[Feature] C++ Extension: Introduce DocNode, TextSplitterBase, SentenseSplitter
#1022
There are no files selected for viewing
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
|
|
@@ -748,30 +748,103 @@ jobs: | |
| export LAZYLLM_K8S_CONFIG_PATH="/mnt/nfs_share/k8s_config.yaml" | ||
| export LAZYLLM_HOME="${{ env.K8S_CI_PATH }}/${{ github.run_id }}-${{ github.job }}" | ||
| mkdir -p $LAZYLLM_HOME | ||
| source /mnt/nfs_share/env.sh | ||
| pytest --lf --last-failed-no-failures=all --durations=0 --reruns=2 -v tests/k8s_tests | ||
|
|
||
| cpp_ext_test: | ||
| name: C++ Build + Python Regression (${{ matrix.os }}) | ||
| needs: [ clone ] | ||
| if: always() | ||
| runs-on: ${{ matrix.os }} | ||
| strategy: | ||
| fail-fast: false | ||
| matrix: | ||
| os: [ubuntu-latest, macos-latest, windows-latest] | ||
| timeout-minutes: 120 | ||
| defaults: | ||
| run: | ||
| shell: bash | ||
| env: | ||
| PYTHONNOUSERSITE: "1" | ||
| PYTHONPATH: ${{ github.workspace }} | ||
| LAZYLLM_ENABLE_CPP_OVERRIDE: "1" | ||
| LAZYLLM_EXPECTED_LOG_MODULES: "all" | ||
| LAZYLLM_DEFAULT_LAUNCHER: "empty" | ||
| LAZYLLM_OPENAI_API_KEY: ${{ secrets.LAZYLLM_OPENAI_API_KEY }} | ||
| LAZYLLM_KIMI_API_KEY: ${{ secrets.LAZYLLM_KIMI_API_KEY }} | ||
| LAZYLLM_AIPING_API_KEY: ${{ secrets.LAZYLLM_AIPING_API_KEY }} | ||
| LAZYLLM_GLM_API_KEY: ${{ secrets.LAZYLLM_GLM_API_KEY }} | ||
| LAZYLLM_GLM_MODEL_NAME: ${{ secrets.LAZYLLM_GLM_MODEL_NAME }} | ||
| LAZYLLM_QWEN_API_KEY: ${{ secrets.LAZYLLM_QWEN_API_KEY }} | ||
| LAZYLLM_QWEN_MODEL_NAME: ${{ secrets.LAZYLLM_QWEN_MODEL_NAME }} | ||
| LAZYLLM_QWEN_TEXT2IMAGE_MODEL_NAME: ${{ secrets.LAZYLLM_QWEN_TEXT2IMAGE_MODEL_NAME }} | ||
| LAZYLLM_SENSENOVA_API_KEY: ${{ secrets.LAZYLLM_SENSENOVA_API_KEY }} | ||
| LAZYLLM_SENSENOVA_SECRET_KEY: ${{ secrets.LAZYLLM_SENSENOVA_SECRET_KEY }} | ||
| LAZYLLM_DOUBAO_API_KEY: ${{ secrets.LAZYLLM_DOUBAO_API_KEY }} | ||
| LAZYLLM_DOUBAO_MODEL_NAME: ${{ secrets.LAZYLLM_DOUBAO_MODEL_NAME }} | ||
| LAZYLLM_SILICONFLOW_API_KEY: ${{ secrets.LAZYLLM_SILICONFLOW_API_KEY }} | ||
| LAZYLLM_SILICONFLOW_MODEL_NAME: ${{ secrets.LAZYLLM_SILICONFLOW_MODEL_NAME }} | ||
| LAZYLLM_MINIMAX_API_KEY: ${{ secrets.LAZYLLM_MINIMAX_API_KEY }} | ||
| LAZYLLM_MINIMAX_MODEL_NAME: ${{ secrets.LAZYLLM_MINIMAX_MODEL_NAME }} | ||
| LAZYLLM_PPOP_API_KEY: ${{ secrets.LAZYLLM_PPOP_API_KEY }} | ||
|
|
||
| steps: | ||
| - name: Checkout | ||
| uses: actions/checkout@v4 | ||
| with: | ||
| submodules: false | ||
|
|
||
| - name: Setup | ||
| uses: ./.github/actions/setup | ||
|
|
||
| - name: Install test requirements | ||
| run: | | ||
| pip install -r tests/requirements.txt | ||
| if [[ "${{ runner.os }}" == "Linux" ]]; then | ||
| pip install -r tests/requirements_linux.txt | ||
| elif [[ "${{ runner.os }}" == "macOS" ]]; then | ||
| pip install -r tests/requirements_mac.txt | ||
| fi | ||
|
|
||
| - name: Download test dataset | ||
| run: | | ||
| set -euo pipefail | ||
| DATA_DIR="$GITHUB_WORKSPACE/.ci_data" | ||
| rm -rf "$DATA_DIR" | ||
| export GIT_TERMINAL_PROMPT=0 | ||
| git clone --depth 1 https://[email protected]/LazyAGI/LazyLLM-Data.git "$DATA_DIR" | ||
| echo "LAZYLLM_DATA_PATH=$DATA_DIR" >> "$GITHUB_ENV" | ||
| env: | ||
| GITHUB_TOKEN: ${{ secrets.PERSONAL_GITHUB_TOKEN || github.token }} | ||
|
|
||
| - name: Build and run C++ tests | ||
| run: | | ||
| pip install pybind11 | ||
| bash csrc/scripts/build_test.sh | ||
|
|
||
| - name: Install C++ extension artifacts into workspace | ||
| run: | | ||
| cmake --install build --prefix . --component lazyllm_cpp | ||
| ls -al lazyllm | rg "lazyllm_cpp|cpp_lib" || true | ||
|
|
||
| - name: Run basic tests | ||
|
Contributor
