kreuzberg 4.0.0.pre.rc.6
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.
- checksums.yaml +7 -0
- data/.gitignore +14 -0
- data/.rspec +3 -0
- data/.rubocop.yaml +1 -0
- data/.rubocop.yml +538 -0
- data/Gemfile +8 -0
- data/Gemfile.lock +157 -0
- data/README.md +426 -0
- data/Rakefile +25 -0
- data/Steepfile +47 -0
- data/examples/async_patterns.rb +341 -0
- data/ext/kreuzberg_rb/extconf.rb +45 -0
- data/ext/kreuzberg_rb/native/Cargo.lock +6535 -0
- data/ext/kreuzberg_rb/native/Cargo.toml +44 -0
- data/ext/kreuzberg_rb/native/README.md +425 -0
- data/ext/kreuzberg_rb/native/build.rs +15 -0
- data/ext/kreuzberg_rb/native/include/ieeefp.h +11 -0
- data/ext/kreuzberg_rb/native/include/msvc_compat/strings.h +14 -0
- data/ext/kreuzberg_rb/native/include/strings.h +20 -0
- data/ext/kreuzberg_rb/native/include/unistd.h +47 -0
- data/ext/kreuzberg_rb/native/src/lib.rs +2998 -0
- data/extconf.rb +28 -0
- data/kreuzberg.gemspec +148 -0
- data/lib/kreuzberg/api_proxy.rb +142 -0
- data/lib/kreuzberg/cache_api.rb +46 -0
- data/lib/kreuzberg/cli.rb +55 -0
- data/lib/kreuzberg/cli_proxy.rb +127 -0
- data/lib/kreuzberg/config.rb +691 -0
- data/lib/kreuzberg/error_context.rb +32 -0
- data/lib/kreuzberg/errors.rb +118 -0
- data/lib/kreuzberg/extraction_api.rb +85 -0
- data/lib/kreuzberg/mcp_proxy.rb +186 -0
- data/lib/kreuzberg/ocr_backend_protocol.rb +113 -0
- data/lib/kreuzberg/post_processor_protocol.rb +86 -0
- data/lib/kreuzberg/result.rb +216 -0
- data/lib/kreuzberg/setup_lib_path.rb +80 -0
- data/lib/kreuzberg/validator_protocol.rb +89 -0
- data/lib/kreuzberg/version.rb +5 -0
- data/lib/kreuzberg.rb +103 -0
- data/sig/kreuzberg/internal.rbs +184 -0
- data/sig/kreuzberg.rbs +520 -0
- data/spec/binding/cache_spec.rb +227 -0
- data/spec/binding/cli_proxy_spec.rb +85 -0
- data/spec/binding/cli_spec.rb +55 -0
- data/spec/binding/config_spec.rb +345 -0
- data/spec/binding/config_validation_spec.rb +283 -0
- data/spec/binding/error_handling_spec.rb +213 -0
- data/spec/binding/errors_spec.rb +66 -0
- data/spec/binding/plugins/ocr_backend_spec.rb +307 -0
- data/spec/binding/plugins/postprocessor_spec.rb +269 -0
- data/spec/binding/plugins/validator_spec.rb +274 -0
- data/spec/fixtures/config.toml +39 -0
- data/spec/fixtures/config.yaml +41 -0
- data/spec/fixtures/invalid_config.toml +4 -0
- data/spec/smoke/package_spec.rb +178 -0
- data/spec/spec_helper.rb +42 -0
- data/vendor/kreuzberg/Cargo.toml +204 -0
- data/vendor/kreuzberg/README.md +175 -0
- data/vendor/kreuzberg/benches/otel_overhead.rs +48 -0
- data/vendor/kreuzberg/build.rs +474 -0
- data/vendor/kreuzberg/src/api/error.rs +81 -0
- data/vendor/kreuzberg/src/api/handlers.rs +199 -0
- data/vendor/kreuzberg/src/api/mod.rs +79 -0
- data/vendor/kreuzberg/src/api/server.rs +353 -0
- data/vendor/kreuzberg/src/api/types.rs +170 -0
- data/vendor/kreuzberg/src/cache/mod.rs +1167 -0
- data/vendor/kreuzberg/src/chunking/mod.rs +677 -0
- data/vendor/kreuzberg/src/core/batch_mode.rs +95 -0
- data/vendor/kreuzberg/src/core/config.rs +1032 -0
- data/vendor/kreuzberg/src/core/extractor.rs +1024 -0
- data/vendor/kreuzberg/src/core/io.rs +329 -0
- data/vendor/kreuzberg/src/core/mime.rs +605 -0
- data/vendor/kreuzberg/src/core/mod.rs +45 -0
- data/vendor/kreuzberg/src/core/pipeline.rs +984 -0
- data/vendor/kreuzberg/src/embeddings.rs +432 -0
- data/vendor/kreuzberg/src/error.rs +431 -0
- data/vendor/kreuzberg/src/extraction/archive.rs +954 -0
- data/vendor/kreuzberg/src/extraction/docx.rs +40 -0
- data/vendor/kreuzberg/src/extraction/email.rs +854 -0
- data/vendor/kreuzberg/src/extraction/excel.rs +688 -0
