objToJSON.c

/*
Copyright (c) 2011-2013, ESN Social Software AB and Jonas Tarnstrom
All rights reserved.

Redistribution and use in source and binary forms, with or without
modification, are permitted provided that the following conditions are met:
* Redistributions of source code must retain the above copyright
notice, this list of conditions and the following disclaimer.
* Redistributions in binary form must reproduce the above copyright
notice, this list of conditions and the following disclaimer in the
documentation and/or other materials provided with the distribution.
* Neither the name of the ESN Social Software AB nor the
names of its contributors may be used to endorse or promote products
derived from this software without specific prior written permission.

THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND
ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
DISCLAIMED. IN NO EVENT SHALL ESN SOCIAL SOFTWARE AB OR JONAS TARNSTROM BE
LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE
GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF
THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.


Portions of code from MODP_ASCII - Ascii transformations (upper/lower, etc)
https://github.com/client9/stringencoders
Copyright (c) 2007  Nick Galbreath -- nickg [at] modp [dot] com. All rights
reserved.

Numeric decoder derived from TCL library
https://www.opensource.apple.com/source/tcl/tcl-14/tcl/license.terms
* Copyright (c) 1988-1993 The Regents of the University of California.
* Copyright (c) 1994 Sun Microsystems, Inc.
*/

// Licence at LICENSES/ULTRAJSON_LICENSE

#define PY_SSIZE_T_CLEAN
#include <Python.h>

#define NO_IMPORT_ARRAY
#define PY_ARRAY_UNIQUE_SYMBOL UJSON_NUMPY
#include "datetime.h"
#include "pandas/datetime/pd_datetime.h"
#include "pandas/vendored/ujson/lib/ultrajson.h"
#include <numpy/arrayobject.h>
#include <numpy/arrayscalars.h>
#include <numpy/ndarraytypes.h>
#include <numpy/npy_math.h>

static const int CSTR_SIZE = 20;

npy_int64 get_nat(void) { return NPY_MIN_INT64; }

typedef const char *(*PFN_PyTypeToUTF8)(JSOBJ obj, JSONTypeContext *ti,
                                        size_t *_outLen);

int object_is_decimal_type(PyObject *obj);
int object_is_dataframe_type(PyObject *obj);
int object_is_series_type(PyObject *obj);
int object_is_index_type(PyObject *obj);
int object_is_nat_type(PyObject *obj);
int object_is_na_type(PyObject *obj);

typedef struct __NpyArrContext {
  PyObject *array;
  char *dataptr;
  npy_intp curdim;    // current dimension in array's order
  npy_intp stridedim; // dimension we are striding over
  int inc;            // stride dimension increment (+/- 1)
  npy_intp dim;
  npy_intp stride;
  npy_intp ndim;
  npy_intp index[NPY_MAXDIMS];
  int type_num;

  char **rowLabels;
  char **columnLabels;
} NpyArrContext;

typedef struct __PdBlockContext {
  Py_ssize_t colIdx;
  Py_ssize_t ncols;
  int transpose;

  NpyArrContext **npyCtxts; // NpyArrContext for each column
} PdBlockContext;

typedef struct __TypeContext {
  JSPFN_ITERBEGIN iterBegin;
  JSPFN_ITEREND iterEnd;
  JSPFN_ITERNEXT iterNext;
  JSPFN_ITERGETNAME iterGetName;
  JSPFN_ITERGETVALUE iterGetValue;
  PFN_PyTypeToUTF8 PyTypeToUTF8;
  PyObject *newObj;
  PyObject *dictObj;
  Py_ssize_t index;
  Py_ssize_t size;
  PyObject *itemValue;
  PyObject *itemName;
  PyObject *attrList;
  PyObject *iterator;

  double doubleValue;
  JSINT64 longValue;

  char *cStr;
  NpyArrContext *npyarr;
  PdBlockContext *pdblock;
  int transpose;
  char **rowLabels;
  char **columnLabels;
  npy_intp rowLabelsLen;
  npy_intp columnLabelsLen;
} TypeContext;

typedef struct __PyObjectEncoder {
  JSONObjectEncoder enc;

  // pass through the NpyArrContext when encoding multi-dimensional arrays
  NpyArrContext *npyCtxtPassthru;

  // pass through the PdBlockContext when encoding blocks
  PdBlockContext *blkCtxtPassthru;

  // pass-through to encode numpy data directly
  int npyType;
  void *npyValue;

  // whether to encode datetime as ISO strings (0=epoch int, 1=ISO str)
  int datetimeIso;
  // the unit to encode to, governs epoch and ISO8601 precision (default ms)
  NPY_DATETIMEUNIT datetimeUnit;
  // pass-through: the unit of the actual integer value being encoded
  // (has to be set when calling NpyDateTimeToIsoCallback or
  // NpyTimeDeltaToIsoCallback)
  NPY_DATETIMEUNIT valueUnit;

  // output format style for pandas data types
  int outputFormat;
  int originalOutputFormat;

  PyObject *defaultHandler;
} PyObjectEncoder;

#define GET_TC(__ptrtc) ((TypeContext *)((__ptrtc)->prv))

enum PANDAS_FORMAT { SPLIT, RECORDS, INDEX, COLUMNS, VALUES };

static int PdBlock_iterNext(JSOBJ, JSONTypeContext *);

static TypeContext *createTypeContext(void) {
  TypeContext *pc = PyObject_Malloc(sizeof(TypeContext));
  if (!pc) {
    PyErr_NoMemory();
    return NULL;
  }
  pc->newObj = NULL;
  pc->dictObj = NULL;
  pc->itemValue = NULL;
  pc->itemName = NULL;
  pc->attrList = NULL;
  pc->index = 0;
  pc->size = 0;
  pc->longValue = 0;
  pc->doubleValue = 0.0;
  pc->cStr = NULL;
  pc->npyarr = NULL;
  pc->pdblock = NULL;
  pc->rowLabels = NULL;
  pc->columnLabels = NULL;
  pc->transpose = 0;
  pc->rowLabelsLen = 0;
  pc->columnLabelsLen = 0;

  return pc;
}

static PyObject *get_values(PyObject *obj) {
  PyObject *values = NULL;

  if (object_is_index_type(obj) || object_is_series_type(obj)) {
    // The special cases to worry about are dt64tz and category[dt64tz].
    //  In both cases we want the UTC-localized datetime64 ndarray,
    //  without going through and object array of Timestamps.
    if (PyObject_HasAttrString(obj, "tz")) {
      PyObject *tz = PyObject_GetAttrString(obj, "tz");
      if (tz != Py_None) {
        // Go through object array if we have dt64tz, since tz info will
        // be lost if values is used directly.
        Py_DECREF(tz);
        values = PyObject_CallMethod(obj, "__array__", NULL);
        return values;
      }
      Py_DECREF(tz);
    }
    values = PyObject_GetAttrString(obj, "values");
    if (values == NULL) {
      // Clear so we can subsequently try another method
      PyErr_Clear();
    } else if (PyObject_HasAttrString(values, "_values_for_json")) {
      // We have gotten an ExtensionArray
      PyObject *array_values =
          PyObject_CallMethod(values, "_values_for_json", NULL);
      Py_DECREF(values);
      values = array_values;
    } else if (PyObject_HasAttrString(values, "__array__")) {
      // We may have gotten a Categorical or Sparse array so call np.array
      PyObject *array_values = PyObject_CallMethod(values, "__array__", NULL);
      Py_DECREF(values);
      values = array_values;
    } else if (!PyArray_CheckExact(values)) {
      // Didn't get a numpy array, so keep trying
      Py_DECREF(values);
      values = NULL;
    }
  }

  if (values == NULL) {
    PyObject *typeRepr = PyObject_Repr((PyObject *)Py_TYPE(obj));
    PyObject *repr;
    if (PyObject_HasAttrString(obj, "dtype")) {
      PyObject *dtype = PyObject_GetAttrString(obj, "dtype");
      repr = PyObject_Repr(dtype);
      Py_DECREF(dtype);
    } else {
      repr = PyUnicode_FromString("<unknown dtype>");
    }

    PyErr_Format(PyExc_ValueError, "%R or %R are not JSON serializable yet",
                 repr, typeRepr);
    Py_DECREF(repr);
    Py_DECREF(typeRepr);

    return NULL;
  }

  return values;
}

static PyObject *get_sub_attr(PyObject *obj, char *attr, char *subAttr) {
  PyObject *tmp = PyObject_GetAttrString(obj, attr);
  if (tmp == 0) {
    return 0;
  }
  PyObject *ret = PyObject_GetAttrString(tmp, subAttr);
  Py_DECREF(tmp);

  return ret;
}

static Py_ssize_t get_attr_length(PyObject *obj, char *attr) {
  PyObject *tmp = PyObject_GetAttrString(obj, attr);
  if (tmp == 0) {
    return 0;
  }
  Py_ssize_t ret = PyObject_Length(tmp);
  Py_DECREF(tmp);

  if (ret == -1) {
    return 0;
  }

  return ret;
}

static npy_int64 get_long_attr(PyObject *o, const char *attr) {
  // NB we are implicitly assuming that o is a Timedelta or Timestamp, or NaT

  PyObject *value = PyObject_GetAttrString(o, attr);
  const npy_int64 long_val =
      (PyLong_Check(value) ? PyLong_AsLongLong(value) : PyLong_AsLong(value));

  Py_DECREF(value);

  if (object_is_nat_type(o)) {
    // i.e. o is NaT, long_val will be NPY_MIN_INT64
    return long_val;
  }

