Extract Python naming helper functions into a separate names library.

PiperOrigin-RevId: 961053716

cmake/installed_include_golden.txt

@@ -35,6 +35,7 @@ google/protobuf/compiler/php/names.h
 google/protobuf/compiler/plugin.h
 google/protobuf/compiler/plugin.pb.h
 google/protobuf/compiler/plugin.proto
+google/protobuf/compiler/python/names.h
 google/protobuf/compiler/retention.h
 google/protobuf/compiler/scc.h
 google/protobuf/compiler/subprocess.h

pkg/BUILD.bazel

@@ -222,6 +222,7 @@ cc_dist_library(
         "//src/google/protobuf/compiler/java:names",
         "//src/google/protobuf/compiler/objectivec:names",
         "//src/google/protobuf/compiler/php:names",
+        "//src/google/protobuf/compiler/python:names",
     ],
 )
 

src/google/protobuf/compiler/python/BUILD.bazel

@@ -27,6 +27,22 @@ cc_binary(
     ],
 )
 
+cc_library(
+    name = "names",
+    srcs = ["names.cc"],
+    hdrs = ["names.h"],
+    copts = COPTS,
+    strip_include_prefix = "/src",
+    visibility = ["//visibility:public"],
+    deps = [
+        "//src/google/protobuf",
+        "//src/google/protobuf:port",
+        "//src/google/protobuf/compiler:code_generator",
+        "@abseil-cpp//absl/log:absl_check",
+        "@abseil-cpp//absl/strings",
+    ],
+)
+
 cc_library(
     name = "python",
     srcs = [
@@ -47,6 +63,7 @@ cc_library(
         "@com_github_grpc_grpc//tools/distrib/python/grpcio_tools:__subpackages__",
     ],
     deps = [
+        ":names",
         "//src/google/protobuf",
         "//src/google/protobuf:port",
         "//src/google/protobuf/compiler:code_generator",

src/google/protobuf/compiler/python/generator.cc

@@ -44,6 +44,7 @@
 #include "absl/strings/substitute.h"
 #include "google/protobuf/compiler/code_generator.h"
 #include "google/protobuf/compiler/python/helpers.h"
+#include "google/protobuf/compiler/python/names.h"
 #include "google/protobuf/compiler/python/pyi_generator.h"
 #include "google/protobuf/compiler/retention.h"
 #include "google/protobuf/compiler/versions.h"
@@ -62,19 +63,6 @@ namespace compiler {
 namespace python {
 
 namespace {
-// Returns the alias we assign to the module of the given .proto filename
-// when importing. See testPackageInitializationImport in
-// third_party/py/google/protobuf/internal/reflection_test.py
-// to see why we need the alias.
-std::string ModuleAlias(absl::string_view filename) {
-  std::string module_name = ModuleName(filename);
-  // We can't have dots in the module name, so we replace each with _dot_.
-  // But that could lead to a collision between a.b and a_dot_b, so we also
-  // duplicate each underscore.
-  absl::StrReplaceAll({{"_", "__"}}, &module_name);
-  absl::StrReplaceAll({{".", "_dot_"}}, &module_name);
-  return module_name;
-}
 
 // Name of the class attribute where we store the Python
 // descriptor.Descriptor instance for the generated class.
@@ -311,7 +299,7 @@ bool Generator::Generate(const FileDescriptor* file,
   if (GeneratingDescriptorProto()) {
     printer_->Outdent();
   }
-  std::string module_name = ModuleName(file->name());
+  std::string module_name = ModuleName(file);
   if (!opensource_runtime_) {
     module_name =
         std::string(absl::StripPrefix(module_name, kThirdPartyPrefix));
@@ -396,8 +384,8 @@ void Generator::PrintTopBoilerplate() const {
   printer_->Print("\n\n");
 }
 
-std::string Generator::ImportModuleName(absl::string_view filename) const {
-  std::string module_name = ModuleName(filename);
+std::string Generator::ImportModuleName(const FileDescriptor* file) const {
+  std::string module_name = ModuleName(file);
   if (!opensource_runtime_) {
     module_name =
         std::string(absl::StripPrefix(module_name, kThirdPartyPrefix));
@@ -409,10 +397,8 @@ std::string Generator::ImportModuleName(absl::string_view filename) const {
 void Generator::PrintImports() const {
   bool has_importlib = false;
   for (int i = 0; i < file_->dependency_count(); ++i) {
-    absl::string_view filename = file_->dependency(i)->name();
-
-    std::string module_name = ImportModuleName(filename);
-    std::string module_alias = ModuleAlias(filename);
+    std::string module_name = ImportModuleName(file_->dependency(i));
+    std::string module_alias = ModuleAlias(file_->dependency(i));
     if (ContainsPythonKeyword(module_name)) {
       // If the module path contains a Python keyword, we have to quote the
       // module name and import it using importlib. Otherwise the usual kind of
@@ -446,8 +432,7 @@ void Generator::PrintImports() const {
 