- data/vendor/kreuzberg/src/extraction/html.rs +553 -0
- data/vendor/kreuzberg/src/extraction/image.rs +368 -0
- data/vendor/kreuzberg/src/extraction/libreoffice.rs +563 -0
- data/vendor/kreuzberg/src/extraction/markdown.rs +213 -0
- data/vendor/kreuzberg/src/extraction/mod.rs +81 -0
- data/vendor/kreuzberg/src/extraction/office_metadata/app_properties.rs +398 -0
- data/vendor/kreuzberg/src/extraction/office_metadata/core_properties.rs +247 -0
- data/vendor/kreuzberg/src/extraction/office_metadata/custom_properties.rs +240 -0
- data/vendor/kreuzberg/src/extraction/office_metadata/mod.rs +130 -0
- data/vendor/kreuzberg/src/extraction/office_metadata/odt_properties.rs +287 -0
- data/vendor/kreuzberg/src/extraction/pptx.rs +3000 -0
- data/vendor/kreuzberg/src/extraction/structured.rs +490 -0
- data/vendor/kreuzberg/src/extraction/table.rs +328 -0
- data/vendor/kreuzberg/src/extraction/text.rs +269 -0
- data/vendor/kreuzberg/src/extraction/xml.rs +333 -0
- data/vendor/kreuzberg/src/extractors/archive.rs +446 -0
- data/vendor/kreuzberg/src/extractors/bibtex.rs +469 -0
- data/vendor/kreuzberg/src/extractors/docbook.rs +502 -0
- data/vendor/kreuzberg/src/extractors/docx.rs +367 -0
- data/vendor/kreuzberg/src/extractors/email.rs +143 -0
- data/vendor/kreuzberg/src/extractors/epub.rs +707 -0
- data/vendor/kreuzberg/src/extractors/excel.rs +343 -0
- data/vendor/kreuzberg/src/extractors/fictionbook.rs +491 -0
- data/vendor/kreuzberg/src/extractors/fictionbook.rs.backup2 +738 -0
- data/vendor/kreuzberg/src/extractors/html.rs +393 -0
- data/vendor/kreuzberg/src/extractors/image.rs +198 -0
- data/vendor/kreuzberg/src/extractors/jats.rs +1051 -0
- data/vendor/kreuzberg/src/extractors/jupyter.rs +367 -0
- data/vendor/kreuzberg/src/extractors/latex.rs +652 -0
- data/vendor/kreuzberg/src/extractors/markdown.rs +700 -0
- data/vendor/kreuzberg/src/extractors/mod.rs +365 -0
- data/vendor/kreuzberg/src/extractors/odt.rs +628 -0
- data/vendor/kreuzberg/src/extractors/opml.rs +634 -0
- data/vendor/kreuzberg/src/extractors/orgmode.rs +528 -0
- data/vendor/kreuzberg/src/extractors/pdf.rs +493 -0
- data/vendor/kreuzberg/src/extractors/pptx.rs +248 -0
- data/vendor/kreuzberg/src/extractors/rst.rs +576 -0
- data/vendor/kreuzberg/src/extractors/rtf.rs +810 -0
- data/vendor/kreuzberg/src/extractors/security.rs +484 -0
- data/vendor/kreuzberg/src/extractors/security_tests.rs +367 -0
- data/vendor/kreuzberg/src/extractors/structured.rs +140 -0
- data/vendor/kreuzberg/src/extractors/text.rs +260 -0
- data/vendor/kreuzberg/src/extractors/typst.rs +650 -0
- data/vendor/kreuzberg/src/extractors/xml.rs +135 -0
- data/vendor/kreuzberg/src/image/dpi.rs +164 -0
- data/vendor/kreuzberg/src/image/mod.rs +6 -0
- data/vendor/kreuzberg/src/image/preprocessing.rs +417 -0
- data/vendor/kreuzberg/src/image/resize.rs +89 -0
- data/vendor/kreuzberg/src/keywords/config.rs +154 -0
- data/vendor/kreuzberg/src/keywords/mod.rs +237 -0
- data/vendor/kreuzberg/src/keywords/processor.rs +267 -0
- data/vendor/kreuzberg/src/keywords/rake.rs +293 -0
- data/vendor/kreuzberg/src/keywords/types.rs +68 -0
- data/vendor/kreuzberg/src/keywords/yake.rs +163 -0
- data/vendor/kreuzberg/src/language_detection/mod.rs +942 -0
- data/vendor/kreuzberg/src/lib.rs +105 -0
- data/vendor/kreuzberg/src/mcp/mod.rs +32 -0
- data/vendor/kreuzberg/src/mcp/server.rs +1968 -0
- data/vendor/kreuzberg/src/ocr/cache.rs +469 -0
- data/vendor/kreuzberg/src/ocr/error.rs +37 -0
- data/vendor/kreuzberg/src/ocr/hocr.rs +216 -0
- data/vendor/kreuzberg/src/ocr/mod.rs +58 -0