  // ensure we are in nanoseconds, similar to Timestamp._as_creso or _as_unit
  PyObject *reso = PyObject_GetAttrString(o, "_creso");
  if (!PyLong_Check(reso)) {
    // https://github.com/pandas-dev/pandas/pull/49034#discussion_r1023165139
    Py_DECREF(reso);
    return -1;
  }

  long cReso = PyLong_AsLong(reso);
  Py_DECREF(reso);
  if (cReso == -1 && PyErr_Occurred()) {
    return -1;
  }

  if (cReso == NPY_FR_us) {
    return long_val * 1000L;
  } else if (cReso == NPY_FR_ms) {
    return long_val * 1000000L;
  } else if (cReso == NPY_FR_s) {
    return long_val * 1000000000L;
  }

  return long_val;
}

static npy_float64 total_seconds(PyObject *td) {
  PyObject *value = PyObject_CallMethod(td, "total_seconds", NULL);
  const npy_float64 double_val = PyFloat_AS_DOUBLE(value);
  Py_DECREF(value);
  return double_val;
}

static const char *PyBytesToUTF8(JSOBJ _obj, JSONTypeContext *Py_UNUSED(tc),
                                 size_t *_outLen) {
  PyObject *obj = (PyObject *)_obj;
  *_outLen = PyBytes_GET_SIZE(obj);
  return PyBytes_AS_STRING(obj);
}

static const char *PyUnicodeToUTF8(JSOBJ _obj, JSONTypeContext *tc,
                                   size_t *_outLen) {
  char *encoded = (char *)PyUnicode_AsUTF8AndSize(_obj, (Py_ssize_t *)_outLen);
  if (encoded == NULL) {
    /* Something went wrong.
      Set errorMsg(to tell encoder to stop),
      and let Python exception propagate. */
    JSONObjectEncoder *enc = (JSONObjectEncoder *)tc->encoder;
    enc->errorMsg = "Encoding failed.";
  }
  return encoded;
}

/* JSON callback. returns a char* and mutates the pointer to *len */
static const char *NpyDateTimeToIsoCallback(JSOBJ Py_UNUSED(unused),
                                            JSONTypeContext *tc, size_t *len) {
  NPY_DATETIMEUNIT base = ((PyObjectEncoder *)tc->encoder)->datetimeUnit;
  NPY_DATETIMEUNIT valueUnit = ((PyObjectEncoder *)tc->encoder)->valueUnit;
  GET_TC(tc)->cStr = int64ToIso(GET_TC(tc)->longValue, valueUnit, base, len);
  return GET_TC(tc)->cStr;
}

/* JSON callback. returns a char* and mutates the pointer to *len */
static const char *NpyTimeDeltaToIsoCallback(JSOBJ Py_UNUSED(unused),
                                             JSONTypeContext *tc, size_t *len) {
  NPY_DATETIMEUNIT valueUnit = ((PyObjectEncoder *)tc->encoder)->valueUnit;
  GET_TC(tc)->cStr = int64ToIsoDuration(GET_TC(tc)->longValue, valueUnit, len);
  return GET_TC(tc)->cStr;
}

/* JSON callback */
static const char *PyDateTimeToIsoCallback(JSOBJ obj, JSONTypeContext *tc,
                                           size_t *len) {
  if (!PyDate_Check(obj) && !PyDateTime_Check(obj)) {
    PyErr_SetString(PyExc_TypeError, "Expected date or datetime object");
    ((JSONObjectEncoder *)tc->encoder)->errorMsg = "";
    return NULL;
  }

  NPY_DATETIMEUNIT base = ((PyObjectEncoder *)tc->encoder)->datetimeUnit;
  GET_TC(tc)->cStr = PyDateTimeToIso(obj, base, len);
  return GET_TC(tc)->cStr;
}

static const char *PyTimeToJSON(JSOBJ _obj, JSONTypeContext *tc,
                                size_t *outLen) {
  PyObject *obj = (PyObject *)_obj;
  PyObject *str = PyObject_CallMethod(obj, "isoformat", NULL);
  if (str == NULL) {
    *outLen = 0;
    if (!PyErr_Occurred()) {
      PyErr_SetString(PyExc_ValueError, "Failed to convert time");
    }
    ((JSONObjectEncoder *)tc->encoder)->errorMsg = "";
    return NULL;
  }
  if (PyUnicode_Check(str)) {
    PyObject *tmp = str;
    str = PyUnicode_AsUTF8String(str);
    Py_DECREF(tmp);
  }

  GET_TC(tc)->newObj = str;

  *outLen = PyBytes_GET_SIZE(str);
  char *outValue = PyBytes_AS_STRING(str);
  return outValue;
}

static const char *PyDecimalToUTF8Callback(JSOBJ _obj, JSONTypeContext *tc,
                                           size_t *len) {
  PyObject *obj = (PyObject *)_obj;
  PyObject *format_spec = PyUnicode_FromStringAndSize("f", 1);
  PyObject *str = PyObject_Format(obj, format_spec);
  Py_DECREF(format_spec);

  if (str == NULL) {
    ((JSONObjectEncoder *)tc->encoder)->errorMsg = "";
    return NULL;
  }

  GET_TC(tc)->newObj = str;

  Py_ssize_t s_len;
  char *outValue = (char *)PyUnicode_AsUTF8AndSize(str, &s_len);
  *len = s_len;

  return outValue;
}

//=============================================================================
// Numpy array iteration functions
//=============================================================================

static void NpyArr_freeItemValue(JSOBJ Py_UNUSED(_obj), JSONTypeContext *tc) {
  if (GET_TC(tc)->npyarr &&
      GET_TC(tc)->itemValue != GET_TC(tc)->npyarr->array) {
    Py_XDECREF(GET_TC(tc)->itemValue);
    GET_TC(tc)->itemValue = NULL;
  }
}

static int NpyArr_iterNextNone(JSOBJ Py_UNUSED(_obj),
                               JSONTypeContext *Py_UNUSED(tc)) {
  return 0;
}

static void NpyArr_iterBegin(JSOBJ _obj, JSONTypeContext *tc) {
  PyArrayObject *obj =
      (PyArrayObject *)(GET_TC(tc)->newObj ? GET_TC(tc)->newObj : _obj);

  NpyArrContext *npyarr = PyObject_Malloc(sizeof(NpyArrContext));
  GET_TC(tc)->npyarr = npyarr;

  if (!npyarr) {
    PyErr_NoMemory();
    GET_TC(tc)->iterNext = NpyArr_iterNextNone;
    return;
  }

  npyarr->array = (PyObject *)obj;
  npyarr->dataptr = PyArray_DATA(obj);
  npyarr->ndim = PyArray_NDIM(obj) - 1;
  npyarr->curdim = 0;
  npyarr->type_num = PyArray_DESCR(obj)->type_num;

  if (GET_TC(tc)->transpose) {
    npyarr->dim = PyArray_DIM(obj, (int)npyarr->ndim);
    npyarr->stride = PyArray_STRIDE(obj, (int)npyarr->ndim);
    npyarr->stridedim = npyarr->ndim;
    npyarr->index[npyarr->ndim] = 0;
    npyarr->inc = -1;
  } else {
    npyarr->dim = PyArray_DIM(obj, 0);
    npyarr->stride = PyArray_STRIDE(obj, 0);
    npyarr->stridedim = 0;
    npyarr->index[0] = 0;
    npyarr->inc = 1;
  }

  npyarr->columnLabels = GET_TC(tc)->columnLabels;
  npyarr->rowLabels = GET_TC(tc)->rowLabels;
}

static void NpyArr_iterEnd(JSOBJ obj, JSONTypeContext *tc) {
  NpyArrContext *npyarr = GET_TC(tc)->npyarr;

  if (npyarr) {
    NpyArr_freeItemValue(obj, tc);
    PyObject_Free(npyarr);
  }
}

static void NpyArrPassThru_iterBegin(JSOBJ Py_UNUSED(obj),
                                     JSONTypeContext *Py_UNUSED(tc)) {}

static void NpyArrPassThru_iterEnd(JSOBJ obj, JSONTypeContext *tc) {
  NpyArrContext *npyarr = GET_TC(tc)->npyarr;
  // finished this dimension, reset the data pointer
  npyarr->curdim--;
  npyarr->dataptr -= npyarr->stride * npyarr->index[npyarr->stridedim];
  npyarr->stridedim -= npyarr->inc;

  if (!PyArray_Check(npyarr->array)) {
    PyErr_SetString(PyExc_TypeError,
                    "NpyArrayPassThru_iterEnd received a non-array object");
    return;
  }
  const PyArrayObject *arrayobj = (const PyArrayObject *)npyarr->array;
  npyarr->dim = PyArray_DIM(arrayobj, (int)npyarr->stridedim);
  npyarr->stride = PyArray_STRIDE(arrayobj, (int)npyarr->stridedim);
  npyarr->dataptr += npyarr->stride;

  NpyArr_freeItemValue(obj, tc);
}

static int NpyArr_iterNextItem(JSOBJ obj, JSONTypeContext *tc) {
  NpyArrContext *npyarr = GET_TC(tc)->npyarr;

  if (PyErr_Occurred()) {
    return 0;
  }

  if (npyarr->index[npyarr->stridedim] >= npyarr->dim) {
    return 0;
  }

  NpyArr_freeItemValue(obj, tc);

  if (!PyArray_Check(npyarr->array)) {
    PyErr_SetString(PyExc_TypeError,
                    "NpyArr_iterNextItem received a non-array object");
    return 0;
  }
  PyArrayObject *arrayobj = (PyArrayObject *)npyarr->array;

  if (PyArray_ISDATETIME(arrayobj)) {
    GET_TC(tc)->itemValue = obj;
    Py_INCREF(obj);
    ((PyObjectEncoder *)tc->encoder)->npyType = PyArray_TYPE(arrayobj);
    // Also write the resolution (unit) of the ndarray
    PyArray_Descr *dtype = PyArray_DESCR(arrayobj);
    ((PyObjectEncoder *)tc->encoder)->valueUnit =
        get_datetime_metadata_from_dtype(dtype).base;
    ((PyObjectEncoder *)tc->encoder)->npyValue = npyarr->dataptr;
    ((PyObjectEncoder *)tc->encoder)->npyCtxtPassthru = npyarr;
  } else {
    GET_TC(tc)->itemValue = PyArray_GETITEM(arrayobj, npyarr->dataptr);
  }

  npyarr->dataptr += npyarr->stride;
  npyarr->index[npyarr->stridedim]++;
  return 1;
}

static int NpyArr_iterNext(JSOBJ _obj, JSONTypeContext *tc) {
  NpyArrContext *npyarr = GET_TC(tc)->npyarr;

  if (PyErr_Occurred()) {
    return 0;
  }

  if (npyarr->curdim >= npyarr->ndim ||
      npyarr->index[npyarr->stridedim] >= npyarr->dim) {
    // innermost dimension, start retrieving item values
    GET_TC(tc)->iterNext = NpyArr_iterNextItem;
    return NpyArr_iterNextItem(_obj, tc);
  }

