Use a JSON library to feed FillableArray.

.ci/azure-buildtest-awkward.yml

@@ -10,6 +10,7 @@ trigger:
       - .ci/azure-deploy-awkward.yml
       - .ci/linux-build.sh
       - docs/*
+      - studies/*
 
 pr:
   branches:

.gitignore

@@ -1,3 +1,5 @@
+studies/**/sample-*
+
 ############################################################# IDEs
 
 # ...

.gitmodules

@@ -1,6 +1,6 @@
 [submodule "pybind11"]
 	path = pybind11
 	url = https://github.com/pybind/pybind11.git
-[submodule "simdjson"]
-	path = simdjson
-	url = https://github.com/lemire/simdjson.git
+[submodule "rapidjson"]
+	path = rapidjson
+	url = https://github.com/Tencent/rapidjson.git

CMakeLists.txt

@@ -19,8 +19,9 @@ add_definitions(-DVERSION_INFO="${VERSION_INFO}")
 set(CMAKE_MACOSX_RPATH 1)
 
 file(GLOB CPU_KERNEL_SOURCES "src/cpu-kernels/*.cpp")
-file(GLOB LIBAWKWARD_SOURCES "src/libawkward/*.cpp" "src/libawkward/array/*.cpp" "src/libawkward/fillable/*.cpp" "src/libawkward/type/*.cpp")
+file(GLOB LIBAWKWARD_SOURCES "src/libawkward/*.cpp" "src/libawkward/array/*.cpp" "src/libawkward/fillable/*.cpp" "src/libawkward/type/*.cpp" "src/libawkward/io/*.cpp")
 include_directories(include)
+include_directories(rapidjson/include)
 
 add_subdirectory(pybind11)
 
@@ -44,6 +45,7 @@ add_library(awkward        SHARED $<TARGET_OBJECTS:awkward-objects>)
 target_link_libraries(awkward-static PRIVATE awkward-cpu-kernels-static)
 target_link_libraries(awkward        PRIVATE awkward-cpu-kernels-static)
 addtest(PR016 tests/test_PR016_finish_getitem_for_rawarray.cpp)
+addtest(PR019 tests/test_PR019_use_json_library.cpp)
 
 pybind11_add_module(layout src/pyawkward.cpp)
 set_target_properties(layout PROPERTIES CXX_VISIBILITY_PRESET default)

README.md

@@ -119,6 +119,7 @@ Completed items are ☑check-marked. See [closed PRs](https://github.com/scikit-
    * [ ] Translation to and from Apache Arrow and Parquet in C++.
    * [ ] Persistence to any medium that stores named binary blobs, as before, but accessible via C++ (especially for writing). The persistence format might differ slightly from the existing one (break backward compatibility, if needed).
    * [ ] Universal `array.get[...]` as a softer form of `array[...]` that skips non-existent indexes, rather than raising errors.
+   * [ ] Explicit interface with [NumExpr](https://numexpr.readthedocs.io/en/latest/index.html).
 
 ### At some point in the future
 

awkward1/__init__.py

@@ -2,6 +2,7 @@
 
 import awkward1.layout
 import awkward1._numba
+
 from awkward1.operations.convert import *
 
 __version__ = awkward1.layout.__version__

awkward1/operations/convert.py

@@ -1,12 +1,23 @@
 # BSD 3-Clause License; see https://github.com/jpivarski/awkward-1.0/blob/master/LICENSE
 
 import numbers
+import json
+try:
+    from collections.abc import Iterable
+except ImportError:
+    from collections import Iterable
 
 import numpy
 
 import awkward1.util
 import awkward1.layout
 
+def fromiter(iterable, initial=1024, resize=2.0):
+    out = awkward1.layout.FillableArray(initial=initial, resize=resize)
+    for x in iterable:
+        out.fill(x)
+    return out.snapshot()
+
 def tolist(array):
     if array is None or isinstance(array, (bool, str, bytes, numbers.Number)):
         return array
@@ -26,4 +37,22 @@ def tolist(array):
     else:
         raise TypeError("unrecognized array type: {0}".format(repr(array)))
 