- data/vendor/kreuzberg/src/ocr/processor.rs +863 -0
- data/vendor/kreuzberg/src/ocr/table/mod.rs +4 -0
- data/vendor/kreuzberg/src/ocr/table/tsv_parser.rs +144 -0
- data/vendor/kreuzberg/src/ocr/tesseract_backend.rs +450 -0
- data/vendor/kreuzberg/src/ocr/types.rs +393 -0
- data/vendor/kreuzberg/src/ocr/utils.rs +47 -0
- data/vendor/kreuzberg/src/ocr/validation.rs +206 -0
- data/vendor/kreuzberg/src/panic_context.rs +154 -0
- data/vendor/kreuzberg/src/pdf/error.rs +122 -0
- data/vendor/kreuzberg/src/pdf/images.rs +139 -0
- data/vendor/kreuzberg/src/pdf/metadata.rs +346 -0
- data/vendor/kreuzberg/src/pdf/mod.rs +50 -0
- data/vendor/kreuzberg/src/pdf/rendering.rs +369 -0
- data/vendor/kreuzberg/src/pdf/table.rs +393 -0
- data/vendor/kreuzberg/src/pdf/text.rs +158 -0
- data/vendor/kreuzberg/src/plugins/extractor.rs +1013 -0
- data/vendor/kreuzberg/src/plugins/mod.rs +209 -0
- data/vendor/kreuzberg/src/plugins/ocr.rs +620 -0
- data/vendor/kreuzberg/src/plugins/processor.rs +642 -0
- data/vendor/kreuzberg/src/plugins/registry.rs +1337 -0
- data/vendor/kreuzberg/src/plugins/traits.rs +258 -0
- data/vendor/kreuzberg/src/plugins/validator.rs +956 -0
- data/vendor/kreuzberg/src/stopwords/mod.rs +1470 -0
- data/vendor/kreuzberg/src/text/mod.rs +19 -0
- data/vendor/kreuzberg/src/text/quality.rs +697 -0
- data/vendor/kreuzberg/src/text/string_utils.rs +217 -0
- data/vendor/kreuzberg/src/text/token_reduction/cjk_utils.rs +164 -0
- data/vendor/kreuzberg/src/text/token_reduction/config.rs +100 -0
- data/vendor/kreuzberg/src/text/token_reduction/core.rs +796 -0
- data/vendor/kreuzberg/src/text/token_reduction/filters.rs +902 -0
- data/vendor/kreuzberg/src/text/token_reduction/mod.rs +160 -0
- data/vendor/kreuzberg/src/text/token_reduction/semantic.rs +619 -0
- data/vendor/kreuzberg/src/text/token_reduction/simd_text.rs +147 -0
- data/vendor/kreuzberg/src/types.rs +903 -0
- data/vendor/kreuzberg/src/utils/mod.rs +17 -0
- data/vendor/kreuzberg/src/utils/quality.rs +959 -0
- data/vendor/kreuzberg/src/utils/string_utils.rs +381 -0
- data/vendor/kreuzberg/stopwords/af_stopwords.json +53 -0
- data/vendor/kreuzberg/stopwords/ar_stopwords.json +482 -0
- data/vendor/kreuzberg/stopwords/bg_stopwords.json +261 -0
- data/vendor/kreuzberg/stopwords/bn_stopwords.json +400 -0
- data/vendor/kreuzberg/stopwords/br_stopwords.json +1205 -0
- data/vendor/kreuzberg/stopwords/ca_stopwords.json +280 -0
- data/vendor/kreuzberg/stopwords/cs_stopwords.json +425 -0
- data/vendor/kreuzberg/stopwords/da_stopwords.json +172 -0
- data/vendor/kreuzberg/stopwords/de_stopwords.json +622 -0
- data/vendor/kreuzberg/stopwords/el_stopwords.json +849 -0
- data/vendor/kreuzberg/stopwords/en_stopwords.json +1300 -0
- data/vendor/kreuzberg/stopwords/eo_stopwords.json +175 -0
- data/vendor/kreuzberg/stopwords/es_stopwords.json +734 -0
- data/vendor/kreuzberg/stopwords/et_stopwords.json +37 -0
- data/vendor/kreuzberg/stopwords/eu_stopwords.json +100 -0
- data/vendor/kreuzberg/stopwords/fa_stopwords.json +801 -0
- data/vendor/kreuzberg/stopwords/fi_stopwords.json +849 -0
- data/vendor/kreuzberg/stopwords/fr_stopwords.json +693 -0
- data/vendor/kreuzberg/stopwords/ga_stopwords.json +111 -0
- data/vendor/kreuzberg/stopwords/gl_stopwords.json +162 -0
- data/vendor/kreuzberg/stopwords/gu_stopwords.json +226 -0
- data/vendor/kreuzberg/stopwords/ha_stopwords.json +41 -0
- data/vendor/kreuzberg/stopwords/he_stopwords.json +196 -0
- data/vendor/kreuzberg/stopwords/hi_stopwords.json +227 -0
- data/vendor/kreuzberg/stopwords/hr_stopwords.json +181 -0
- data/vendor/kreuzberg/stopwords/hu_stopwords.json +791 -0