  // dig a dimension deeper
  npyarr->index[npyarr->stridedim]++;

  npyarr->curdim++;
  npyarr->stridedim += npyarr->inc;
  if (!PyArray_Check(npyarr->array)) {
    PyErr_SetString(PyExc_TypeError,
                    "NpyArr_iterNext received a non-array object");
    return 0;
  }
  const PyArrayObject *arrayobj = (const PyArrayObject *)npyarr->array;

  npyarr->dim = PyArray_DIM(arrayobj, (int)npyarr->stridedim);
  npyarr->stride = PyArray_STRIDE(arrayobj, (int)npyarr->stridedim);
  npyarr->index[npyarr->stridedim] = 0;

  ((PyObjectEncoder *)tc->encoder)->npyCtxtPassthru = npyarr;
  GET_TC(tc)->itemValue = npyarr->array;
  return 1;
}

static JSOBJ NpyArr_iterGetValue(JSOBJ Py_UNUSED(obj), JSONTypeContext *tc) {
  return GET_TC(tc)->itemValue;
}

static const char *NpyArr_iterGetName(JSOBJ Py_UNUSED(obj), JSONTypeContext *tc,
                                      size_t *outLen) {
  NpyArrContext *npyarr = GET_TC(tc)->npyarr;
  const char *cStr;

  if (GET_TC(tc)->iterNext == NpyArr_iterNextItem) {
    const npy_intp idx = npyarr->index[npyarr->stridedim] - 1;
    cStr = npyarr->columnLabels[idx];
  } else {
    const npy_intp idx = npyarr->index[npyarr->stridedim - npyarr->inc] - 1;
    cStr = npyarr->rowLabels[idx];
  }

  *outLen = strlen(cStr);

  return cStr;
}

//=============================================================================
// Pandas block iteration functions
//
// Serialises a DataFrame column by column to avoid unnecessary data copies and
// more representative serialisation when dealing with mixed dtypes.
//
// Uses a dedicated NpyArrContext for each column.
//=============================================================================

static void PdBlockPassThru_iterEnd(JSOBJ obj, JSONTypeContext *tc) {
  PdBlockContext *blkCtxt = GET_TC(tc)->pdblock;

  if (blkCtxt->transpose) {
    blkCtxt->colIdx++;
  } else {
    blkCtxt->colIdx = 0;
  }

  NpyArr_freeItemValue(obj, tc);
}

static int PdBlock_iterNextItem(JSOBJ obj, JSONTypeContext *tc) {
  PdBlockContext *blkCtxt = GET_TC(tc)->pdblock;

  if (blkCtxt->colIdx >= blkCtxt->ncols) {
    return 0;
  }

  GET_TC(tc)->npyarr = blkCtxt->npyCtxts[blkCtxt->colIdx];
  blkCtxt->colIdx++;
  return NpyArr_iterNextItem(obj, tc);
}

static const char *PdBlock_iterGetName(JSOBJ Py_UNUSED(obj),
                                       JSONTypeContext *tc, size_t *outLen) {
  PdBlockContext *blkCtxt = GET_TC(tc)->pdblock;
  NpyArrContext *npyarr = blkCtxt->npyCtxts[0];
  const char *cStr;

  if (GET_TC(tc)->iterNext == PdBlock_iterNextItem) {
    const npy_intp idx = blkCtxt->colIdx - 1;
    cStr = npyarr->columnLabels[idx];
  } else {
    const npy_intp idx =
        GET_TC(tc)->iterNext != PdBlock_iterNext
            ? npyarr->index[npyarr->stridedim - npyarr->inc] - 1
            : npyarr->index[npyarr->stridedim];

    cStr = npyarr->rowLabels[idx];
  }

  *outLen = strlen(cStr);
  return cStr;
}

static const char *PdBlock_iterGetName_Transpose(JSOBJ Py_UNUSED(obj),
                                                 JSONTypeContext *tc,
                                                 size_t *outLen) {
  PdBlockContext *blkCtxt = GET_TC(tc)->pdblock;
  NpyArrContext *npyarr = blkCtxt->npyCtxts[blkCtxt->colIdx];
  const char *cStr;

  if (GET_TC(tc)->iterNext == NpyArr_iterNextItem) {
    const npy_intp idx = npyarr->index[npyarr->stridedim] - 1;
    cStr = npyarr->columnLabels[idx];
  } else {
    const npy_intp idx = blkCtxt->colIdx;
    cStr = npyarr->rowLabels[idx];
  }

  *outLen = strlen(cStr);
  return cStr;
}

static int PdBlock_iterNext(JSOBJ obj, JSONTypeContext *tc) {
  PdBlockContext *blkCtxt = GET_TC(tc)->pdblock;

  if (PyErr_Occurred() || ((JSONObjectEncoder *)tc->encoder)->errorMsg) {
    return 0;
  }

  if (blkCtxt->transpose) {
    if (blkCtxt->colIdx >= blkCtxt->ncols) {
      return 0;
    }
  } else {
    const NpyArrContext *npyarr = blkCtxt->npyCtxts[0];
    if (npyarr->index[npyarr->stridedim] >= npyarr->dim) {
      return 0;
    }
  }

  ((PyObjectEncoder *)tc->encoder)->blkCtxtPassthru = blkCtxt;
  GET_TC(tc)->itemValue = obj;

  return 1;
}

static void PdBlockPassThru_iterBegin(JSOBJ Py_UNUSED(obj),
                                      JSONTypeContext *tc) {
  PdBlockContext *blkCtxt = GET_TC(tc)->pdblock;

  if (blkCtxt->transpose) {
    // if transposed we exhaust each column before moving to the next
    GET_TC(tc)->iterNext = NpyArr_iterNextItem;
    GET_TC(tc)->iterGetName = PdBlock_iterGetName_Transpose;
    GET_TC(tc)->npyarr = blkCtxt->npyCtxts[blkCtxt->colIdx];
  }
}

static void PdBlock_iterBegin(JSOBJ _obj, JSONTypeContext *tc) {
  PyObject *obj = (PyObject *)_obj;

  GET_TC(tc)->iterGetName = GET_TC(tc)->transpose
                                ? PdBlock_iterGetName_Transpose
                                : PdBlock_iterGetName;

  PdBlockContext *blkCtxt = PyObject_Malloc(sizeof(PdBlockContext));
  if (!blkCtxt) {
    PyErr_NoMemory();
    GET_TC(tc)->iterNext = NpyArr_iterNextNone;
    return;
  }
  GET_TC(tc)->pdblock = blkCtxt;

  blkCtxt->colIdx = 0;
  blkCtxt->transpose = GET_TC(tc)->transpose;
  blkCtxt->ncols = get_attr_length(obj, "columns");

  if (blkCtxt->ncols == 0) {
    blkCtxt->npyCtxts = NULL;

    GET_TC(tc)->iterNext = NpyArr_iterNextNone;
    return;
  }

  blkCtxt->npyCtxts = PyObject_Malloc(sizeof(NpyArrContext *) * blkCtxt->ncols);
  if (!blkCtxt->npyCtxts) {
    PyErr_NoMemory();
    GET_TC(tc)->iterNext = NpyArr_iterNextNone;
    return;
  }

  PyObject *arrays = get_sub_attr(obj, "_mgr", "column_arrays");
  if (!arrays) {
    GET_TC(tc)->iterNext = NpyArr_iterNextNone;
    return;
  }

  for (Py_ssize_t i = 0; i < PyObject_Length(arrays); i++) {
    PyObject *array = PyList_GET_ITEM(arrays, i);
    if (!array) {
      GET_TC(tc)->iterNext = NpyArr_iterNextNone;
      goto ARR_RET;
    }

    // ensure we have a numpy array (i.e. np.asarray)
    PyObject *values = PyObject_CallMethod(array, "__array__", NULL);
    if ((!values) || (!PyArray_CheckExact(values))) {
      // Didn't get a numpy array
      ((JSONObjectEncoder *)tc->encoder)->errorMsg = "";
      GET_TC(tc)->iterNext = NpyArr_iterNextNone;
      goto ARR_RET;
    }

    GET_TC(tc)->newObj = values;

    // init a dedicated context for this column
    NpyArr_iterBegin(obj, tc);

    GET_TC(tc)->itemValue = NULL;
    ((PyObjectEncoder *)tc->encoder)->npyCtxtPassthru = NULL;

    blkCtxt->npyCtxts[i] = GET_TC(tc)->npyarr;
    GET_TC(tc)->newObj = NULL;
  }
  GET_TC(tc)->npyarr = blkCtxt->npyCtxts[0];
  goto ARR_RET;

ARR_RET:
  Py_DECREF(arrays);
}

static void PdBlock_iterEnd(JSOBJ obj, JSONTypeContext *tc) {
  GET_TC(tc)->itemValue = NULL;
  NpyArrContext *npyarr = GET_TC(tc)->npyarr;
  PdBlockContext *blkCtxt = GET_TC(tc)->pdblock;

  if (blkCtxt) {
    for (int i = 0; i < blkCtxt->ncols; i++) {
      npyarr = blkCtxt->npyCtxts[i];
      if (npyarr) {
        if (npyarr->array) {
          Py_DECREF(npyarr->array);
          npyarr->array = NULL;
        }