-__all__ = [x for x in list(globals()) if not x.startswith("_") and x not in ("awkward1", "numpy")]
+fromjson = awkward1.layout.fromjson
+
+def tojson(array, *args, **kwargs):
+    if array is None or isinstance(array, (bool, str, bytes, numbers.Number)):
+        return json.dumps(array)
+
+    elif isinstance(array, numpy.ndarray):
+        return awkward1.layout.NumpyArray(array).tojson(*args, **kwargs)
+
+    elif isinstance(array, awkward1.layout.FillableArray):
+        return array.snapshot().tojson(*args, **kwargs)
+
+    elif isinstance(array, awkward1.layout.Content):
+        return array.tojson(*args, **kwargs)
+
+    else:
+        raise TypeError("unrecognized array type: {0}".format(repr(array)))
+
+__all__ = [x for x in list(globals()) if not x.startswith("_") and x not in ("numbers", "json", "Iterable", "numpy", "awkward1")]

include/awkward/Content.h

@@ -3,9 +3,12 @@
 #ifndef AWKWARD_CONTENT_H_
 #define AWKWARD_CONTENT_H_
 
+#include <cstdio>
+
 #include "awkward/cpu-kernels/util.h"
 #include "awkward/Identity.h"
 #include "awkward/Slice.h"
+#include "awkward/io/json.h"
 
 namespace awkward {
   class Content {
@@ -17,6 +20,7 @@ namespace awkward {
     virtual void setid() = 0;
     virtual void setid(const std::shared_ptr<Identity> id) = 0;
     virtual const std::string tostring_part(const std::string indent, const std::string pre, const std::string post) const = 0;
+    virtual void tojson_part(ToJson& builder) const = 0;
     virtual int64_t length() const = 0;
     virtual const std::shared_ptr<Content> shallow_copy() const = 0;
     virtual void checksafe() const = 0;
@@ -30,6 +34,8 @@ namespace awkward {
     virtual const std::pair<int64_t, int64_t> minmax_depth() const = 0;
 
     const std::string tostring() const;
+    const std::string tojson(bool pretty, int64_t maxdecimals) const;
+    void tojson(FILE* destination, bool pretty, int64_t maxdecimals, int64_t buffersize) const;
     const std::shared_ptr<Content> getitem_ellipsis(const Slice& tail, const Index64& advanced) const;
     const std::shared_ptr<Content> getitem_newaxis(const Slice& tail, const Index64& advanced) const;
   };

include/awkward/Index.h

@@ -35,6 +35,7 @@ namespace awkward {
     const std::string tostring_part(const std::string indent, const std::string pre, const std::string post) const;
     T getitem_at(int64_t at) const;
     T getitem_at_unsafe(int64_t at) const;
+    void setitem_at_unsafe(int64_t at, T value) const;
     IndexOf<T> getitem_range(int64_t start, int64_t stop) const;
     IndexOf<T> getitem_range_unsafe(int64_t start, int64_t stop) const;
     virtual const std::shared_ptr<Index> shallow_copy() const;

include/awkward/Slice.h

@@ -36,7 +36,7 @@ namespace awkward {
 
   class SliceRange: public SliceItem {
   public:
-    SliceRange(int64_t start, int64_t stop, int64_t step): start_(start), stop_(stop), step_(step) {
+    SliceRange(int64_t start, int64_t stop, int64_t step): start_(start), stop_(stop), step_(step == none() ? 1 : step) {
       assert(step_ != 0);
     }
     int64_t start() const { return start_; }
@@ -114,6 +114,12 @@ namespace awkward {
     const Slice tail() const;
     const std::string tostring() const;
     void append(const std::shared_ptr<SliceItem>& item);
+    void append(const SliceAt& item);
+    void append(const SliceRange& item);
+    void append(const SliceEllipsis& item);
+    void append(const SliceNewAxis& item);
+    template <typename T>
+    void append(const SliceArrayOf<T>& item);
     void become_sealed();
     bool isadvanced() const;
 

include/awkward/array/ListArray.h

@@ -29,6 +29,7 @@ namespace awkward {
     virtual void setid();
     virtual void setid(const std::shared_ptr<Identity> id);
     virtual const std::string tostring_part(const std::string indent, const std::string pre, const std::string post) const;
+    virtual void tojson_part(ToJson& builder) const;
     virtual int64_t length() const;
     virtual const std::shared_ptr<Content> shallow_copy() const;
     virtual void checksafe() const;

include/awkward/array/ListOffsetArray.h

@@ -27,6 +27,7 @@ namespace awkward {
     virtual void setid();
     virtual void setid(const std::shared_ptr<Identity> id);
     virtual const std::string tostring_part(const std::string indent, const std::string pre, const std::string post) const;
+    virtual void tojson_part(ToJson& builder) const;
     virtual int64_t length() const;
     virtual const std::shared_ptr<Content> shallow_copy() const;
     virtual void checksafe() const;

include/awkward/array/NumpyArray.h

@@ -37,6 +37,7 @@ namespace awkward {
     bool isscalar() const;
     bool isempty() const;
     void* byteptr() const;
+    void* byteptr(ssize_t at) const;
     ssize_t bytelength() const;
     uint8_t getbyte(ssize_t at) const;
 