- data/vendor/kreuzberg/stopwords/hy_stopwords.json +47 -0
- data/vendor/kreuzberg/stopwords/id_stopwords.json +760 -0
- data/vendor/kreuzberg/stopwords/it_stopwords.json +634 -0
- data/vendor/kreuzberg/stopwords/ja_stopwords.json +136 -0
- data/vendor/kreuzberg/stopwords/kn_stopwords.json +84 -0
- data/vendor/kreuzberg/stopwords/ko_stopwords.json +681 -0
- data/vendor/kreuzberg/stopwords/ku_stopwords.json +64 -0
- data/vendor/kreuzberg/stopwords/la_stopwords.json +51 -0
- data/vendor/kreuzberg/stopwords/lt_stopwords.json +476 -0
- data/vendor/kreuzberg/stopwords/lv_stopwords.json +163 -0
- data/vendor/kreuzberg/stopwords/ml_stopwords.json +1 -0
- data/vendor/kreuzberg/stopwords/mr_stopwords.json +101 -0
- data/vendor/kreuzberg/stopwords/ms_stopwords.json +477 -0
- data/vendor/kreuzberg/stopwords/ne_stopwords.json +490 -0
- data/vendor/kreuzberg/stopwords/nl_stopwords.json +415 -0
- data/vendor/kreuzberg/stopwords/no_stopwords.json +223 -0
- data/vendor/kreuzberg/stopwords/pl_stopwords.json +331 -0
- data/vendor/kreuzberg/stopwords/pt_stopwords.json +562 -0
- data/vendor/kreuzberg/stopwords/ro_stopwords.json +436 -0
- data/vendor/kreuzberg/stopwords/ru_stopwords.json +561 -0
- data/vendor/kreuzberg/stopwords/si_stopwords.json +193 -0
- data/vendor/kreuzberg/stopwords/sk_stopwords.json +420 -0
- data/vendor/kreuzberg/stopwords/sl_stopwords.json +448 -0
- data/vendor/kreuzberg/stopwords/so_stopwords.json +32 -0
- data/vendor/kreuzberg/stopwords/st_stopwords.json +33 -0
- data/vendor/kreuzberg/stopwords/sv_stopwords.json +420 -0
- data/vendor/kreuzberg/stopwords/sw_stopwords.json +76 -0
- data/vendor/kreuzberg/stopwords/ta_stopwords.json +129 -0
- data/vendor/kreuzberg/stopwords/te_stopwords.json +54 -0
- data/vendor/kreuzberg/stopwords/th_stopwords.json +118 -0
- data/vendor/kreuzberg/stopwords/tl_stopwords.json +149 -0
- data/vendor/kreuzberg/stopwords/tr_stopwords.json +506 -0
- data/vendor/kreuzberg/stopwords/uk_stopwords.json +75 -0
- data/vendor/kreuzberg/stopwords/ur_stopwords.json +519 -0
- data/vendor/kreuzberg/stopwords/vi_stopwords.json +647 -0
- data/vendor/kreuzberg/stopwords/yo_stopwords.json +62 -0
- data/vendor/kreuzberg/stopwords/zh_stopwords.json +796 -0
- data/vendor/kreuzberg/stopwords/zu_stopwords.json +31 -0
- data/vendor/kreuzberg/tests/api_extract_multipart.rs +52 -0
- data/vendor/kreuzberg/tests/api_tests.rs +966 -0
- data/vendor/kreuzberg/tests/archive_integration.rs +543 -0
- data/vendor/kreuzberg/tests/batch_orchestration.rs +556 -0
- data/vendor/kreuzberg/tests/batch_processing.rs +316 -0
- data/vendor/kreuzberg/tests/bibtex_parity_test.rs +421 -0
- data/vendor/kreuzberg/tests/concurrency_stress.rs +525 -0
- data/vendor/kreuzberg/tests/config_features.rs +598 -0
- data/vendor/kreuzberg/tests/config_loading_tests.rs +415 -0
- data/vendor/kreuzberg/tests/core_integration.rs +510 -0
- data/vendor/kreuzberg/tests/csv_integration.rs +414 -0
- data/vendor/kreuzberg/tests/docbook_extractor_tests.rs +498 -0
- data/vendor/kreuzberg/tests/docx_metadata_extraction_test.rs +122 -0
- data/vendor/kreuzberg/tests/docx_vs_pandoc_comparison.rs +370 -0
- data/vendor/kreuzberg/tests/email_integration.rs +325 -0
- data/vendor/kreuzberg/tests/epub_native_extractor_tests.rs +275 -0
- data/vendor/kreuzberg/tests/error_handling.rs +393 -0
- data/vendor/kreuzberg/tests/fictionbook_extractor_tests.rs +228 -0
- data/vendor/kreuzberg/tests/format_integration.rs +159 -0
- data/vendor/kreuzberg/tests/helpers/mod.rs +142 -0
- data/vendor/kreuzberg/tests/html_table_test.rs +551 -0