        GET_TC(tc)->npyarr = npyarr;
        NpyArr_iterEnd(obj, tc);

        blkCtxt->npyCtxts[i] = NULL;
      }
    }

    if (blkCtxt->npyCtxts) {
      PyObject_Free(blkCtxt->npyCtxts);
    }
    PyObject_Free(blkCtxt);
  }
}

//=============================================================================
// Tuple iteration functions
// itemValue is borrowed reference, no ref counting
//=============================================================================
static void Tuple_iterBegin(JSOBJ obj, JSONTypeContext *tc) {
  GET_TC(tc)->index = 0;
  GET_TC(tc)->size = PyTuple_GET_SIZE((PyObject *)obj);
  GET_TC(tc)->itemValue = NULL;
}

static int Tuple_iterNext(JSOBJ obj, JSONTypeContext *tc) {

  if (GET_TC(tc)->index >= GET_TC(tc)->size) {
    return 0;
  }

  PyObject *item = PyTuple_GET_ITEM(obj, GET_TC(tc)->index);

  GET_TC(tc)->itemValue = item;
  GET_TC(tc)->index++;
  return 1;
}

static void Tuple_iterEnd(JSOBJ Py_UNUSED(obj),
                          JSONTypeContext *Py_UNUSED(tc)) {}

static JSOBJ Tuple_iterGetValue(JSOBJ Py_UNUSED(obj), JSONTypeContext *tc) {
  return GET_TC(tc)->itemValue;
}

static const char *Tuple_iterGetName(JSOBJ Py_UNUSED(obj),
                                     JSONTypeContext *Py_UNUSED(tc),
                                     size_t *Py_UNUSED(outLen)) {
  return NULL;
}

//=============================================================================
// Set iteration functions
// itemValue is borrowed reference, no ref counting
//=============================================================================
static void Set_iterBegin(JSOBJ obj, JSONTypeContext *tc) {
  GET_TC(tc)->itemValue = NULL;
  GET_TC(tc)->iterator = PyObject_GetIter(obj);
}

static int Set_iterNext(JSOBJ Py_UNUSED(obj), JSONTypeContext *tc) {
  if (GET_TC(tc)->itemValue) {
    Py_DECREF(GET_TC(tc)->itemValue);
    GET_TC(tc)->itemValue = NULL;
  }

  PyObject *item = PyIter_Next(GET_TC(tc)->iterator);

  if (item == NULL) {
    return 0;
  }

  GET_TC(tc)->itemValue = item;
  return 1;
}

static void Set_iterEnd(JSOBJ Py_UNUSED(obj), JSONTypeContext *tc) {
  if (GET_TC(tc)->itemValue) {
    Py_DECREF(GET_TC(tc)->itemValue);
    GET_TC(tc)->itemValue = NULL;
  }

  if (GET_TC(tc)->iterator) {
    Py_DECREF(GET_TC(tc)->iterator);
    GET_TC(tc)->iterator = NULL;
  }
}

static JSOBJ Set_iterGetValue(JSOBJ Py_UNUSED(obj), JSONTypeContext *tc) {
  return GET_TC(tc)->itemValue;
}

static const char *Set_iterGetName(JSOBJ Py_UNUSED(obj),
                                   JSONTypeContext *Py_UNUSED(tc),
                                   size_t *Py_UNUSED(outLen)) {
  return NULL;
}

//=============================================================================
// Dir iteration functions
// itemName ref is borrowed from PyObject_Dir (attrList). No refcount
// itemValue ref is from PyObject_GetAttr. Ref counted
//=============================================================================
static void Dir_iterBegin(JSOBJ obj, JSONTypeContext *tc) {
  GET_TC(tc)->attrList = PyObject_Dir(obj);
  GET_TC(tc)->index = 0;
  GET_TC(tc)->size = PyList_GET_SIZE(GET_TC(tc)->attrList);
}

static void Dir_iterEnd(JSOBJ Py_UNUSED(obj), JSONTypeContext *tc) {
  if (GET_TC(tc)->itemValue) {
    Py_DECREF(GET_TC(tc)->itemValue);
    GET_TC(tc)->itemValue = NULL;
  }

  if (GET_TC(tc)->itemName) {
    Py_DECREF(GET_TC(tc)->itemName);
    GET_TC(tc)->itemName = NULL;
  }

  Py_DECREF((PyObject *)GET_TC(tc)->attrList);
}

static int Dir_iterNext(JSOBJ _obj, JSONTypeContext *tc) {
  PyObject *obj = (PyObject *)_obj;
  PyObject *itemValue = GET_TC(tc)->itemValue;
  PyObject *itemName = GET_TC(tc)->itemName;

  if (PyErr_Occurred() || ((JSONObjectEncoder *)tc->encoder)->errorMsg) {
    return 0;
  }

  if (itemValue) {
    Py_DECREF(GET_TC(tc)->itemValue);
    GET_TC(tc)->itemValue = itemValue = NULL;
  }

  if (itemName) {
    Py_DECREF(GET_TC(tc)->itemName);
    GET_TC(tc)->itemName = itemName = NULL;
  }

  for (; GET_TC(tc)->index < GET_TC(tc)->size; GET_TC(tc)->index++) {
    PyObject *attrName =
        PyList_GET_ITEM(GET_TC(tc)->attrList, GET_TC(tc)->index);
    PyObject *attr = PyUnicode_AsUTF8String(attrName);
    const char *attrStr = PyBytes_AS_STRING(attr);

    if (attrStr[0] == '_') {
      Py_DECREF(attr);
      continue;
    }

    itemValue = PyObject_GetAttr(obj, attrName);
    if (itemValue == NULL) {
      PyErr_Clear();
      Py_DECREF(attr);
      continue;
    }

    if (PyCallable_Check(itemValue)) {
      Py_DECREF(itemValue);
      Py_DECREF(attr);
      continue;
    }

    GET_TC(tc)->itemName = itemName;
    GET_TC(tc)->itemValue = itemValue;

    itemName = attr;
    break;
  }

  if (itemName == NULL) {
    GET_TC(tc)->index = GET_TC(tc)->size;
    GET_TC(tc)->itemValue = NULL;
    return 0;
  }

  GET_TC(tc)->itemName = itemName;
  GET_TC(tc)->itemValue = itemValue;
  GET_TC(tc)->index++;

  return 1;
}

static JSOBJ Dir_iterGetValue(JSOBJ Py_UNUSED(obj), JSONTypeContext *tc) {
  return GET_TC(tc)->itemValue;
}

static const char *Dir_iterGetName(JSOBJ Py_UNUSED(obj), JSONTypeContext *tc,
                                   size_t *outLen) {
  *outLen = PyBytes_GET_SIZE(GET_TC(tc)->itemName);
  return PyBytes_AS_STRING(GET_TC(tc)->itemName);
}

//=============================================================================
// List iteration functions
// itemValue is borrowed from object (which is list). No refcounting
//=============================================================================
static void List_iterBegin(JSOBJ obj, JSONTypeContext *tc) {
  GET_TC(tc)->index = 0;
  GET_TC(tc)->size = PyList_GET_SIZE((PyObject *)obj);
}

static int List_iterNext(JSOBJ obj, JSONTypeContext *tc) {
  if (GET_TC(tc)->index >= GET_TC(tc)->size) {
    return 0;
  }

  GET_TC(tc)->itemValue = PyList_GET_ITEM(obj, GET_TC(tc)->index);
  GET_TC(tc)->index++;
  return 1;
}

static void List_iterEnd(JSOBJ Py_UNUSED(obj), JSONTypeContext *Py_UNUSED(tc)) {
}

static JSOBJ List_iterGetValue(JSOBJ Py_UNUSED(obj), JSONTypeContext *tc) {
  return GET_TC(tc)->itemValue;
}

static const char *List_iterGetName(JSOBJ Py_UNUSED(obj),
                                    JSONTypeContext *Py_UNUSED(tc),
                                    size_t *Py_UNUSED(outLen)) {
  return NULL;
}

//=============================================================================
// pandas Index iteration functions
//=============================================================================
static void Index_iterBegin(JSOBJ Py_UNUSED(obj), JSONTypeContext *tc) {
  GET_TC(tc)->index = 0;
  GET_TC(tc)->cStr = PyObject_Malloc(CSTR_SIZE);
  if (!GET_TC(tc)->cStr) {
    PyErr_NoMemory();
  }
}

static int Index_iterNext(JSOBJ obj, JSONTypeContext *tc) {
  const Py_ssize_t index = GET_TC(tc)->index;
  Py_XDECREF(GET_TC(tc)->itemValue);
  if (!GET_TC(tc)->cStr) {
    return 0;
  }

  if (index == 0) {
    strcpy(GET_TC(tc)->cStr, "name");
    GET_TC(tc)->itemValue = PyObject_GetAttrString(obj, "name");
  } else if (index == 1) {
    strcpy(GET_TC(tc)->cStr, "data");
    GET_TC(tc)->itemValue = get_values(obj);
    if (!GET_TC(tc)->itemValue) {
      return 0;
    }
  } else {
    return 0;
  }

  GET_TC(tc)->index++;
  return 1;
}

static void Index_iterEnd(JSOBJ Py_UNUSED(obj),
                          JSONTypeContext *Py_UNUSED(tc)) {}

static JSOBJ Index_iterGetValue(JSOBJ Py_UNUSED(obj), JSONTypeContext *tc) {
  return GET_TC(tc)->itemValue;
}

static const char *Index_iterGetName(JSOBJ Py_UNUSED(obj), JSONTypeContext *tc,
                                     size_t *outLen) {
  *outLen = strlen(GET_TC(tc)->cStr);
  return GET_TC(tc)->cStr;
}

