Fix OSS builds with the following changes:

1. Use CTAD to allow TriStatePtr to be used with non-unique_ptr data.
2. Migrate from old WORKSPACE to MODULE.bazel
3. Use pip python rule to manage pip dependencies required by array_record_datasource.

PiperOrigin-RevId: 709169059

BUILD

@@ -1,6 +1,13 @@
+load("@rules_python//python:pip.bzl", "compile_pip_requirements")
 
 
 py_library(
     name = "setup",
     srcs = ["setup.py"],
 )
+
+compile_pip_requirements(
+    name = "requirements",
+    src = "requirements.in",
+    requirements_txt = "requirements_lock.txt",
+)

MODULE.bazel

@@ -0,0 +1,65 @@
+###############################################################################
+# Bazel now uses Bzlmod by default to manage external dependencies.
+# Please consider migrating your external dependencies from WORKSPACE to MODULE.bazel.
+#
+# For more details, please check https://github.com/bazelbuild/bazel/issues/18958
+###############################################################################
+
+# TODO(fchern): automate version string alignment with setup.py
+VERSION = "0.6.0"
+
+module(
+    name = "array_record",
+    version = VERSION,
+    repo_name = "com_google_array_record",
+)
+
+bazel_dep(name = "abseil-cpp", version = "20240722.0", repo_name = "com_google_absl")
+bazel_dep(name = "abseil-py", version = "2.1.0", repo_name = "com_google_absl_py")
+bazel_dep(name = "pybind11_bazel", version = "2.12.0")
+bazel_dep(name = "rules_proto", version = "7.0.2")
+bazel_dep(name = "googletest", version = "1.15.2", repo_name = "com_google_googletest")
+bazel_dep(name = "riegeli", version = "0.0.0-20240927-cdfb25a", repo_name = "com_google_riegeli")
+bazel_dep(name = "platforms", version = "0.0.10")
+bazel_dep(name = "protobuf", version = "29.0", repo_name = "com_google_protobuf")
+bazel_dep(name = "rules_python", version = "0.40.0")
+
+http_archive = use_repo_rule("@bazel_tools//tools/build_defs/repo:http.bzl", "http_archive")
+
+# V3.4.0, 20210818
+http_archive(
+    name = "eigen3",
+    build_file_content =
+        """
+cc_library(
+    name = 'eigen3',
+    srcs = [],
+    includes = ['.'],
+    hdrs = glob(['Eigen/**', 'unsupported/Eigen/**']),
+    visibility = ['//visibility:public'],
+)
+""",
+    sha256 = "b4c198460eba6f28d34894e3a5710998818515104d6e74e5cc331ce31e46e626",
+    strip_prefix = "eigen-3.4.0",
+    urls = [
+        "https://gitlab.com/libeigen/eigen/-/archive/3.4.0/eigen-3.4.0.tar.bz2",
+    ],
+)
+
+PYTHON_VERSION = "3.10"
+
+python = use_extension("@rules_python//python/extensions:python.bzl", "python")
+python.toolchain(
+    python_version = PYTHON_VERSION,
+)
+
+pip = use_extension("@rules_python//python/extensions:pip.bzl", "pip")
+
+# requirements_lock.txt is generated by
+# bazel run //:requirements.update
+pip.parse(
+    hub_name = "pypi",
+    python_version = PYTHON_VERSION,
+    requirements_lock = "//:requirements_lock.txt",
+)
+use_repo(pip, "pypi")

cpp/BUILD

@@ -77,7 +77,10 @@ cc_library(
     deps = [
         ":common",
         "@com_google_absl//absl/base:core_headers",
+        "@com_google_absl//absl/base:nullability",
         "@com_google_absl//absl/synchronization",
+        "@com_google_riegeli//riegeli/base:dependency",
+        "@com_google_riegeli//riegeli/base:initializer",
     ],
 )
 