- data/vendor/kreuzberg/tests/image_integration.rs +253 -0
- data/vendor/kreuzberg/tests/instrumentation_test.rs +139 -0
- data/vendor/kreuzberg/tests/jats_extractor_tests.rs +639 -0
- data/vendor/kreuzberg/tests/jupyter_extractor_tests.rs +704 -0
- data/vendor/kreuzberg/tests/keywords_integration.rs +479 -0
- data/vendor/kreuzberg/tests/keywords_quality.rs +509 -0
- data/vendor/kreuzberg/tests/latex_extractor_tests.rs +496 -0
- data/vendor/kreuzberg/tests/markdown_extractor_tests.rs +490 -0
- data/vendor/kreuzberg/tests/mime_detection.rs +428 -0
- data/vendor/kreuzberg/tests/ocr_configuration.rs +510 -0
- data/vendor/kreuzberg/tests/ocr_errors.rs +676 -0
- data/vendor/kreuzberg/tests/ocr_quality.rs +627 -0
- data/vendor/kreuzberg/tests/ocr_stress.rs +469 -0
- data/vendor/kreuzberg/tests/odt_extractor_tests.rs +695 -0
- data/vendor/kreuzberg/tests/opml_extractor_tests.rs +616 -0
- data/vendor/kreuzberg/tests/orgmode_extractor_tests.rs +822 -0
- data/vendor/kreuzberg/tests/pdf_integration.rs +43 -0
- data/vendor/kreuzberg/tests/pipeline_integration.rs +1411 -0
- data/vendor/kreuzberg/tests/plugin_ocr_backend_test.rs +771 -0
- data/vendor/kreuzberg/tests/plugin_postprocessor_test.rs +560 -0
- data/vendor/kreuzberg/tests/plugin_system.rs +921 -0
- data/vendor/kreuzberg/tests/plugin_validator_test.rs +783 -0
- data/vendor/kreuzberg/tests/registry_integration_tests.rs +586 -0
- data/vendor/kreuzberg/tests/rst_extractor_tests.rs +692 -0
- data/vendor/kreuzberg/tests/rtf_extractor_tests.rs +776 -0
- data/vendor/kreuzberg/tests/security_validation.rs +415 -0
- data/vendor/kreuzberg/tests/stopwords_integration_test.rs +888 -0
- data/vendor/kreuzberg/tests/test_fastembed.rs +609 -0
- data/vendor/kreuzberg/tests/typst_behavioral_tests.rs +1259 -0
- data/vendor/kreuzberg/tests/typst_extractor_tests.rs +647 -0
- data/vendor/kreuzberg/tests/xlsx_metadata_extraction_test.rs +87 -0
- data/vendor/rb-sys/.cargo-ok +1 -0
- data/vendor/rb-sys/.cargo_vcs_info.json +6 -0
- data/vendor/rb-sys/Cargo.lock +393 -0
- data/vendor/rb-sys/Cargo.toml +70 -0
- data/vendor/rb-sys/Cargo.toml.orig +57 -0
- data/vendor/rb-sys/LICENSE-APACHE +190 -0
- data/vendor/rb-sys/LICENSE-MIT +21 -0
- data/vendor/rb-sys/bin/release.sh +21 -0
- data/vendor/rb-sys/build/features.rs +108 -0
- data/vendor/rb-sys/build/main.rs +246 -0
- data/vendor/rb-sys/build/stable_api_config.rs +153 -0
- data/vendor/rb-sys/build/version.rs +48 -0
- data/vendor/rb-sys/readme.md +36 -0
- data/vendor/rb-sys/src/bindings.rs +21 -0
- data/vendor/rb-sys/src/hidden.rs +11 -0
- data/vendor/rb-sys/src/lib.rs +34 -0
- data/vendor/rb-sys/src/macros.rs +371 -0
- data/vendor/rb-sys/src/memory.rs +53 -0
- data/vendor/rb-sys/src/ruby_abi_version.rs +38 -0
- data/vendor/rb-sys/src/special_consts.rs +31 -0
- data/vendor/rb-sys/src/stable_api/compiled.c +179 -0
- data/vendor/rb-sys/src/stable_api/compiled.rs +257 -0
- data/vendor/rb-sys/src/stable_api/ruby_2_6.rs +316 -0
- data/vendor/rb-sys/src/stable_api/ruby_2_7.rs +316 -0
- data/vendor/rb-sys/src/stable_api/ruby_3_0.rs +324 -0
- data/vendor/rb-sys/src/stable_api/ruby_3_1.rs +317 -0
- data/vendor/rb-sys/src/stable_api/ruby_3_2.rs +315 -0
- data/vendor/rb-sys/src/stable_api/ruby_3_3.rs +326 -0
- data/vendor/rb-sys/src/stable_api/ruby_3_4.rs +327 -0
- data/vendor/rb-sys/src/stable_api.rs +261 -0
- data/vendor/rb-sys/src/symbol.rs +31 -0
- data/vendor/rb-sys/src/tracking_allocator.rs +332 -0
- data/vendor/rb-sys/src/utils.rs +89 -0
- data/vendor/rb-sys/src/value_type.rs +7 -0
- metadata +536 -0
|
@@ -0,0 +1,258 @@
|
|
|
1
|
+
//! Base plugin trait definition.