//=============================================================================
// pandas Series iteration functions
//=============================================================================
static void Series_iterBegin(JSOBJ Py_UNUSED(obj), JSONTypeContext *tc) {
  PyObjectEncoder *enc = (PyObjectEncoder *)tc->encoder;
  GET_TC(tc)->index = 0;
  enc->outputFormat = VALUES; // for contained series
  GET_TC(tc)->cStr = PyObject_Malloc(CSTR_SIZE);
  if (!GET_TC(tc)->cStr) {
    PyErr_NoMemory();
  }
}

static int Series_iterNext(JSOBJ obj, JSONTypeContext *tc) {
  const Py_ssize_t index = GET_TC(tc)->index;
  Py_XDECREF(GET_TC(tc)->itemValue);
  if (!GET_TC(tc)->cStr) {
    return 0;
  }

  if (index == 0) {
    strcpy(GET_TC(tc)->cStr, "name");
    GET_TC(tc)->itemValue = PyObject_GetAttrString(obj, "name");
  } else if (index == 1) {
    strcpy(GET_TC(tc)->cStr, "index");
    GET_TC(tc)->itemValue = PyObject_GetAttrString(obj, "index");
  } else if (index == 2) {
    strcpy(GET_TC(tc)->cStr, "data");
    GET_TC(tc)->itemValue = get_values(obj);
    if (!GET_TC(tc)->itemValue) {
      return 0;
    }
  } else {
    return 0;
  }

  GET_TC(tc)->index++;
  return 1;
}

static void Series_iterEnd(JSOBJ Py_UNUSED(obj), JSONTypeContext *tc) {
  PyObjectEncoder *enc = (PyObjectEncoder *)tc->encoder;
  enc->outputFormat = enc->originalOutputFormat;
}

static JSOBJ Series_iterGetValue(JSOBJ Py_UNUSED(obj), JSONTypeContext *tc) {
  return GET_TC(tc)->itemValue;
}

static const char *Series_iterGetName(JSOBJ Py_UNUSED(obj), JSONTypeContext *tc,
                                      size_t *outLen) {
  *outLen = strlen(GET_TC(tc)->cStr);
  return GET_TC(tc)->cStr;
}

//=============================================================================
// pandas DataFrame iteration functions
//=============================================================================
static void DataFrame_iterBegin(JSOBJ Py_UNUSED(obj), JSONTypeContext *tc) {
  PyObjectEncoder *enc = (PyObjectEncoder *)tc->encoder;
  GET_TC(tc)->index = 0;
  enc->outputFormat = VALUES; // for contained series & index
  GET_TC(tc)->cStr = PyObject_Malloc(CSTR_SIZE);
  if (!GET_TC(tc)->cStr) {
    PyErr_NoMemory();
  }
}

static int DataFrame_iterNext(JSOBJ obj, JSONTypeContext *tc) {
  const Py_ssize_t index = GET_TC(tc)->index;
  Py_XDECREF(GET_TC(tc)->itemValue);
  if (!GET_TC(tc)->cStr) {
    return 0;
  }

  if (index == 0) {
    strcpy(GET_TC(tc)->cStr, "columns");
    GET_TC(tc)->itemValue = PyObject_GetAttrString(obj, "columns");
  } else if (index == 1) {
    strcpy(GET_TC(tc)->cStr, "index");
    GET_TC(tc)->itemValue = PyObject_GetAttrString(obj, "index");
  } else if (index == 2) {
    strcpy(GET_TC(tc)->cStr, "data");
    Py_INCREF(obj);
    GET_TC(tc)->itemValue = obj;
  } else {
    return 0;
  }

  GET_TC(tc)->index++;
  return 1;
}

static void DataFrame_iterEnd(JSOBJ Py_UNUSED(obj), JSONTypeContext *tc) {
  PyObjectEncoder *enc = (PyObjectEncoder *)tc->encoder;
  enc->outputFormat = enc->originalOutputFormat;
}

static JSOBJ DataFrame_iterGetValue(JSOBJ Py_UNUSED(obj), JSONTypeContext *tc) {
  return GET_TC(tc)->itemValue;
}

static const char *DataFrame_iterGetName(JSOBJ Py_UNUSED(obj),
                                         JSONTypeContext *tc, size_t *outLen) {
  *outLen = strlen(GET_TC(tc)->cStr);
  return GET_TC(tc)->cStr;
}

//=============================================================================
// Dict iteration functions
// itemName might converted to string (Python_Str). Do refCounting
// itemValue is borrowed from object (which is dict). No refCounting
//=============================================================================
static void Dict_iterBegin(JSOBJ Py_UNUSED(obj), JSONTypeContext *tc) {
  GET_TC(tc)->index = 0;
}

static int Dict_iterNext(JSOBJ Py_UNUSED(obj), JSONTypeContext *tc) {
  if (GET_TC(tc)->itemName) {
    Py_DECREF(GET_TC(tc)->itemName);
    GET_TC(tc)->itemName = NULL;
  }

  if (!PyDict_Next((PyObject *)GET_TC(tc)->dictObj, &GET_TC(tc)->index,
                   &GET_TC(tc)->itemName, &GET_TC(tc)->itemValue)) {
    return 0;
  }

  if (PyUnicode_Check(GET_TC(tc)->itemName)) {
    GET_TC(tc)->itemName = PyUnicode_AsUTF8String(GET_TC(tc)->itemName);
  } else if (!PyBytes_Check(GET_TC(tc)->itemName)) {
    GET_TC(tc)->itemName = PyObject_Str(GET_TC(tc)->itemName);
    PyObject *itemNameTmp = GET_TC(tc)->itemName;
    GET_TC(tc)->itemName = PyUnicode_AsUTF8String(GET_TC(tc)->itemName);
    Py_DECREF(itemNameTmp);
  } else {
    Py_INCREF(GET_TC(tc)->itemName);
  }
  return 1;
}

static void Dict_iterEnd(JSOBJ Py_UNUSED(obj), JSONTypeContext *tc) {
  if (GET_TC(tc)->itemName) {
    Py_DECREF(GET_TC(tc)->itemName);
    GET_TC(tc)->itemName = NULL;
  }
  Py_DECREF(GET_TC(tc)->dictObj);
}

static JSOBJ Dict_iterGetValue(JSOBJ Py_UNUSED(obj), JSONTypeContext *tc) {
  return GET_TC(tc)->itemValue;
}

static const char *Dict_iterGetName(JSOBJ Py_UNUSED(obj), JSONTypeContext *tc,
                                    size_t *outLen) {
  *outLen = PyBytes_GET_SIZE(GET_TC(tc)->itemName);
  return PyBytes_AS_STRING(GET_TC(tc)->itemName);
}

static void NpyArr_freeLabels(char **labels, npy_intp len) {
  if (labels) {
    for (npy_intp i = 0; i < len; i++) {
      PyObject_Free(labels[i]);
    }
    PyObject_Free(labels);
  }
}

/*
 * Function: NpyArr_encodeLabels
 * -----------------------------
 *
 * Builds an array of "encoded" labels.
 *
 * labels: PyArrayObject pointer for labels to be "encoded"
 * num : number of labels
 *
 * "encode" is quoted above because we aren't really doing encoding
 * For historical reasons this function would actually encode the entire
 * array into a separate buffer with a separate call to JSON_Encode
 * and would leave it to complex pointer manipulation from there to
 * unpack values as needed. To make things simpler and more idiomatic
 * this has instead just stringified any input save for datetime values,
 * which may need to be represented in various formats.
 */
static char **NpyArr_encodeLabels(PyArrayObject *labels, PyObjectEncoder *enc,
                                  npy_intp num) {
  // NOTE this function steals a reference to labels.
  PyObject *item = NULL;
  const NPY_DATETIMEUNIT targetUnit = enc->datetimeUnit;

  if (!labels) {
    return 0;
  }

  if (PyArray_SIZE(labels) < num) {
    PyErr_SetString(PyExc_ValueError,
                    "Label array sizes do not match corresponding data shape");
    Py_DECREF(labels);
    return 0;
  }

  char **ret = PyObject_Malloc(sizeof(char *) * num);
  if (!ret) {
    PyErr_NoMemory();
    Py_DECREF(labels);
    return 0;
  }

  for (npy_intp i = 0; i < num; i++) {
    ret[i] = NULL;
  }

  const npy_intp stride = PyArray_STRIDE(labels, 0);
  char *dataptr = PyArray_DATA(labels);
  const int type_num = PyArray_TYPE(labels);
  PyArray_Descr *dtype = PyArray_DESCR(labels);

  for (npy_intp i = 0; i < num; i++) {
    item = PyArray_GETITEM(labels, dataptr);
    if (!item) {
      NpyArr_freeLabels(ret, num);
      ret = 0;
      break;
    }

    int is_datetimelike = 0;
    int from_numpy = 0;
    int64_t i8date;
    NPY_DATETIMEUNIT valueUnit = NPY_FR_ns;
    if (PyTypeNum_ISDATETIME(type_num)) {
      is_datetimelike = 1;
      from_numpy = 1;
      i8date = *(int64_t *)dataptr;
      valueUnit = get_datetime_metadata_from_dtype(dtype).base;
    } else if (PyDate_Check(item) || PyDelta_Check(item)) {
      // convert datetime/timedelta object to nanos
      is_datetimelike = 1;
      if (PyObject_HasAttrString(item, "_value")) {
        // pd.Timestamp object or pd.NaT
        // see test_date_index_and_values for case with non-nano
        i8date = get_long_attr(item, "_value");
      } else {
        if (PyDelta_Check(item)) {
          // TODO(anyone): cast below loses precision if total_seconds return
          // value exceeds number of bits that significand can hold
          // also liable to overflow
          i8date = (int64_t)(total_seconds(item) *
                             1000000000LL); // nanoseconds per second
        } else {
          // datetime.* objects don't follow above rules
          i8date = PyDateTimeToEpoch(item, NPY_FR_ns);
        }
      }
    }

