Add a C++ program that prints operator document in JSON format (#4981)

* Add print_operators_doc.cc

* Update Escape

* Correct a bug

* Remove OpInfoMap::Iterate

* Update the print_operators_doc.cc

* Escape tab

* Use auto&

* Use auto&

* Remove trailing ,

* clang-format C++

paddle/framework/op_info.h

@@ -87,11 +87,8 @@ class OpInfoMap {
     }
   }
 
-  template <typename Callback>
-  void IterAllInfo(Callback callback) {
-    for (auto& it : map_) {
-      callback(it.first, it.second);
-    }
+  const std::unordered_map<std::string, const OpInfo>& map() const {
+    return map_;
   }
 
  private:

paddle/pybind/CMakeLists.txt

@@ -4,3 +4,5 @@ if(WITH_PYTHON)
     DEPS pybind python backward proto_desc tensor_array paddle_memory executor
     ${GLOB_OP_LIB})
 endif(WITH_PYTHON)
+
+cc_binary(print_operators_doc SRCS print_operators_doc.cc DEPS ${GLOB_OP_LIB} tensor_array)

paddle/pybind/print_operators_doc.cc

+#include <iostream>
+#include <sstream>  // std::stringstream
+#include <string>
+
+#include "paddle/framework/op_info.h"
+#include "paddle/framework/op_registry.h"
+#include "paddle/pybind/pybind.h"
+
+std::string Escape(const std::string& s) {
+  std::string r;
+  for (size_t i = 0; i < s.size(); i++) {
+    switch (s[i]) {
+      case '\"':
+        r += "\\\"";
+        break;
+      case '\\':
+        r += "\\\\";
+        break;
+      case '\n':
+        r += "\\n";
+        break;
+      case '\t':
+        r += "\\t";
+      case '\r':
+        break;
+      default:
+        r += s[i];
+        break;
+    }
+  }
+  return r;
+}
+
+std::string AttrType(paddle::framework::AttrType at) {
+  switch (at) {
+    case paddle::framework::INT:
+      return "int";
+    case paddle::framework::FLOAT:
+      return "float";
+    case paddle::framework::STRING:
+      return "string";
+    case paddle::framework::BOOLEAN:
+      return "bool";
+    case paddle::framework::INTS:
+      return "int array";
+    case paddle::framework::FLOATS:
+      return "float array";
+    case paddle::framework::STRINGS:
+      return "string array";
+    case paddle::framework::BOOLEANS:
+      return "bool array";
+    case paddle::framework::BLOCK:
+      return "block id";
+  }
+  return "UNKNOWN";  // not possible
+}
+
+void PrintVar(const paddle::framework::OpProto::Var& v, std::stringstream& ss) {
+  ss << " { "
+     << "\n"
+     << "   \"name\" : \"" << Escape(v.name()) << "\",\n"
+     << "   \"comment\" : \"" << Escape(v.comment()) << "\",\n"
+     << "   \"duplicable\" : " << v.duplicable() << ",\n"
+     << "   \"intermediate\" : " << v.intermediate() << "\n"
+     << " },";
+}
+
+void PrintAttr(const paddle::framework::OpProto::Attr& a,
+               std::stringstream& ss) {
+  ss << " { "
+     << "\n"
+     << "   \"name\" : \"" << Escape(a.name()) << "\",\n"
+     << "   \"type\" : \"" << AttrType(a.type()) << "\",\n"
+     << "   \"comment\" : \"" << Escape(a.comment()) << "\",\n"
+     << "   \"generated\" : " << a.generated() << "\n"
+     << " },";
+}
+
+void PrintOpProto(const std::string& type,
+                  const paddle::framework::OpInfo& opinfo,
+                  std::stringstream& ss) {
+  std::cerr << "Processing " << type << "\n";
+
+  const paddle::framework::OpProto* p = opinfo.proto_;
+  if (p == nullptr) {
+    return;  // It is possible that an operator doesn't have OpProto.
+  }
+
+  ss << "{\n"
+     << " \"type\" : \"" << Escape(p->type()) << "\",\n"
+     << " \"comment\" : \"" << Escape(p->comment()) << "\",\n";
+
+  ss << " \"inputs\" : [ "
+     << "\n";
+  for (int i = 0; i < p->inputs_size(); i++) {
+    PrintVar(p->inputs(i), ss);
+  }
+  ss.seekp(-1, ss.cur);  // remove the trailing comma
+  ss << " ], "
+     << "\n";
+
+  ss << " \"outputs\" : [ "
+     << "\n";
+  for (int i = 0; i < p->outputs_size(); i++) {
+    PrintVar(p->outputs(i), ss);
+  }
+  ss.seekp(-1, ss.cur);  // remove the trailing comma
+  ss << " ], "
+     << "\n";
+
+  ss << " \"attrs\" : [ "
+     << "\n";
+  for (int i = 0; i < p->attrs_size(); i++) {
+    PrintAttr(p->attrs(i), ss);
+  }
+  ss.seekp(-1, ss.cur);  // remove the trailing comma
+  ss << " ] "
+     << "\n";
+
+  ss << "},";
+}
+
+int main() {
+  std::stringstream ss;
+  ss << "[\n";
+  for (auto& iter : paddle::framework::OpInfoMap::Instance().map()) {
+    PrintOpProto(iter.first, iter.second, ss);
+  }
+  ss.seekp(-1, ss.cur);  // remove the trailing comma
+  ss << "]\n";
+  std::cout << ss.str();
+}

paddle/pybind/pybind.cc

@@ -225,15 +225,16 @@ All parameter, weight, gradient are variables in Paddle.
   //! Python str. If you want a str object, you should cast them in Python.
   m.def("get_all_op_protos", []() -> std::vector<py::bytes> {
     std::vector<py::bytes> ret_values;
-
-    OpInfoMap::Instance().IterAllInfo([&ret_values](const std::string &type,
-                                                    const OpInfo &info) {
-      if (!info.HasOpProtoAndChecker()) return;
+    for (auto &iter : OpInfoMap::Instance().map()) {
+      auto &info = iter.second;
+      if (info.HasOpProtoAndChecker()) {
         std::string str;
-      PADDLE_ENFORCE(info.Proto().SerializeToString(&str),
+        PADDLE_ENFORCE(
+            info.Proto().SerializeToString(&str),
             "Serialize OpProto Error. This could be a bug of Paddle.");
         ret_values.emplace_back(str);
-    });
+      }
+    }
     return ret_values;
   });
   m.def_submodule(