|
|
2
|
+
//!
|
|
3
|
+
//! All plugins must implement the `Plugin` trait, which provides basic lifecycle
|
|
4
|
+
//! management and metadata methods.
|
|
5
|
+
|
|
6
|
+
use crate::Result;
|
|
7
|
+
|
|
8
|
+
/// Base trait that all plugins must implement.
|
|
9
|
+
///
|
|
10
|
+
/// This trait provides common functionality for plugin lifecycle management,
|
|
11
|
+
/// identification, and metadata.
|
|
12
|
+
///
|
|
13
|
+
/// # Thread Safety
|
|
14
|
+
///
|
|
15
|
+
/// All plugins must be `Send + Sync` to support concurrent usage across threads.
|
|
16
|
+
///
|
|
17
|
+
/// # Example
|
|
18
|
+
///
|
|
19
|
+
/// ```rust
|
|
20
|
+
/// use kreuzberg::plugins::Plugin;
|
|
21
|
+
/// use kreuzberg::Result;
|
|
22
|
+
/// use std::sync::atomic::{AtomicBool, Ordering};
|
|
23
|
+
///
|
|
24
|
+
/// struct MyPlugin {
|
|
25
|
+
/// initialized: AtomicBool,
|
|
26
|
+
/// }
|
|
27
|
+
///
|
|
28
|
+
/// impl Plugin for MyPlugin {
|
|
29
|
+
/// fn name(&self) -> &str {
|
|
30
|
+
/// "my-plugin"
|
|
31
|
+
/// }
|
|
32
|
+
///
|
|
33
|
+
/// fn version(&self) -> String {
|
|
34
|
+
/// "1.0.0".to_string()
|
|
35
|
+
/// }
|
|
36
|
+
///
|
|
37
|
+
/// fn initialize(&self) -> Result<()> {
|
|
38
|
+
/// self.initialized.store(true, Ordering::Release);
|
|
39
|
+
/// println!("Plugin initialized!");
|
|
40
|
+
/// Ok(())
|
|
41
|
+
/// }
|
|
42
|
+
///
|
|
43
|
+
/// fn shutdown(&self) -> Result<()> {
|
|
44
|
+
/// self.initialized.store(false, Ordering::Release);
|
|
45
|
+
/// println!("Plugin shutdown!");
|
|
46
|
+
/// Ok(())
|
|
47
|
+
/// }
|
|
48
|
+
/// }
|
|
49
|
+
/// ```
|
|
50
|
+
pub trait Plugin: Send + Sync {
|
|
51
|
+
/// Returns the unique name/identifier for this plugin.
|
|
52
|
+
///
|
|
53
|
+
/// The name should be:
|
|
54
|
+
/// - Unique across all plugins
|
|
55
|
+
/// - Lowercase with hyphens (e.g., "my-custom-plugin")
|
|
56
|
+
/// - URL-safe characters only
|
|
57
|
+
///
|
|
58
|
+
/// # Example
|
|
59
|
+
///
|
|
60
|
+
/// ```rust
|
|
61
|
+
/// # use kreuzberg::plugins::Plugin;
|
|
62
|
+
/// # use kreuzberg::Result;
|
|
63
|
+
/// # struct MyPlugin;
|
|
64
|
+
/// # impl Plugin for MyPlugin {
|
|
65
|
+
/// # fn version(&self) -> String { "1.0.0".to_string() }
|
|
66
|
+
/// # fn initialize(&self) -> Result<()> { Ok(()) }
|
|
67
|
+
/// # fn shutdown(&self) -> Result<()> { Ok(()) }
|
|
68
|
+
/// fn name(&self) -> &str {
|
|
69
|
+
/// "pdf-extractor"
|
|
70
|
+
/// }
|
|
71
|
+
/// # }
|
|
72
|
+
/// ```
|
|
73
|
+
fn name(&self) -> &str;
|
|
74
|
+
|
|
75
|
+
/// Returns the semantic version of this plugin.