    size_t len;
    char *cLabel;
    if (is_datetimelike) {
      if (i8date == get_nat()) {
        len = 4;
        cLabel = PyObject_Malloc(len + 1);
        strncpy(cLabel, "null", len + 1);
      } else {
        if (enc->datetimeIso) {
          if ((type_num == NPY_TIMEDELTA) || (PyDelta_Check(item))) {
            cLabel = int64ToIsoDuration(i8date, valueUnit, &len);
          } else {
            if (type_num == NPY_DATETIME) {
              cLabel = int64ToIso(i8date, valueUnit, targetUnit, &len);
            } else {
              cLabel = PyDateTimeToIso(item, targetUnit, &len);
            }
          }
          if (cLabel == NULL) {
            Py_DECREF(item);
            NpyArr_freeLabels(ret, num);
            ret = 0;
            break;
          }
        } else {
          int size_of_cLabel = 21; // 21 chars for int 64
          cLabel = PyObject_Malloc(size_of_cLabel);
          if (!from_numpy) {
            // numpy arrays are already scaled to the correct unit
            // only need to scale if coming from a datetime/timedelta object
            if (scaleNanosecToUnit(&i8date, targetUnit) == -1) {
              NpyArr_freeLabels(ret, num);
              ret = 0;
              break;
            }
          }
          snprintf(cLabel, size_of_cLabel, "%" PRId64, i8date);
          len = strlen(cLabel);
        }
      }
    } else { // Fallback to string representation
      // Replace item with the string to keep it alive.
      Py_SETREF(item, PyObject_Str(item));
      if (item == NULL) {
        NpyArr_freeLabels(ret, num);
        ret = 0;
        break;
      }

      cLabel = (char *)PyUnicode_AsUTF8(item);
      len = strlen(cLabel);
    }

    // Add 1 to include NULL terminator
    ret[i] = PyObject_Malloc(len + 1);
    memcpy(ret[i], cLabel, len + 1);
    Py_DECREF(item);

    if (is_datetimelike) {
      PyObject_Free(cLabel);
    }

    if (PyErr_Occurred()) {
      NpyArr_freeLabels(ret, num);
      ret = 0;
      break;
    }

    if (!ret[i]) {
      PyErr_NoMemory();
      ret = 0;
      break;
    }

    dataptr += stride;
  }

  Py_DECREF(labels);
  return ret;
}

static void Object_invokeDefaultHandler(PyObject *obj, PyObjectEncoder *enc) {
  PyObject *tmpObj = NULL;
  tmpObj = PyObject_CallFunctionObjArgs(enc->defaultHandler, obj, NULL);
  if (!PyErr_Occurred()) {
    if (tmpObj == NULL) {
      PyErr_SetString(PyExc_TypeError, "Failed to execute default handler");
    } else {
      encode(tmpObj, (JSONObjectEncoder *)enc, NULL, 0);
    }
  }
  Py_XDECREF(tmpObj);
  return;
}

static void Object_beginTypeContext(JSOBJ _obj, JSONTypeContext *tc) {
  tc->prv = NULL;

  if (!_obj) {
    tc->type = JT_INVALID;
    return;
  }

  PyObject *obj = (PyObject *)_obj;
  PyObjectEncoder *enc = (PyObjectEncoder *)tc->encoder;

  if (PyBool_Check(obj)) {
    tc->type = (obj == Py_True) ? JT_TRUE : JT_FALSE;
    return;
  } else if (obj == Py_None) {
    tc->type = JT_NULL;
    return;
  }

  TypeContext *pc = createTypeContext();
  if (!pc) {
    tc->type = JT_INVALID;
    return;
  }
  tc->prv = pc;

  if (PyTypeNum_ISDATETIME(enc->npyType)) {
    int64_t longVal = *(npy_int64 *)enc->npyValue;
    if (longVal == get_nat()) {
      tc->type = JT_NULL;
    } else {
      if (enc->datetimeIso) {
        if (enc->npyType == NPY_TIMEDELTA) {
          pc->PyTypeToUTF8 = NpyTimeDeltaToIsoCallback;
        } else {
          pc->PyTypeToUTF8 = NpyDateTimeToIsoCallback;
        }
        // Currently no way to pass longVal to iso function, so use
        // state management
        pc->longValue = longVal;
        tc->type = JT_UTF8;
      } else {
        // numpy array was already scaled to unit, so just use int value for
        // epoch
        pc->longValue = longVal;
        tc->type = JT_LONG;
      }
    }

    // TODO(username): this prevents infinite loop with
    // mixed-type DataFrames;
    // refactor
    enc->npyCtxtPassthru = NULL;
    enc->npyType = -1;
    return;
  }

  if (PyIter_Check(obj) || (PyArray_Check(obj) && !PyArray_CheckScalar(obj))) {
    goto ISITERABLE;
  }

  if (PyLong_Check(obj)) {
    tc->type = JT_LONG;
    int overflow = 0;
    pc->longValue = PyLong_AsLongLongAndOverflow(obj, &overflow);
    int err;
    err = (pc->longValue == -1) && PyErr_Occurred();

    if (overflow) {
      tc->type = JT_BIGNUM;
    } else if (err) {
      goto INVALID;
    }

    return;
  } else if (PyFloat_Check(obj)) {
    const double val = PyFloat_AS_DOUBLE(obj);
    if (npy_isnan(val) || npy_isinf(val)) {
      tc->type = JT_NULL;
    } else {
      pc->doubleValue = val;
      tc->type = JT_DOUBLE;
    }
    return;
  } else if (PyBytes_Check(obj)) {
    pc->PyTypeToUTF8 = PyBytesToUTF8;
    tc->type = JT_UTF8;
    return;
  } else if (PyUnicode_Check(obj)) {
    pc->PyTypeToUTF8 = PyUnicodeToUTF8;
    tc->type = JT_UTF8;
    return;
  } else if (object_is_decimal_type(obj)) {
    PyObject *is_nan_py = PyObject_RichCompare(obj, obj, Py_NE);
    if (is_nan_py == NULL) {
      goto INVALID;
    }
    int is_nan = (is_nan_py == Py_True);
    Py_DECREF(is_nan_py);
    if (is_nan) {
      tc->type = JT_NULL;
      return;
    }
    pc->PyTypeToUTF8 = PyDecimalToUTF8Callback;
    tc->type = JT_UTF8;
    return;
  } else if (PyDateTime_Check(obj) || PyDate_Check(obj)) {
    if (object_is_nat_type(obj)) {
      tc->type = JT_NULL;
      return;
    }

    if (enc->datetimeIso) {
      pc->PyTypeToUTF8 = PyDateTimeToIsoCallback;
      tc->type = JT_UTF8;
    } else {
      NPY_DATETIMEUNIT base = ((PyObjectEncoder *)tc->encoder)->datetimeUnit;
      pc->longValue = PyDateTimeToEpoch(obj, base);
      tc->type = JT_LONG;
    }
    return;
  } else if (PyTime_Check(obj)) {
    pc->PyTypeToUTF8 = PyTimeToJSON;
    tc->type = JT_UTF8;
    return;
  } else if (PyArray_IsScalar(obj, Datetime)) {
    npy_int64 longVal;
    if (((PyDatetimeScalarObject *)obj)->obval == get_nat()) {
      tc->type = JT_NULL;
      return;
    }
    PyArray_Descr *dtype = PyArray_DescrFromScalar(obj);
    if (!PyTypeNum_ISDATETIME(dtype->type_num)) {
      PyErr_Format(PyExc_ValueError, "Could not get resolution of datetime");
      return;
    }

    PyArray_Descr *outcode = PyArray_DescrFromType(NPY_INT64);
    PyArray_CastScalarToCtype(obj, &longVal, outcode);
    Py_DECREF(outcode);

    if (enc->datetimeIso) {
      GET_TC(tc)->longValue = longVal;
      pc->PyTypeToUTF8 = NpyDateTimeToIsoCallback;
      enc->valueUnit = get_datetime_metadata_from_dtype(dtype).base;
      tc->type = JT_UTF8;
    } else {
      NPY_DATETIMEUNIT base = ((PyObjectEncoder *)tc->encoder)->datetimeUnit;
      pc->longValue = PyDateTimeToEpoch(obj, base);
      tc->type = JT_LONG;
    }
    return;
  } else if (PyDelta_Check(obj)) {
    // pd.Timedelta object or pd.NaT should evaluate true here
    // fallback to nanoseconds per sec for other objects
    // TODO(anyone): cast below loses precision if total_seconds return
    // value exceeds number of bits that significand can hold
    // also liable to overflow
    int64_t value = PyObject_HasAttrString(obj, "_value")
                        ? get_long_attr(obj, "_value")
                        : (int64_t)(total_seconds(obj) * 1000000000LL);

    if (value == get_nat()) {
      tc->type = JT_NULL;
      return;
    } else if (enc->datetimeIso) {
      pc->PyTypeToUTF8 = NpyTimeDeltaToIsoCallback;
      enc->valueUnit = NPY_FR_ns;
      tc->type = JT_UTF8;
    } else {
      const int unit = ((PyObjectEncoder *)tc->encoder)->datetimeUnit;
      if (scaleNanosecToUnit(&value, unit) != 0) {
        // TODO(username): Add some kind of error handling here
      }

      if (PyErr_Occurred() && PyErr_ExceptionMatches(PyExc_OverflowError)) {
        goto INVALID;
      }

      tc->type = JT_LONG;
    }
    pc->longValue = value;
    return;
  } else if (PyArray_IsScalar(obj, Integer)) {
    tc->type = JT_LONG;
    PyArray_CastScalarToCtype(obj, &(pc->longValue),
                              PyArray_DescrFromType(NPY_INT64));

    if (PyErr_Occurred() && PyErr_ExceptionMatches(PyExc_OverflowError)) {
      goto INVALID;
    }

    return;
  } else if (PyArray_IsScalar(obj, Bool)) {
    PyArray_CastScalarToCtype(obj, &(pc->longValue),
                              PyArray_DescrFromType(NPY_BOOL));
    tc->type = (pc->longValue) ? JT_TRUE : JT_FALSE;
    return;
  } else if (PyArray_IsScalar(obj, Float) || PyArray_IsScalar(obj, Double)) {
    PyArray_CastScalarToCtype(obj, &(pc->doubleValue),
                              PyArray_DescrFromType(NPY_DOUBLE));
    tc->type = JT_DOUBLE;
    return;
  } else if (PyArray_CheckScalar(obj)) {
    PyErr_Format(PyExc_TypeError,
                 "%R (numpy-scalar) is not JSON serializable at the moment",
                 obj);
    goto INVALID;
  } else if (object_is_na_type(obj)) {
    tc->type = JT_NULL;
    return;
  }

