!7549 [MSLITE][Develop] add lite string interface

Merge pull request !7549 from sunsuodong/add_lite_interface

mindspore/lite/include/lite_utils.h

@@ -20,6 +20,7 @@
 #include <string>
 #include <memory>
 #include "schema/model_generated.h"
+#include "include/ms_tensor.h"
 
 namespace mindspore::lite {
 /// \brief Allocator defined a memory pool for malloc memory and free memory dynamically.
@@ -36,5 +37,18 @@ using Uint32Vector = std::vector<uint32_t>;
 using String = std::string;
 using NodeType = schema::NodeType;
 using AllocatorPtr = std::shared_ptr<Allocator>;
+
+/// \brief Set data of MSTensor from string vector.
+///
+/// \param[in] input string vector.
+/// \param[out] MSTensor.
+///
+/// \return STATUS as an error code of this interface, STATUS is defined in errorcode.h.
+int StringsToMSTensor(const std::vector<std::string> &inputs, tensor::MSTensor *tensor);
+
+/// \brief Get string vector from MSTensor.
+/// \param[in] MSTensor.
+/// \return string vector.
+std::vector<std::string> MSTensorToStrings(const tensor::MSTensor *tensor);
 }  // namespace mindspore::lite
 #endif  // MINDSPORE_LITE_INCLUDE_LITE_UTILS_H_

mindspore/lite/src/common/string_util.cc

@@ -15,6 +15,8 @@
  */
 
 #include "src/common/string_util.h"
+#include <algorithm>
+#include "include/ms_tensor.h"
 
 namespace mindspore {
 namespace lite {
@@ -28,9 +30,13 @@ std::vector<StringPack> ParseTensorBuffer(Tensor *tensor) {
 }
 
 std::vector<StringPack> ParseStringBuffer(const void *data) {
+  std::vector<StringPack> buffer;
+  if (data == nullptr) {
+    MS_LOG(ERROR) << "data is nullptr";
+    return buffer;
+  }
   const int32_t *offset = reinterpret_cast<const int32_t *>(data);
   int32_t num = *offset;
-  std::vector<StringPack> buffer;
   for (int i = 0; i < num; i++) {
     offset += 1;
     buffer.push_back(StringPack{(*(offset + 1)) - (*offset), reinterpret_cast<const char *>(data) + (*offset)});
@@ -108,6 +114,26 @@ int GetStringCount(const void *data) { return *(static_cast<const int32_t *>(dat
 
 int GetStringCount(Tensor *tensor) { return GetStringCount(tensor->MutableData()); }
 
+int StringsToMSTensor(const std::vector<std::string> &inputs, tensor::MSTensor *tensor) {
+  std::vector<StringPack> all_pack;
+  for (auto &input : inputs) {
+    StringPack pack = {static_cast<int>(input.length()), input.data()};
+    all_pack.push_back(pack);
+  }
+  return WriteStringsToTensor(static_cast<Tensor *>(tensor), all_pack);
+}
+
+std::vector<std::string> MSTensorToStrings(const tensor::MSTensor *tensor) {
+  const void *ptr = static_cast<const Tensor *>(tensor)->data_c();
+  std::vector<StringPack> all_pack = ParseStringBuffer(ptr);
+  std::vector<std::string> result(all_pack.size());
+  std::transform(all_pack.begin(), all_pack.end(), result.begin(), [](StringPack &pack) {
+    std::string str(pack.data, pack.len);
+    return str;
+  });
+  return result;
+}
+
 // Some primes between 2^63 and 2^64
 static uint64_t k0 = 0xc3a5c85c97cb3127ULL;
 static uint64_t k1 = 0xb492b66fbe98f273ULL;

mindspore/lite/src/common/string_util.h

@@ -33,12 +33,11 @@ typedef struct {
 } StringPack;
 
 // example of string tensor:
-// 3, 0, 0, 0    # int32, num of strings
-// 20, 0, 0, 0   # int32, offset of 0-th string
-// 23, 0, 0, 0   # int32, offset of 0-th string
-// 26, 0, 0, 0   # int32, offset of 0-th string
-// 29, 0, 0, 0   # int32, total length of tensor data
-// 'h', 'o', 'w', 'a', 'r', 'e', 'y', 'o', 'u'  # char, how are you
+// 2, 0, 0, 0    # int32, num of strings
+// 16, 0, 0, 0   # int32, offset of 0-th string
+// 21, 0, 0, 0   # int32, offset of 1-th string
+// 30, 0, 0, 0   # int32, total length of tensor data
+// 'h', 'e', 'l', 'l', 'o', 'h', 'o', 'w', 'a', 'r', 'e', 'y', 'o', 'u'  # char, "hello", "how are you"
 std::vector<StringPack> ParseTensorBuffer(Tensor *tensor);
 std::vector<StringPack> ParseStringBuffer(const void *data);