There was a problem hiding this comment. This is an AI-generated suggestion; please verify before applying. [medium] [maintainability] basic_tests 和 advanced_tests 的 marker 计算逻辑完全重复，违反 DRY 原则，后续修改容易遗漏。 Suggestion: 将 marker 计算提取到一个独立的前置步骤中，通过环境变量传递给后续步骤： auto reviewed by BOT (claude-opus-4-6) |
||
| run: | | ||
| if [[ "${{ runner.os }}" == "Linux" ]]; then | ||
| MARKER="not skip_on_linux" | ||
| elif [[ "${{ runner.os }}" == "macOS" ]]; then | ||
| MARKER="not skip_on_mac" | ||
| else | ||
| MARKER="not skip_on_win" | ||
| fi | ||
| pytest -v --order-scope=class -m "$MARKER" tests/basic_tests | ||
|
|
||
| - name: Run advanced tests | ||
| run: | | ||
| if [[ "${{ runner.os }}" == "Linux" ]]; then | ||
| MARKER="not skip_on_linux" | ||
| elif [[ "${{ runner.os }}" == "macOS" ]]; then | ||
| MARKER="not skip_on_mac" | ||
| else | ||
| MARKER="not skip_on_win" | ||
| fi | ||
| pytest -v --order-scope=class -m "$MARKER" tests/advanced_tests | ||
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
|
|
@@ -191,12 +191,6 @@ jobs: | |
| name: repo-with-docs | ||
| path: ./repo_artifact | ||
|
|
||
| - name: Extract repo-with-docs | ||
| run: | | ||
| set -ex | ||
|
|
||
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
File renamed without changes.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,131 @@ | ||
| #include "binding_utils.hpp" | ||
|
|
||
| #include <type_traits> | ||
|
|
||
| namespace lazyllm::pybind_utils { | ||
|
|
||
| std::string DumpJson(const py::object& obj) { | ||
| py::object json = py::module_::import("json"); | ||
| py::object dumps = json.attr("dumps"); | ||
| py::object dumped = dumps(obj, py::arg("ensure_ascii") = false); | ||
| return dumped.cast<std::string>(); | ||
| } | ||
|
|
||
| py::object LoadJson(const std::string& text) { | ||
| py::object json = py::module_::import("json"); | ||
| py::object loads = json.attr("loads"); | ||
| return loads(py::str(text)); | ||
| } | ||
|
|
||
| bool ExtractStringSequence(const py::object& obj, std::vector<std::string>* out) { | ||
| if (!py::isinstance<py::sequence>(obj) || py::isinstance<py::str>(obj)) return false; | ||
| py::sequence seq = obj.cast<py::sequence>(); | ||
| out->clear(); | ||
| out->reserve(seq.size()); | ||
| for (py::handle item : seq) { | ||
| if (!py::isinstance<py::str>(item)) { | ||
| out->clear(); | ||
| return false; | ||
| } | ||
| out->push_back(py::cast<std::string>(item)); | ||
| } | ||
| return true; | ||
| } | ||
|
|
||
| lazyllm::MetadataMode ParseMetadataMode(const py::object& mode) { | ||
| if (mode.is_none()) return lazyllm::MetadataMode::NONE; | ||
| try { | ||
| if (py::hasattr(mode, "name")) { | ||
| const auto name = py::cast<std::string>(mode.attr("name")); | ||
| if (name == "ALL") return lazyllm::MetadataMode::ALL; | ||
| if (name == "EMBED") return lazyllm::MetadataMode::EMBED; | ||
| if (name == "LLM") return lazyllm::MetadataMode::LLM; | ||
| if (name == "NONE") return lazyllm::MetadataMode::NONE; | ||
| } | ||
| } catch (const py::error_already_set&) { | ||
| } | ||
|
Contributor
There was a problem hiding this comment. This is an AI-generated suggestion; please verify before applying. [medium] [exception] Silently swallowing Suggestion: At minimum, log or clear the Python error state explicitly. Consider narrowing the catch scope or adding a comment explaining why the exception is intentionally ignored. auto reviewed by BOT (claude-opus-4-6) |
||
| if (py::isinstance<py::str>(mode)) { | ||
| const auto name = mode.cast<std::string>(); | ||
| if (name == "ALL") return lazyllm::MetadataMode::ALL; | ||
| if (name == "EMBED") return lazyllm::MetadataMode::EMBED; | ||