cpp/array_record_reader.cc

@@ -342,8 +342,8 @@ absl::Status ArrayRecordReaderBase::ParallelReadRecords(
         uint64_t chunk_idx_start = buf_idx * state_->chunk_group_size;
         // inclusive index, not the conventional exclusive index.
         uint64_t last_chunk_idx =
-            std::min((buf_idx + 1) * state_->chunk_group_size - 1,
-                     state_->chunk_offsets.size() - 1);
+            std::min<uint64_t>((buf_idx + 1) * state_->chunk_group_size - 1,
+                               state_->chunk_offsets.size() - 1);
         uint64_t buf_len = state_->ChunkEndOffset(last_chunk_idx) -
                            state_->chunk_offsets[chunk_idx_start];
         AR_ENDO_JOB(
@@ -708,8 +708,8 @@ bool ArrayRecordReaderBase::ReadAheadFromBuffer(uint64_t buffer_idx) {
     chunk_offsets.reserve(state_->chunk_group_size);
     uint64_t chunk_start = buffer_to_add * state_->chunk_group_size;
     uint64_t chunk_end =
-        std::min(state_->chunk_offsets.size(),
-                 (buffer_to_add + 1) * state_->chunk_group_size);
+        std::min<uint64_t>(state_->chunk_offsets.size(),
+                           (buffer_to_add + 1) * state_->chunk_group_size);
     for (uint64_t chunk_idx = chunk_start; chunk_idx < chunk_end; ++chunk_idx) {
       chunk_offsets.push_back(state_->chunk_offsets[chunk_idx]);
     }

cpp/array_record_reader.h

@@ -32,6 +32,7 @@ limitations under the License.
 #ifndef ARRAY_RECORD_CPP_ARRAY_RECORD_READER_H_
 #define ARRAY_RECORD_CPP_ARRAY_RECORD_READER_H_
 
+#include <cstddef>
 #include <cstdint>
 #include <memory>
 #include <optional>
@@ -46,8 +47,8 @@ limitations under the License.
 #include "absl/strings/string_view.h"
 #include "absl/types/span.h"
 #include "cpp/common.h"
-#include "cpp/tri_state_ptr.h"
 #include "cpp/thread_pool.h"
+#include "cpp/tri_state_ptr.h"
 #include "google/protobuf/message_lite.h"
 #include "riegeli/base/initializer.h"
 #include "riegeli/base/object.h"
@@ -293,7 +294,7 @@ class ArrayRecordReaderBase : public riegeli::Object {
 
   void Initialize();
 
-  virtual TriStatePtr<riegeli::Reader>::SharedRef get_backing_reader()
+  virtual TriStatePtrBase<riegeli::Reader>::SharedRef get_backing_reader()
       const = 0;
 
  private:
@@ -345,24 +346,27 @@ class ArrayRecordReader : public ArrayRecordReaderBase {
                              Options options = Options(),
                              ARThreadPool* pool = nullptr)
       : ArrayRecordReaderBase(std::move(options), pool),
-        main_reader_(std::make_unique<TriStatePtr<riegeli::Reader>>(
+        main_reader_(std::make_unique<TriStatePtr<riegeli::Reader, Src>>(
             std::move(src))) {
     Initialize();
   }
 
  protected:
-  TriStatePtr<riegeli::Reader>::SharedRef get_backing_reader() const override {
+  TriStatePtrBase<riegeli::Reader>::SharedRef get_backing_reader()
+      const override {
     return main_reader_->MakeShared();
   }
 
   void Done() override {
-    if (main_reader_ == nullptr) return;
+    if (main_reader_ == nullptr) {
+      return;
+    }
     auto unique = main_reader_->WaitAndMakeUnique();
     if (!unique->Close()) Fail(unique->status());
   }
 
  private:
-  std::unique_ptr<TriStatePtr<riegeli::Reader>> main_reader_;
+  std::unique_ptr<TriStatePtr<riegeli::Reader, Src>> main_reader_;
 };
 
 template <typename Src>

cpp/array_record_writer.cc

@@ -265,7 +265,7 @@ class ArrayRecordWriterBase::SubmitChunkCallback
 
   // Aggregate the offsets information and write it to the file.
   void WriteFooterAndPostscript(
-      TriStatePtr<SequencedChunkWriterBase>::SharedRef writer);
+      TriStatePtrBase<SequencedChunkWriterBase>::SharedRef writer);
 
  private:
   const Options options_;
@@ -489,7 +489,7 @@ void ArrayRecordWriterBase::SubmitChunkCallback::operator()(
 }
 
 void ArrayRecordWriterBase::SubmitChunkCallback::WriteFooterAndPostscript(
-    TriStatePtr<SequencedChunkWriterBase>::SharedRef writer) {
+    TriStatePtrBase<SequencedChunkWriterBase>::SharedRef writer) {
   // Flushes prior chunks
   writer->SubmitFutureChunks(true);
   // Footer and postscript must pad to block boundary