@@ -45,6 +46,7 @@ namespace awkward {
     virtual void setid();
     virtual void setid(const std::shared_ptr<Identity> id);
     virtual const std::string tostring_part(const std::string indent, const std::string pre, const std::string post) const;
+    virtual void tojson_part(ToJson& builder) const;
     virtual int64_t length() const;
     virtual const std::shared_ptr<Content> shallow_copy() const;
     virtual void checksafe() const;

include/awkward/array/RawArray.h

@@ -20,6 +20,26 @@
 #include "awkward/Content.h"
 
 namespace awkward {
+  void tojson_boolean(ToJson& builder, bool* array, int64_t length) {
+    for (int i = 0;  i < length;  i++) {
+      builder.boolean((bool)array[i]);
+    }
+  }
+
+  template <typename T>
+  void tojson_integer(ToJson& builder, T* array, int64_t length) {
+    for (int i = 0;  i < length;  i++) {
+      builder.integer((int64_t)array[i]);
+    }
+  }
+
+  template <typename T>
+  void tojson_real(ToJson& builder, T* array, int64_t length) {
+    for (int i = 0;  i < length;  i++) {
+      builder.real((double)array[i]);
+    }
+  }
+
   template <typename T>
   class RawArrayOf: public Content {
   public:
@@ -123,6 +143,42 @@ namespace awkward {
       return out.str();
     }
 
+    virtual void tojson_part(ToJson& builder) const {
+      if (std::is_same<T, double>::value) {
+        tojson_real(builder, reinterpret_cast<double*>(byteptr()), length());
+      }
+      else if (std::is_same<T, float>::value) {
+        tojson_real(builder, reinterpret_cast<float*>(byteptr()), length());
+      }
+      else if (std::is_same<T, int64_t>::value) {
+        tojson_real(builder, reinterpret_cast<int64_t*>(byteptr()), length());
+      }
+      else if (std::is_same<T, uint64_t>::value) {
+        tojson_real(builder, reinterpret_cast<uint64_t*>(byteptr()), length());
+      }
+      else if (std::is_same<T, int32_t>::value) {
+        tojson_real(builder, reinterpret_cast<int32_t*>(byteptr()), length());
+      }
+      else if (std::is_same<T, uint32_t>::value) {
+        tojson_real(builder, reinterpret_cast<uint32_t*>(byteptr()), length());
+      }
+      else if (std::is_same<T, int16_t>::value) {
+        tojson_real(builder, reinterpret_cast<int16_t*>(byteptr()), length());
+      }
+      else if (std::is_same<T, uint16_t>::value) {
+        tojson_real(builder, reinterpret_cast<uint16_t*>(byteptr()), length());
+      }
+      else if (std::is_same<T, int8_t>::value) {
+        tojson_real(builder, reinterpret_cast<int8_t*>(byteptr()), length());
+      }
+      else if (std::is_same<T, uint8_t>::value) {
+        tojson_real(builder, reinterpret_cast<uint8_t*>(byteptr()), length());
+      }
+      else {
+        throw std::invalid_argument(std::string("cannot convert RawArrayOf<") + typeid(T).name() + std::string("> into JSON"));
+      }
+    }
+
     virtual int64_t length() const { return length_; }
 
     virtual const std::shared_ptr<Content> shallow_copy() const { return std::shared_ptr<Content>(new RawArrayOf<T>(id_, ptr_, offset_, length_, itemsize_)); }
@@ -232,7 +288,7 @@ namespace awkward {
           throw std::runtime_error("array.ndim != 1");
         }
         Index64 flathead = array->ravel();
-        Error err = awkward_regularize_arrayslice_64(
+        struct Error err = awkward_regularize_arrayslice_64(
           flathead.ptr().get(),
           flathead.length(),
           length_);
@@ -247,7 +303,7 @@ namespace awkward {
 
     virtual const std::shared_ptr<Content> carry(const Index64& carry) const {
       std::shared_ptr<T> ptr(new T[(size_t)carry.length()], awkward::util::array_deleter<T>());
-      Error err = awkward_numpyarray_getitem_next_null_64(
+      struct Error err = awkward_numpyarray_getitem_next_null_64(
         reinterpret_cast<uint8_t*>(ptr.get()),
         reinterpret_cast<uint8_t*>(ptr_.get()),
         carry.length(),