ISITERABLE:

  if (object_is_index_type(obj)) {
    if (enc->outputFormat == SPLIT) {
      tc->type = JT_OBJECT;
      pc->iterBegin = Index_iterBegin;
      pc->iterEnd = Index_iterEnd;
      pc->iterNext = Index_iterNext;
      pc->iterGetValue = Index_iterGetValue;
      pc->iterGetName = Index_iterGetName;
      return;
    }

    pc->newObj = get_values(obj);
    if (pc->newObj) {
      tc->type = JT_ARRAY;
      pc->iterBegin = NpyArr_iterBegin;
      pc->iterEnd = NpyArr_iterEnd;
      pc->iterNext = NpyArr_iterNext;
      pc->iterGetValue = NpyArr_iterGetValue;
      pc->iterGetName = NpyArr_iterGetName;
    } else {
      goto INVALID;
    }

    return;
  } else if (object_is_series_type(obj)) {
    if (enc->outputFormat == SPLIT) {
      tc->type = JT_OBJECT;
      pc->iterBegin = Series_iterBegin;
      pc->iterEnd = Series_iterEnd;
      pc->iterNext = Series_iterNext;
      pc->iterGetValue = Series_iterGetValue;
      pc->iterGetName = Series_iterGetName;
      return;
    }

    pc->newObj = get_values(obj);
    if (!pc->newObj) {
      goto INVALID;
    }

    if (enc->outputFormat == INDEX || enc->outputFormat == COLUMNS) {
      tc->type = JT_OBJECT;
      PyObject *tmpObj = PyObject_GetAttrString(obj, "index");
      if (!tmpObj) {
        goto INVALID;
      }
      PyObject *values = get_values(tmpObj);
      Py_DECREF(tmpObj);
      if (!values) {
        goto INVALID;
      }

      if (!PyArray_Check(pc->newObj)) {
        PyErr_SetString(PyExc_TypeError,
                        "Object_beginTypeContext received a non-array object");
        goto INVALID;
      }
      const PyArrayObject *arrayobj = (const PyArrayObject *)pc->newObj;
      pc->columnLabelsLen = PyArray_DIM(arrayobj, 0);
      pc->columnLabels = NpyArr_encodeLabels((PyArrayObject *)values, enc,
                                             pc->columnLabelsLen);
      if (!pc->columnLabels) {
        goto INVALID;
      }
    } else {
      tc->type = JT_ARRAY;
    }
    pc->iterBegin = NpyArr_iterBegin;
    pc->iterEnd = NpyArr_iterEnd;
    pc->iterNext = NpyArr_iterNext;
    pc->iterGetValue = NpyArr_iterGetValue;
    pc->iterGetName = NpyArr_iterGetName;
    return;
  } else if (PyArray_Check(obj)) {
    if (enc->npyCtxtPassthru) {
      pc->npyarr = enc->npyCtxtPassthru;
      tc->type = (pc->npyarr->columnLabels ? JT_OBJECT : JT_ARRAY);

      pc->iterBegin = NpyArrPassThru_iterBegin;
      pc->iterNext = NpyArr_iterNext;
      pc->iterEnd = NpyArrPassThru_iterEnd;
      pc->iterGetValue = NpyArr_iterGetValue;
      pc->iterGetName = NpyArr_iterGetName;

      enc->npyCtxtPassthru = NULL;
      return;
    }

    tc->type = JT_ARRAY;
    pc->iterBegin = NpyArr_iterBegin;
    pc->iterEnd = NpyArr_iterEnd;
    pc->iterNext = NpyArr_iterNext;
    pc->iterGetValue = NpyArr_iterGetValue;
    pc->iterGetName = NpyArr_iterGetName;
    return;
  } else if (object_is_dataframe_type(obj)) {
    if (enc->blkCtxtPassthru) {
      pc->pdblock = enc->blkCtxtPassthru;
      tc->type =
          (pc->pdblock->npyCtxts[0]->columnLabels ? JT_OBJECT : JT_ARRAY);

      pc->iterBegin = PdBlockPassThru_iterBegin;
      pc->iterEnd = PdBlockPassThru_iterEnd;
      pc->iterNext = PdBlock_iterNextItem;
      pc->iterGetName = PdBlock_iterGetName;
      pc->iterGetValue = NpyArr_iterGetValue;

      enc->blkCtxtPassthru = NULL;
      return;
    }

    if (enc->outputFormat == SPLIT) {
      tc->type = JT_OBJECT;
      pc->iterBegin = DataFrame_iterBegin;
      pc->iterEnd = DataFrame_iterEnd;
      pc->iterNext = DataFrame_iterNext;
      pc->iterGetValue = DataFrame_iterGetValue;
      pc->iterGetName = DataFrame_iterGetName;
      return;
    }

    pc->iterBegin = PdBlock_iterBegin;
    pc->iterEnd = PdBlock_iterEnd;
    pc->iterNext = PdBlock_iterNext;
    pc->iterGetName = PdBlock_iterGetName;
    pc->iterGetValue = NpyArr_iterGetValue;

    if (enc->outputFormat == VALUES) {
      tc->type = JT_ARRAY;
    } else if (enc->outputFormat == RECORDS) {
      tc->type = JT_ARRAY;
      PyObject *tmpObj = PyObject_GetAttrString(obj, "columns");
      if (!tmpObj) {
        goto INVALID;
      }
      PyObject *values = get_values(tmpObj);
      if (!values) {
        Py_DECREF(tmpObj);
        goto INVALID;
      }
      pc->columnLabelsLen = PyObject_Size(tmpObj);
      pc->columnLabels = NpyArr_encodeLabels((PyArrayObject *)values, enc,
                                             pc->columnLabelsLen);
      Py_DECREF(tmpObj);
      if (!pc->columnLabels) {
        goto INVALID;
      }
    } else if (enc->outputFormat == INDEX || enc->outputFormat == COLUMNS) {
      tc->type = JT_OBJECT;
      PyObject *tmpObj =
          (enc->outputFormat == INDEX ? PyObject_GetAttrString(obj, "index")
                                      : PyObject_GetAttrString(obj, "columns"));
      if (!tmpObj) {
        goto INVALID;
      }
      PyObject *values = get_values(tmpObj);
      if (!values) {
        Py_DECREF(tmpObj);
        goto INVALID;
      }
      pc->rowLabelsLen = PyObject_Size(tmpObj);
      pc->rowLabels =
          NpyArr_encodeLabels((PyArrayObject *)values, enc, pc->rowLabelsLen);
      Py_DECREF(tmpObj);
      tmpObj =
          (enc->outputFormat == INDEX ? PyObject_GetAttrString(obj, "columns")
                                      : PyObject_GetAttrString(obj, "index"));
      if (!tmpObj) {
        NpyArr_freeLabels(pc->rowLabels, pc->rowLabelsLen);
        pc->rowLabels = NULL;
        goto INVALID;
      }
      values = get_values(tmpObj);
      if (!values) {
        Py_DECREF(tmpObj);
        NpyArr_freeLabels(pc->rowLabels, pc->rowLabelsLen);
        pc->rowLabels = NULL;
        goto INVALID;
      }
      pc->columnLabelsLen = PyObject_Size(tmpObj);
      pc->columnLabels = NpyArr_encodeLabels((PyArrayObject *)values, enc,
                                             pc->columnLabelsLen);
      Py_DECREF(tmpObj);
      if (!pc->columnLabels) {
        NpyArr_freeLabels(pc->rowLabels, pc->rowLabelsLen);
        pc->rowLabels = NULL;
        goto INVALID;
      }

      if (enc->outputFormat == COLUMNS) {
        pc->transpose = 1;
      }
    } else {
      goto INVALID;
    }
    return;
  } else if (PyDict_Check(obj)) {
    tc->type = JT_OBJECT;
    pc->iterBegin = Dict_iterBegin;
    pc->iterEnd = Dict_iterEnd;
    pc->iterNext = Dict_iterNext;
    pc->iterGetValue = Dict_iterGetValue;
    pc->iterGetName = Dict_iterGetName;
    pc->dictObj = obj;
    Py_INCREF(obj);

    return;
  } else if (PyList_Check(obj)) {
    tc->type = JT_ARRAY;
    pc->iterBegin = List_iterBegin;
    pc->iterEnd = List_iterEnd;
    pc->iterNext = List_iterNext;
    pc->iterGetValue = List_iterGetValue;
    pc->iterGetName = List_iterGetName;
    return;
  } else if (PyTuple_Check(obj)) {
    tc->type = JT_ARRAY;
    pc->iterBegin = Tuple_iterBegin;
    pc->iterEnd = Tuple_iterEnd;
    pc->iterNext = Tuple_iterNext;
    pc->iterGetValue = Tuple_iterGetValue;
    pc->iterGetName = Tuple_iterGetName;
    return;
  } else if (PyAnySet_Check(obj)) {
    tc->type = JT_ARRAY;
    pc->iterBegin = Set_iterBegin;
    pc->iterEnd = Set_iterEnd;
    pc->iterNext = Set_iterNext;
    pc->iterGetValue = Set_iterGetValue;
    pc->iterGetName = Set_iterGetName;
    return;
  }