|
|
76
|
+
///
|
|
77
|
+
/// Should follow semver format: `MAJOR.MINOR.PATCH`
|
|
78
|
+
///
|
|
79
|
+
/// # Example
|
|
80
|
+
///
|
|
81
|
+
/// ```rust
|
|
82
|
+
/// # use kreuzberg::plugins::Plugin;
|
|
83
|
+
/// # use kreuzberg::Result;
|
|
84
|
+
/// # struct MyPlugin;
|
|
85
|
+
/// # impl Plugin for MyPlugin {
|
|
86
|
+
/// # fn name(&self) -> &str { "my-plugin" }
|
|
87
|
+
/// # fn initialize(&self) -> Result<()> { Ok(()) }
|
|
88
|
+
/// # fn shutdown(&self) -> Result<()> { Ok(()) }
|
|
89
|
+
/// fn version(&self) -> String {
|
|
90
|
+
/// "1.2.3".to_string()
|
|
91
|
+
/// }
|
|
92
|
+
/// # }
|
|
93
|
+
/// ```
|
|
94
|
+
fn version(&self) -> String;
|
|
95
|
+
|
|
96
|
+
/// Initialize the plugin.
|
|
97
|
+
///
|
|
98
|
+
/// Called once when the plugin is registered. Use this to:
|
|
99
|
+
/// - Load configuration
|
|
100
|
+
/// - Initialize resources (connections, caches, etc.)
|
|
101
|
+
/// - Validate dependencies
|
|
102
|
+
///
|
|
103
|
+
/// # Thread Safety
|
|
104
|
+
///
|
|
105
|
+
/// This method takes `&self` instead of `&mut self` to work with `Arc<dyn Plugin>`.
|
|
106
|
+
/// Plugins needing mutable state during initialization should use interior mutability
|
|
107
|
+
/// patterns (Mutex, RwLock, OnceCell, etc.).
|
|
108
|
+
///
|
|
109
|
+
/// # Errors
|
|
110
|
+
///
|
|
111
|
+
/// Should return an error if initialization fails. The plugin will not be
|
|
112
|
+
/// registered if this method returns an error.
|
|
113
|
+
///
|
|
114
|
+
/// # Example
|
|
115
|
+
///
|
|
116
|
+
/// ```rust
|
|
117
|
+
/// # use kreuzberg::plugins::Plugin;
|
|
118
|
+
/// # use kreuzberg::Result;
|
|
119
|
+
/// # use std::sync::Mutex;
|
|
120
|
+
/// # struct MyPlugin { config: Mutex<Option<String>> }
|
|
121
|
+
/// # impl Plugin for MyPlugin {
|
|
122
|
+
/// # fn name(&self) -> &str { "my-plugin" }
|
|
123
|
+
/// # fn version(&self) -> String { "1.0.0".to_string() }
|
|
124
|
+
/// # fn shutdown(&self) -> Result<()> { Ok(()) }
|
|
125
|
+
/// fn initialize(&self) -> Result<()> {
|
|
126
|
+
/// // Load configuration using interior mutability
|
|
127
|
+
/// let mut config = self.config.lock().unwrap();
|
|
128
|
+
/// *config = Some("loaded".to_string());
|
|
129
|
+
///
|
|
130
|
+
/// // Perform any initialization work
|
|
131
|
+
/// println!("Plugin initialized successfully");
|
|
132
|
+
///
|
|
133
|
+
/// Ok(())
|
|
134
|
+
/// }
|
|
135
|
+
/// # }
|
|
136
|
+
/// ```
|
|
137
|
+
fn initialize(&self) -> Result<()>;
|
|
138
|
+
|
|
139
|
+
/// Shutdown the plugin.
|
|
140
|
+
///
|
|
141
|
+
/// Called when the plugin is being unregistered or the application is shutting down.
|
|
142
|
+
/// Use this to:
|
|
143
|
+
/// - Close connections
|
|
144
|
+
/// - Flush caches
|
|
145
|
+
/// - Release resources
|
|
146
|
+
///
|
|
147
|
+
/// # Thread Safety
|
|
148
|
+
///
|
|
149
|
+
/// This method takes `&self` instead of `&mut self` to work with `Arc<dyn Plugin>`.
|
|
150
|
+
/// Plugins needing mutable state during shutdown should use interior mutability
|
|
151
|
+
/// patterns (Mutex, RwLock, etc.).
|
|
152
|
+
///
|
|
153
|
+
/// # Errors
|
|
154
|
+
///
|
|
155
|
+
/// Errors during shutdown are logged but don't prevent the shutdown process.