| if (name == "LLM") return lazyllm::MetadataMode::LLM; | ||
| if (name == "NONE") return lazyllm::MetadataMode::NONE; | ||
| } | ||
| if (py::isinstance<py::int_>(mode)) { | ||
| const auto value = mode.cast<int>(); | ||
| switch (value) { | ||
| case 0: return lazyllm::MetadataMode::ALL; | ||
| case 1: return lazyllm::MetadataMode::EMBED; | ||
| case 2: return lazyllm::MetadataMode::LLM; | ||
| case 3: return lazyllm::MetadataMode::NONE; | ||
| default: break; | ||
| } | ||
| } | ||
| return lazyllm::MetadataMode::NONE; | ||
| } | ||
|
|
||
| lazyllm::MetadataVType PyToMetadataValue(const py::handle& value) { | ||
| if (value.is_none()) return std::string("None"); | ||
| if (py::isinstance<py::bool_>(value)) return static_cast<int>(value.cast<bool>()); | ||
| if (py::isinstance<py::int_>(value)) return value.cast<int>(); | ||
| if (py::isinstance<py::float_>(value)) return value.cast<double>(); | ||
| if (py::isinstance<py::str>(value)) return value.cast<std::string>(); | ||
|
|
||
| if (py::isinstance<py::sequence>(value) && !py::isinstance<py::str>(value)) { | ||
| py::sequence seq = value.cast<py::sequence>(); | ||
| if (seq.empty()) return std::vector<std::string>{}; | ||
|
|
||
| bool all_str = true; | ||
| bool all_int = true; | ||
| bool all_numeric = true; | ||
|
|
||
| for (py::handle item : seq) { | ||
| const bool is_str = py::isinstance<py::str>(item); | ||
| const bool is_int = py::isinstance<py::int_>(item) && !py::isinstance<py::bool_>(item); | ||
| const bool is_numeric = is_int || py::isinstance<py::float_>(item) || py::isinstance<py::bool_>(item); | ||
| all_str = all_str && is_str; | ||
| all_int = all_int && is_int; | ||
| all_numeric = all_numeric && is_numeric; | ||
| } | ||
|
|
||
| if (all_str) { | ||
| std::vector<std::string> out; | ||
| out.reserve(seq.size()); | ||
| for (py::handle item : seq) out.push_back(py::cast<std::string>(item)); | ||
| return out; | ||
| } | ||
| if (all_int) { | ||
| std::vector<int> out; | ||
| out.reserve(seq.size()); | ||
| for (py::handle item : seq) out.push_back(py::cast<int>(item)); | ||
| return out; | ||
| } | ||
| if (all_numeric) { | ||
| std::vector<double> out; | ||
| out.reserve(seq.size()); | ||
| for (py::handle item : seq) out.push_back(py::cast<double>(item)); | ||
| return out; | ||
| } | ||
|
|
||
| std::vector<std::string> out; | ||
| out.reserve(seq.size()); | ||
| for (py::handle item : seq) out.push_back(py::str(item).cast<std::string>()); | ||
| return out; | ||
| } | ||
| return py::str(value).cast<std::string>(); | ||
| } | ||
|
|
||
| py::object MetadataValueToPy(const lazyllm::MetadataVType& value) { | ||
| return std::visit([](const auto& v) -> py::object { | ||
| using T = std::decay_t<decltype(v)>; | ||
| if constexpr (std::is_same_v<T, std::string>) return py::str(v); | ||
| if constexpr (std::is_same_v<T, int>) return py::int_(v); | ||
| if constexpr (std::is_same_v<T, double>) return py::float_(v); | ||
| if constexpr (std::is_same_v<T, std::vector<std::string>>) return py::cast(v); | ||
| if constexpr (std::is_same_v<T, std::vector<int>>) return py::cast(v); | ||
| if constexpr (std::is_same_v<T, std::vector<double>>) return py::cast(v); | ||
| return py::none(); | ||
| }, value); | ||
| } | ||
|
|
||
| } // namespace lazyllm::pybind_utils | ||
Oops, something went wrong.
Oops, something went wrong.
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
This is an AI-generated suggestion; please verify before applying.
[critical] [logic]
if: always()使得cpp_ext_test作业在needs: [clone]失败时仍会运行，可能导致在没有正确 clone 的情况下执行后续步骤并产生不可预测的失败。Suggestion: 如果意图是即使其他作业失败也运行，但前提是 clone 成功，应改为：
auto reviewed by BOT (claude-opus-4-6)