  PyObject *toDictFunc = PyObject_GetAttrString(obj, "toDict");

  if (toDictFunc) {
    PyObject *tuple = PyTuple_New(0);
    PyObject *toDictResult = PyObject_Call(toDictFunc, tuple, NULL);
    Py_DECREF(tuple);
    Py_DECREF(toDictFunc);

    if (toDictResult == NULL) {
      PyErr_Clear();
      tc->type = JT_NULL;
      return;
    }

    if (!PyDict_Check(toDictResult)) {
      Py_DECREF(toDictResult);
      tc->type = JT_NULL;
      return;
    }

    tc->type = JT_OBJECT;
    pc->iterBegin = Dict_iterBegin;
    pc->iterEnd = Dict_iterEnd;
    pc->iterNext = Dict_iterNext;
    pc->iterGetValue = Dict_iterGetValue;
    pc->iterGetName = Dict_iterGetName;
    pc->dictObj = toDictResult;
    return;
  }

  PyErr_Clear();

  if (enc->defaultHandler) {
    Object_invokeDefaultHandler(obj, enc);
    goto INVALID;
  }

  tc->type = JT_OBJECT;
  pc->iterBegin = Dir_iterBegin;
  pc->iterEnd = Dir_iterEnd;
  pc->iterNext = Dir_iterNext;
  pc->iterGetValue = Dir_iterGetValue;
  pc->iterGetName = Dir_iterGetName;
  return;

INVALID:
  tc->type = JT_INVALID;
  PyObject_Free(tc->prv);
  tc->prv = NULL;
  return;
}

static void Object_endTypeContext(JSOBJ Py_UNUSED(obj), JSONTypeContext *tc) {
  if (tc->prv) {
    Py_XDECREF(GET_TC(tc)->newObj);
    GET_TC(tc)->newObj = NULL;
    NpyArr_freeLabels(GET_TC(tc)->rowLabels, GET_TC(tc)->rowLabelsLen);
    GET_TC(tc)->rowLabels = NULL;
    NpyArr_freeLabels(GET_TC(tc)->columnLabels, GET_TC(tc)->columnLabelsLen);
    GET_TC(tc)->columnLabels = NULL;
    PyObject_Free(GET_TC(tc)->cStr);
    GET_TC(tc)->cStr = NULL;
    PyObject_Free(tc->prv);
    tc->prv = NULL;
  }
}

static const char *Object_getStringValue(JSOBJ obj, JSONTypeContext *tc,
                                         size_t *_outLen) {
  return GET_TC(tc)->PyTypeToUTF8(obj, tc, _outLen);
}

static JSINT64 Object_getLongValue(JSOBJ Py_UNUSED(obj), JSONTypeContext *tc) {
  return GET_TC(tc)->longValue;
}

static double Object_getDoubleValue(JSOBJ Py_UNUSED(obj), JSONTypeContext *tc) {
  return GET_TC(tc)->doubleValue;
}

static const char *Object_getBigNumStringValue(JSOBJ obj, JSONTypeContext *tc,
                                               size_t *_outLen) {
  PyObject *repr = PyObject_Str(obj);
  const char *str = PyUnicode_AsUTF8AndSize(repr, (Py_ssize_t *)_outLen);
  char *bytes = PyObject_Malloc(*_outLen + 1);
  memcpy(bytes, str, *_outLen + 1);
  GET_TC(tc)->cStr = bytes;

  Py_DECREF(repr);

  return GET_TC(tc)->cStr;
}

static void Object_releaseObject(JSOBJ _obj) { Py_DECREF((PyObject *)_obj); }

static void Object_iterBegin(JSOBJ obj, JSONTypeContext *tc) {
  GET_TC(tc)->iterBegin(obj, tc);
}

static int Object_iterNext(JSOBJ obj, JSONTypeContext *tc) {
  return GET_TC(tc)->iterNext(obj, tc);
}

static void Object_iterEnd(JSOBJ obj, JSONTypeContext *tc) {
  GET_TC(tc)->iterEnd(obj, tc);
}

static JSOBJ Object_iterGetValue(JSOBJ obj, JSONTypeContext *tc) {
  return GET_TC(tc)->iterGetValue(obj, tc);
}

static const char *Object_iterGetName(JSOBJ obj, JSONTypeContext *tc,
                                      size_t *outLen) {
  return GET_TC(tc)->iterGetName(obj, tc, outLen);
}

PyObject *objToJSON(PyObject *Py_UNUSED(self), PyObject *args,
                    PyObject *kwargs) {
  PyDateTime_IMPORT;
  if (PyDateTimeAPI == NULL) {
    return NULL;
  }

  PandasDateTime_IMPORT;
  if (PandasDateTimeAPI == NULL) {
    return NULL;
  }

  static char *kwlist[] = {"obj",
                           "ensure_ascii",
                           "double_precision",
                           "encode_html_chars",
                           "orient",
                           "date_unit",
                           "iso_dates",
                           "default_handler",
                           "indent",
                           NULL};

  PyObject *oinput = NULL;
  PyObject *oensureAscii = NULL;
  int idoublePrecision = 10; // default double precision setting
  PyObject *oencodeHTMLChars = NULL;
  char *sOrient = NULL;
  char *sdateFormat = NULL;
  PyObject *oisoDates = 0;
  PyObject *odefHandler = 0;
  int indent = 0;

  PyObjectEncoder pyEncoder = {
      {
          .beginTypeContext = Object_beginTypeContext,
          .endTypeContext = Object_endTypeContext,
          .getStringValue = Object_getStringValue,
          .getLongValue = Object_getLongValue,
          .getIntValue = NULL,
          .getDoubleValue = Object_getDoubleValue,
          .getBigNumStringValue = Object_getBigNumStringValue,
          .iterBegin = Object_iterBegin,
          .iterNext = Object_iterNext,
          .iterEnd = Object_iterEnd,
          .iterGetValue = Object_iterGetValue,
          .iterGetName = Object_iterGetName,
          .releaseObject = Object_releaseObject,
          .malloc = PyObject_Malloc,
          .realloc = PyObject_Realloc,
          .free = PyObject_Free,
          .recursionMax = -1,
          .doublePrecision = idoublePrecision,
          .forceASCII = 1,
          .encodeHTMLChars = 0,
          .indent = indent,
          .errorMsg = NULL,
      },
      .npyCtxtPassthru = NULL,
      .blkCtxtPassthru = NULL,
      .npyType = -1,
      .npyValue = NULL,
      .datetimeIso = 0,
      .datetimeUnit = NPY_FR_ms,
      .outputFormat = COLUMNS,
      .defaultHandler = NULL,
  };
  JSONObjectEncoder *encoder = (JSONObjectEncoder *)&pyEncoder;

  if (!PyArg_ParseTupleAndKeywords(args, kwargs, "O|OiOssOOi", kwlist, &oinput,
                                   &oensureAscii, &idoublePrecision,
                                   &oencodeHTMLChars, &sOrient, &sdateFormat,
                                   &oisoDates, &odefHandler, &indent)) {
    return NULL;
  }

  if (oensureAscii != NULL && !PyObject_IsTrue(oensureAscii)) {
    encoder->forceASCII = 0;
  }

  if (oencodeHTMLChars != NULL && PyObject_IsTrue(oencodeHTMLChars)) {
    encoder->encodeHTMLChars = 1;
  }

  if (idoublePrecision > JSON_DOUBLE_MAX_DECIMALS || idoublePrecision < 0) {
    PyErr_Format(
        PyExc_ValueError,
        "Invalid value '%d' for option 'double_precision', max is '%u'",
        idoublePrecision, JSON_DOUBLE_MAX_DECIMALS);
    return NULL;
  }
  encoder->doublePrecision = idoublePrecision;

  if (sOrient != NULL) {
    if (strcmp(sOrient, "records") == 0) {
      pyEncoder.outputFormat = RECORDS;
    } else if (strcmp(sOrient, "index") == 0) {
      pyEncoder.outputFormat = INDEX;
    } else if (strcmp(sOrient, "split") == 0) {
      pyEncoder.outputFormat = SPLIT;
    } else if (strcmp(sOrient, "values") == 0) {
      pyEncoder.outputFormat = VALUES;
    } else if (strcmp(sOrient, "columns") != 0) {
      PyErr_Format(PyExc_ValueError, "Invalid value '%s' for option 'orient'",
                   sOrient);
      return NULL;
    }
  }

  if (sdateFormat != NULL) {
    if (strcmp(sdateFormat, "s") == 0) {
      pyEncoder.datetimeUnit = NPY_FR_s;
    } else if (strcmp(sdateFormat, "ms") == 0) {
      pyEncoder.datetimeUnit = NPY_FR_ms;
    } else if (strcmp(sdateFormat, "us") == 0) {
      pyEncoder.datetimeUnit = NPY_FR_us;
    } else if (strcmp(sdateFormat, "ns") == 0) {
      pyEncoder.datetimeUnit = NPY_FR_ns;
    } else {
      PyErr_Format(PyExc_ValueError,
                   "Invalid value '%s' for option 'date_unit'", sdateFormat);
      return NULL;
    }
  }

  if (oisoDates != NULL && PyObject_IsTrue(oisoDates)) {
    pyEncoder.datetimeIso = 1;
  }

  if (odefHandler != NULL && odefHandler != Py_None) {
    if (!PyCallable_Check(odefHandler)) {
      PyErr_SetString(PyExc_TypeError, "Default handler is not callable");
      return NULL;
    }
    pyEncoder.defaultHandler = odefHandler;
  }

  encoder->indent = indent;

  pyEncoder.originalOutputFormat = pyEncoder.outputFormat;

  char buffer[65536];
  char *ret = JSON_EncodeObject(oinput, encoder, buffer, sizeof(buffer));
  if (PyErr_Occurred()) {
    return NULL;
  }

  if (encoder->errorMsg) {
    if (ret != buffer) {
      encoder->free(ret);
    }
    PyErr_Format(PyExc_OverflowError, "%s", encoder->errorMsg);
    return NULL;
  }

  PyObject *newobj = PyUnicode_FromString(ret);

  if (ret != buffer) {
    encoder->free(ret);
  }

  return newobj;
}