   // Print public imports.
   for (int i = 0; i < file_->public_dependency_count(); ++i) {
-    std::string module_name =
-        ImportModuleName(file_->public_dependency(i)->name());
+    std::string module_name = ImportModuleName(file_->public_dependency(i));
     printer_->Print("from $module$ import *\n", "module", module_name);
   }
   printer_->Print("\n");
@@ -601,7 +586,7 @@ void Generator::PrintFileDescriptor() const {
     if (file_->dependency_count() != 0) {
       printer_->Print(",\ndependencies=[");
       for (int i = 0; i < file_->dependency_count(); ++i) {
-        std::string module_alias = ModuleAlias(file_->dependency(i)->name());
+        std::string module_alias = ModuleAlias(file_->dependency(i));
         printer_->Print("$module_alias$.DESCRIPTOR,", "module_alias",
                         module_alias);
       }
@@ -610,8 +595,7 @@ void Generator::PrintFileDescriptor() const {
     if (file_->public_dependency_count() > 0) {
       printer_->Print(",\npublic_dependencies=[");
       for (int i = 0; i < file_->public_dependency_count(); ++i) {
-        std::string module_alias =
-            ModuleAlias(file_->public_dependency(i)->name());
+        std::string module_alias = ModuleAlias(file_->public_dependency(i));
         printer_->Print("$module_alias$.DESCRIPTOR,", "module_alias",
                         module_alias);
       }
@@ -741,7 +725,7 @@ void Generator::PrintDescriptorKeyAndModuleName(
   std::string name = ModuleLevelServiceDescriptorName(descriptor);
   printer_->Print("$descriptor_key$ = $descriptor_name$,\n", "descriptor_key",
                   kDescriptorKey, "descriptor_name", name);
-  std::string module_name = ModuleName(file_->name());
+  std::string module_name = ModuleName(file_);
   if (!opensource_runtime_) {
     module_name =
         std::string(absl::StripPrefix(module_name, kThirdPartyPrefix));
@@ -937,7 +921,7 @@ void Generator::PrintMessage(const Descriptor& message_descriptor,
   m["descriptor_key"] = kDescriptorKey;
   m["descriptor_name"] = ModuleLevelDescriptorName(message_descriptor);
   printer_->Print(m, "'$descriptor_key$' : $descriptor_name$,\n");
-  std::string module_name = ModuleName(file_->name());
+  std::string module_name = ModuleName(file_);
   if (!opensource_runtime_) {
     module_name =
         std::string(absl::StripPrefix(module_name, kThirdPartyPrefix));
@@ -1275,7 +1259,7 @@ std::string Generator::ModuleLevelDescriptorName(
   // We now have the name relative to its own module.  Also qualify with
   // the module name iff this descriptor is from a different .proto file.
   if (descriptor.file() != file_) {
-    name = absl::StrCat(ModuleAlias(descriptor.file()->name()), ".", name);
+    name = absl::StrCat(ModuleAlias(descriptor.file()), ".", name);
   }
   return name;
 }
@@ -1288,7 +1272,7 @@ std::string Generator::ModuleLevelMessageName(
     const Descriptor& descriptor) const {
   std::string name = NamePrefixedWithNestedTypes(descriptor, ".");
   if (descriptor.file() != file_) {
-    name = absl::StrCat(ModuleAlias(descriptor.file()->name()), ".", name);
+    name = absl::StrCat(ModuleAlias(descriptor.file()), ".", name);
   }
   return name;
 }
@@ -1300,7 +1284,7 @@ std::string Generator::ModuleLevelServiceDescriptorName(
   std::string name = absl::StrCat("_", descriptor.name());
   absl::AsciiStrToUpper(&name);
   if (descriptor.file() != file_) {
-    name = absl::StrCat(ModuleAlias(descriptor.file()->name()), ".", name);
+    name = absl::StrCat(ModuleAlias(descriptor.file()), ".", name);
   }
   return name;
 }
@@ -1521,8 +1505,8 @@ void Generator::FixOptionsForMessage(const Descriptor& descriptor,
 void Generator::CopyPublicDependenciesAliases(
     absl::string_view copy_from, const FileDescriptor* file) const {
   for (int i = 0; i < file->public_dependency_count(); ++i) {
-    std::string module_name = ModuleName(file->public_dependency(i)->name());
-    std::string module_alias = ModuleAlias(file->public_dependency(i)->name());
+    std::string module_name = ModuleName(file->public_dependency(i));
+    std::string module_alias = ModuleAlias(file->public_dependency(i));
     // There's no module alias in the dependent file if it was generated by
     // an old protoc (less than 3.0.0-alpha-1). Use module name in this
     // situation.

