24
24
  [features]
25
- 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"]
25
+ 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"]
26
26
  api = ["xberg/api"]
27
27
  api-types = ["xberg/api-types"]
28
28
  auto-rotate-types = ["xberg/auto-rotate-types"]
29
29
  candle-ocr = ["xberg/candle-ocr"]
30
+ captioning = ["xberg/captioning"]
31
+ chunking = ["xberg/chunking"]
30
32
  chunking-tokenizers = ["xberg/chunking-tokenizers"]
31
33
  classification = ["xberg/classification"]
32
34
  diff = ["xberg/diff"]
33
35
  embedding-presets = ["xberg/embedding-presets"]
36
+ embeddings = ["xberg/embeddings"]
37
+ formula-recognition = ["xberg/formula-recognition"]
34
38
  heic = ["xberg/heic"]
35
39
  heuristics = ["xberg/heuristics"]
36
40
  html = ["xberg/html"]
@@ -39,7 +43,11 @@ keywords-yake = ["xberg/keywords-yake"]
39
43
  late-interaction = ["xberg/late-interaction"]
40
44
  late-interaction-presets = ["xberg/late-interaction-presets"]
41
45
  layout-types = ["xberg/layout-types"]
46
+ liter-llm = ["xberg/liter-llm"]
42
47
  markdown-footnotes = ["xberg/markdown-footnotes"]
48
+ mcp = ["xberg/mcp"]
49
+ mcp-http = ["xberg/mcp-http"]
50
+ ner-onnx = ["xberg/ner-onnx"]
43
51
  ocr = ["xberg/ocr"]
44
52
  ocr-wasm = ["xberg/ocr-wasm"]
45
53
  office = ["xberg/office"]
@@ -48,10 +56,14 @@ pdf = ["xberg/pdf"]
48
56
  presets = ["xberg/presets"]
49
57
  quality = ["xberg/quality"]
50
58
  redaction = ["xberg/redaction"]
59
+ redaction-rehydrate = ["xberg/redaction-rehydrate"]
51
60
  reranker = ["xberg/reranker"]
52
61
  reranker-presets = ["xberg/reranker-presets"]
53
62
  sparse-embedding-presets = ["xberg/sparse-embedding-presets"]
63
+ sparse-embeddings = ["xberg/sparse-embeddings"]
54
64
  svg = ["xberg/svg"]
65
+ tokio-runtime = ["xberg/tokio-runtime"]
66
+ transcription = ["xberg/transcription"]
55
67
  transcription-types = ["xberg/transcription-types"]
56
68
  tree-sitter = ["xberg/tree-sitter"]
57
69
  url-config-types = ["xberg/url-config-types"]
@@ -70,13 +82,13 @@ tokio = { version = "1", features = ["rt-multi-thread"] }
70
82
  tracing = "0.1"
71
83
 
72
84
  [target.'cfg(all(target_os = "macos", target_arch = "x86_64"))'.dependencies]
73
- xberg = { version = "1.1.2", default-features = false, features = ["macos-intel-target", "simd-utf8"] }
85
+ xberg = { version = "1.1.4", default-features = false, features = ["macos-intel-target", "simd-utf8"] }
74
86
 
75
87
  [target.'cfg(not(any(target_os = "windows", all(target_os = "macos", target_arch = "x86_64"))))'.dependencies]
76
- xberg = { version = "1.1.2", 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"] }
88
+ xberg = { version = "1.1.4", 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"] }
77
89
 
78
90
  [target.'cfg(target_os = "windows")'.dependencies]
79
- xberg = { version = "1.1.2", default-features = false, features = ["windows-gnu-target", "simd-utf8"] }
91
+ xberg = { version = "1.1.4", default-features = false, features = ["windows-gnu-target", "simd-utf8"] }
80
92
 
81
93
  # This crate deliberately does not use `[lints]` / `workspace = true`: its C-ABI /
82
94
  # PyO3 / napi / ext-php-rs / NIF boundary requires `unsafe` code, and the workspace's