|
|
156
|
+
///
|
|
157
|
+
/// # Example
|
|
158
|
+
///
|
|
159
|
+
/// ```rust
|
|
160
|
+
/// # use kreuzberg::plugins::Plugin;
|
|
161
|
+
/// # use kreuzberg::Result;
|
|
162
|
+
/// # use std::sync::Mutex;
|
|
163
|
+
/// # struct MyPlugin { cache: Mutex<Option<Vec<String>>> }
|
|
164
|
+
/// # impl Plugin for MyPlugin {
|
|
165
|
+
/// # fn name(&self) -> &str { "my-plugin" }
|
|
166
|
+
/// # fn version(&self) -> String { "1.0.0".to_string() }
|
|
167
|
+
/// # fn initialize(&self) -> Result<()> { Ok(()) }
|
|
168
|
+
/// fn shutdown(&self) -> Result<()> {
|
|
169
|
+
/// // Flush caches using interior mutability
|
|
170
|
+
/// let mut cache = self.cache.lock().unwrap();
|
|
171
|
+
/// if let Some(data) = cache.take() {
|
|
172
|
+
/// // Persist cache to disk
|
|
173
|
+
/// }
|
|
174
|
+
///
|
|
175
|
+
/// Ok(())
|
|
176
|
+
/// }
|
|
177
|
+
/// # }
|
|
178
|
+
/// ```
|
|
179
|
+
fn shutdown(&self) -> Result<()>;
|
|
180
|
+
|
|
181
|
+
/// Optional plugin description for debugging and logging.
|
|
182
|
+
///
|
|
183
|
+
/// Defaults to empty string if not overridden.
|
|
184
|
+
fn description(&self) -> &str {
|
|
185
|
+
""
|
|
186
|
+
}
|
|
187
|
+
|
|
188
|
+
/// Optional plugin author information.
|
|
189
|
+
///
|
|
190
|
+
/// Defaults to empty string if not overridden.
|
|
191
|
+
fn author(&self) -> &str {
|
|
192
|
+
""
|
|
193
|
+
}
|
|
194
|
+
}
|
|
195
|
+
|
|
196
|
+
#[cfg(test)]
|
|
197
|
+
mod tests {
|
|
198
|
+
use super::*;
|
|
199
|
+
use std::sync::atomic::{AtomicBool, Ordering};
|
|
200
|
+
|
|
201
|
+
struct TestPlugin {
|
|
202
|
+
initialized: AtomicBool,
|
|
203
|
+
}
|
|
204
|
+
|
|
205
|
+
impl Plugin for TestPlugin {
|
|
206
|
+
fn name(&self) -> &str {
|
|
207
|
+
"test-plugin"
|
|
208
|
+
}
|
|
209
|
+
|
|
210
|
+
fn version(&self) -> String {
|
|
211
|
+
"1.0.0".to_string()
|
|
212
|
+
}
|
|
213
|
+
|
|
214
|
+
fn initialize(&self) -> Result<()> {
|
|
215
|
+
self.initialized.store(true, Ordering::Release);
|
|
216
|
+
Ok(())
|
|
217
|
+
}
|
|
218
|
+
|
|
219
|
+
fn shutdown(&self) -> Result<()> {
|
|
220
|
+
self.initialized.store(false, Ordering::Release);
|
|
221
|
+
Ok(())
|
|
222
|
+
}
|
|
223
|
+
|
|
224
|
+
fn description(&self) -> &str {
|
|
225
|
+
"A test plugin"
|
|
226
|
+
}
|
|
227
|
+
|
|
228
|
+
fn author(&self) -> &str {
|
|
229
|
+
"Test Author"
|
|
230
|
+
}
|
|
231
|
+
}
|
|
232
|
+
|
|
233
|
+
#[test]
|
|
234
|
+
fn test_plugin_metadata() {
|
|
235
|
+
let plugin = TestPlugin {
|
|
236
|
+
initialized: AtomicBool::new(false),
|
|
237
|
+
};
|
|
238
|
+
assert_eq!(plugin.name(), "test-plugin");
|
|
239
|
+
assert_eq!(plugin.version(), "1.0.0");
|
|
240
|
+
assert_eq!(plugin.description(), "A test plugin");
|
|
241
|
+
assert_eq!(plugin.author(), "Test Author");
|
|
242
|
+
}
|
|
243
|
+
|
|
244
|
+
#[test]
|
|
245
|
+
fn test_plugin_lifecycle() {
|
|
246
|
+
let plugin = TestPlugin {
|
|
247
|
+
initialized: AtomicBool::new(false),
|
|
248
|
+
};
|
|
249
|
+
|
|
250
|
+
assert!(!plugin.initialized.load(Ordering::Acquire));
|
|
251
|
+
|
|
252
|
+
plugin.initialize().unwrap();
|
|
253
|
+
assert!(plugin.initialized.load(Ordering::Acquire));
|
|
254
|
+
|
|
255
|
+
plugin.shutdown().unwrap();
|
|
256
|
+
assert!(!plugin.initialized.load(Ordering::Acquire));
|
|
257
|
+
}
|
|
258
|
+
}
|