src/google/protobuf/compiler/python/generator.h

@@ -82,7 +82,7 @@ class PROTOC_EXPORT Generator : public CodeGenerator {
  private:
   GeneratorOptions ParseParameter(absl::string_view parameter,
                                   std::string* error) const;
-  std::string ImportModuleName(absl::string_view filename) const;
+  std::string ImportModuleName(const FileDescriptor* file) const;
   void PrintImports() const;
   template <typename DescriptorT>
   std::string GetResolvedFeatures(const DescriptorT& descriptor) const;

src/google/protobuf/compiler/python/helpers.cc

@@ -14,11 +14,12 @@
 #include "absl/log/absl_check.h"
 #include "absl/strings/escaping.h"
 #include "absl/strings/match.h"
+#include "absl/strings/str_cat.h"
 #include "absl/strings/str_replace.h"
 #include "absl/strings/str_split.h"
 #include "absl/strings/string_view.h"
 #include "absl/strings/strip.h"
-#include "google/protobuf/compiler/code_generator.h"
+#include "google/protobuf/compiler/python/names.h"
 #include "google/protobuf/descriptor.h"
 #include "google/protobuf/descriptor.pb.h"
 
@@ -27,18 +28,6 @@ namespace protobuf {
 namespace compiler {
 namespace python {
 
-// Returns the Python module name expected for a given .proto filename.
-std::string ModuleName(absl::string_view filename) {
-  std::string basename = StripProto(filename);
-  absl::StrReplaceAll({{"-", "_"}, {"/", "."}}, &basename);
-  return absl::StrCat(basename, "_pb2");
-}
-
-std::string StrippedModuleName(absl::string_view filename) {
-  std::string module_name = ModuleName(filename);
-  return module_name;
-}
-
 // Keywords reserved by the Python language.
 const char* const kKeywords[] = {
     "False",  "None",     "True",  "and",    "as",       "assert",
@@ -74,7 +63,7 @@ std::string ResolveKeyword(absl::string_view name) {
 
 std::string GetFileName(const FileDescriptor* file_des,
                         absl::string_view suffix) {
-  std::string module_name = ModuleName(file_des->name());
+  std::string module_name = ModuleName(file_des);
   std::string filename = module_name;
   absl::StrReplaceAll({{".", "/"}}, &filename);
   absl::StrAppend(&filename, suffix);

src/google/protobuf/compiler/python/helpers.h

@@ -10,18 +10,18 @@
 
 #include <string>
 
-#include "absl/strings/str_replace.h"
 #include "absl/strings/string_view.h"
 #include "google/protobuf/descriptor.h"
 #include "google/protobuf/descriptor.pb.h"
 
+// Must be included last.
+#include "google/protobuf/port_def.inc"
+
 namespace google {
 namespace protobuf {
 namespace compiler {
 namespace python {
 
-std::string ModuleName(absl::string_view filename);
-std::string StrippedModuleName(absl::string_view filename);
 bool ContainsPythonKeyword(absl::string_view module_name);
 bool IsPythonKeyword(absl::string_view name);
 std::string ResolveKeyword(absl::string_view name);
@@ -40,4 +40,6 @@ std::string NamePrefixedWithNestedTypes(const DescriptorT& descriptor,
 }  // namespace protobuf
 }  // namespace google
 
+#include "google/protobuf/port_undef.inc"
+
 #endif  // GOOGLE_PROTOBUF_COMPILER_PYTHON_HELPERS_H__

src/google/protobuf/compiler/python/names.cc

@@ -0,0 +1,45 @@
+#include "google/protobuf/compiler/python/names.h"
+
+#include <string>
+
+#include "absl/strings/str_cat.h"
+#include "absl/strings/str_replace.h"
+#include "absl/strings/string_view.h"
+#include "absl/strings/strip.h"
+#include "google/protobuf/compiler/code_generator.h"
+
+namespace google {
+namespace protobuf {
+namespace compiler {
+namespace python {
+
+// Returns the Python module name expected for a given .proto filename.
+std::string ModuleName(const FileDescriptor* file) {
+  std::string basename = StripProto(file->name());
+  absl::StrReplaceAll({{"-", "_"}, {"/", "."}}, &basename);
+  return absl::StrCat(basename, "_pb2");
+}
+
+std::string StrippedModuleName(const FileDescriptor* file) {
+  std::string module_name = ModuleName(file);
+  return module_name;
+}
+
+// Returns the alias we assign to the module of the given .proto filename
+// when importing. See testPackageInitializationImport in
+// third_party/py/google/protobuf/internal/reflection_test.py
+// to see why we need the alias.
+std::string ModuleAlias(const FileDescriptor* file) {
+  std::string module_name = ModuleName(file);
+  // We can't have dots in the module name, so we replace each with _dot_.
+  // But that could lead to a collision between a.b and a_dot_b, so we also
+  // duplicate each underscore.
+  absl::StrReplaceAll({{"_", "__"}}, &module_name);
+  absl::StrReplaceAll({{".", "_dot_"}}, &module_name);
+  return module_name;
+}
+
+}  // namespace python
+}  // namespace compiler
+}  // namespace protobuf
+}  // namespace google

src/google/protobuf/compiler/python/names.h

@@ -0,0 +1,36 @@
+#ifndef GOOGLE_PROTOBUF_COMPILER_PYTHON_NAMES_H__
+#define GOOGLE_PROTOBUF_COMPILER_PYTHON_NAMES_H__
+
+#include <string>
+
+// Must be included last.
+#include "google/protobuf/port_def.inc"
+
+namespace google {
+namespace protobuf {
+
+class FileDescriptor;
+
+namespace compiler {
+namespace python {
+
+// Returns the Python module name expected for a given .proto filename.
+PROTOC_EXPORT std::string ModuleName(const FileDescriptor* file);
+
+// Returns the stripped Python module name expected for a given .proto filename.
+PROTOC_EXPORT std::string StrippedModuleName(const FileDescriptor* file);
+
+// Returns the alias we assign to the module of the given .proto filename
+// when importing. See testPackageInitializationImport in
+// third_party/py/google/protobuf/internal/reflection_test.py
+// to see why we need the alias.
+PROTOC_EXPORT std::string ModuleAlias(const FileDescriptor* file);
+
+}  // namespace python
+}  // namespace compiler
+}  // namespace protobuf
+}  // namespace google
+
+#include "google/protobuf/port_undef.inc"
+
+#endif  // GOOGLE_PROTOBUF_COMPILER_PYTHON_NAMES_H__

src/google/protobuf/compiler/python/pyi_generator.cc

@@ -24,6 +24,7 @@
 #include "absl/synchronization/mutex.h"
 #include "google/protobuf/compiler/code_generator.h"
 #include "google/protobuf/compiler/python/helpers.h"
+#include "google/protobuf/compiler/python/names.h"
 #include "google/protobuf/descriptor.h"
 #include "google/protobuf/descriptor.pb.h"
 #include "google/protobuf/io/printer.h"
@@ -45,7 +46,7 @@ std::string PyiGenerator::ModuleLevelName(const DescriptorT& descriptor) const {
     std::string module_alias;
     const absl::string_view filename = descriptor.file()->name();
     if (import_map_.find(filename) == import_map_.end()) {
-      std::string module_name = ModuleName(descriptor.file()->name());
+      std::string module_name = ModuleName(descriptor.file());
       std::vector<absl::string_view> tokens = absl::StrSplit(module_name, '.');
       module_alias = absl::StrCat("_", tokens.back());
     } else {
@@ -151,7 +152,7 @@ void PyiGenerator::PrintImportForDescriptor(
     const FileDescriptor& desc, absl::flat_hash_set<std::string>* seen_aliases,
     bool* has_importlib) const {
   const absl::string_view filename = desc.name();
-  std::string module_name_owned = StrippedModuleName(filename);
+  std::string module_name_owned = StrippedModuleName(&desc);
   absl::string_view module_name(module_name_owned);
   size_t last_dot_pos = module_name.rfind('.');
   std::string alias = absl::StrCat("_", module_name.substr(last_dot_pos + 1));
@@ -296,7 +297,7 @@ void PyiGenerator::PrintImports() const {
   // Public imports
   for (int i = 0; i < file_->public_dependency_count(); ++i) {
     const FileDescriptor* public_dep = file_->public_dependency(i);
-    std::string module_name = StrippedModuleName(public_dep->name());
+    std::string module_name = StrippedModuleName(public_dep);
     // Top level messages in public imports
     for (int i = 0; i < public_dep->message_type_count(); ++i) {
       printer_->Print(
@@ -404,7 +405,7 @@ std::string PyiGenerator::GetFieldType(
       std::string name = ModuleLevelName(*field_des.message_type());
       if ((containing_des.containing_type() != nullptr &&
            name == containing_des.name())) {
-        std::string module = ModuleName(field_des.file()->name());
+        std::string module = ModuleName(field_des.file());
         name = absl::StrCat(module, ".", name);
       }
       return name;