def lower_box_df(context, builder, sig, args):
assert len(sig.args) % 2 == 0, "name and column pairs expected"
n_cols = len(sig.args)//2
col_names = [a.literal_value for a in sig.args[:n_cols]]
col_arrs = [a for a in args[n_cols:]]
arr_typs = [a for a in sig.args[n_cols:]]
pyapi = context.get_python_api(builder)
env_manager = context.get_env_manager(builder)
c = numba.pythonapi._BoxContext(context, builder, pyapi, env_manager)
gil_state = pyapi.gil_ensure() # acquire GIL
mod_name = context.insert_const_string(c.builder.module, "pandas")
class_obj = pyapi.import_module_noblock(mod_name)
res = pyapi.call_method(class_obj, "DataFrame", ())
for cname, arr, arr_typ in zip(col_names, col_arrs, arr_typs):
# df['cname'] = boxed_arr
# TODO: datetime.date, DatetimeIndex?
if arr_typ == string_array_type:
arr_obj = box_str_arr(arr_typ, arr, c)
else:
arr_obj = box_array(arr_typ, arr, c)
# TODO: is incref required?
context.nrt.incref(builder, arr_typ, arr)
name_str = context.insert_const_string(c.builder.module, cname)
cname_obj = pyapi.string_from_string(name_str)
pyapi.object_setitem(res, cname_obj, arr_obj)
# pyapi.decref(arr_obj)
pyapi.decref(cname_obj)
pyapi.decref(class_obj)
pyapi.gil_release(gil_state) # release GIL
return res
def _box_series_data(dtype, data_typ, val, c):
if isinstance(dtype, types.BaseTuple):
np_dtype = np.dtype(
','.join(str(t) for t in dtype.types), align=True)
dtype = numba.numpy_support.from_dtype(np_dtype)
if dtype == string_type:
arr = box_str_arr(string_array_type, val, c)
elif dtype == datetime_date_type:
arr = box_datetime_date_array(data_typ, val, c)
elif isinstance(dtype, PDCategoricalDtype):
arr = box_categorical_array(data_typ, val, c)
elif data_typ == string_array_split_view_type:
arr = box_str_arr_split_view(data_typ, val, c)
elif dtype == types.List(string_type):
arr = box_list(list_string_array_type, val, c)
else:
arr = box_array(data_typ, val, c)
if isinstance(dtype, types.Record):
o_str = c.context.insert_const_string(c.builder.module, "O")
o_str = c.pyapi.string_from_string(o_str)
arr = c.pyapi.call_method(arr, "astype", (o_str,))
return arr
def box_categorical_array(typ, val, c):
dtype = typ.dtype
mod_name = c.context.insert_const_string(c.builder.module, "pandas")
pd_class_obj = c.pyapi.import_module_noblock(mod_name)
# categories list e.g. ['A', 'B', 'C']
item_objs = _get_cat_obj_items(dtype.categories, c)
n = len(item_objs)
list_obj = c.pyapi.list_new(c.context.get_constant(types.intp, n))
for i in range(n):
idx = c.context.get_constant(types.intp, i)
c.pyapi.incref(item_objs[i])
c.pyapi.list_setitem(list_obj, idx, item_objs[i])
int_dtype = get_categories_int_type(dtype)
arr = box_array(types.Array(int_dtype, 1, 'C'), val, c)
pdcat_cls_obj = c.pyapi.object_getattr_string(pd_class_obj, "Categorical")
cat_arr = c.pyapi.call_method(pdcat_cls_obj, "from_codes", (arr, list_obj))
c.pyapi.decref(pdcat_cls_obj)
c.pyapi.decref(arr)
c.pyapi.decref(list_obj)
for obj in item_objs:
c.pyapi.decref(obj)
c.pyapi.decref(pd_class_obj)
return cat_arr
def box_dataframe(typ, val, c):
context = c.context
builder = c.builder
n_cols = len(typ.columns)
col_names = typ.columns
arr_typs = typ.data
dtypes = [a.dtype for a in arr_typs] # TODO: check Categorical
dataframe = cgutils.create_struct_proxy(typ)(context, builder, value=val)
col_arrs = [
builder.extract_value(dataframe.data, i) for i in range(n_cols)
]
# df unboxed from Python
has_parent = cgutils.is_not_null(builder, dataframe.parent)
pyapi = c.pyapi
# gil_state = pyapi.gil_ensure() # acquire GIL
mod_name = context.insert_const_string(c.builder.module, "pandas")
class_obj = pyapi.import_module_noblock(mod_name)
df_obj = pyapi.call_method(class_obj, "DataFrame", ())
for i, cname, arr, arr_typ, dtype in zip(range(n_cols), col_names,
col_arrs, arr_typs, dtypes):
# df['cname'] = boxed_arr
# TODO: datetime.date, DatetimeIndex?
name_str = context.insert_const_string(c.builder.module, cname)
cname_obj = pyapi.string_from_string(name_str)
if dtype == string_type:
arr_obj = box_str_arr(arr_typ, arr, c)
elif isinstance(dtype, PDCategoricalDtype):
arr_obj = box_categorical_array(arr_typ, arr, c)
# context.nrt.incref(builder, arr_typ, arr)
elif arr_typ == string_array_split_view_type:
arr_obj = box_str_arr_split_view(arr_typ, arr, c)
elif dtype == types.List(string_type):
arr_obj = box_list(list_string_array_type, arr, c)
# context.nrt.incref(builder, arr_typ, arr) # TODO required?
# pyapi.print_object(arr_obj)
else:
arr_obj = box_array(arr_typ, arr, c)
# TODO: is incref required?
# context.nrt.incref(builder, arr_typ, arr)
pyapi.object_setitem(df_obj, cname_obj, arr_obj)
# pyapi.decref(arr_obj)
pyapi.decref(cname_obj)
# set df.index if necessary
if typ.index != types.none:
arr_obj = _box_series_data(typ.index.dtype, typ.index, dataframe.index,
c)
pyapi.object_setattr_string(df_obj, 'index', arr_obj)
pyapi.decref(class_obj)
# pyapi.gil_release(gil_state) # release GIL
return df_obj
def box_datetime_date_array(typ, val, c):
ary = box_array(types.Array(types.int64, 1, 'C'), val, c)
hpat_name = c.context.insert_const_string(c.builder.module, 'hpat')
hpat_mod = c.pyapi.import_module_noblock(hpat_name)
hi_mod = c.pyapi.object_getattr_string(hpat_mod, 'hiframes')
pte_mod = c.pyapi.object_getattr_string(hi_mod, 'pd_timestamp_ext')
iatdd = c.pyapi.object_getattr_string(pte_mod, 'int_array_to_datetime_date')
res = c.pyapi.call_function_objargs(iatdd, [ary])
return res
def box_series(typ, val, c):
"""
"""
if typ.dtype == string_type:
arr = box_str_arr(typ, val, c)
else:
arr = box_array(types.Array(typ.dtype, 1, 'C'), val, c)
mod_name = c.context.insert_const_string(c.builder.module, "pandas")
class_obj = c.pyapi.import_module_noblock(mod_name)
res = c.pyapi.call_method(class_obj, "Series", (arr, ))
# class_obj = c.pyapi.unserialize(c.pyapi.serialize_object(pd.Series))
# res = c.pyapi.call_function_objargs(class_obj, (arr,))
c.pyapi.decref(class_obj)
return res
def box_dt_index(typ, val, c):
"""
"""
mod_name = c.context.insert_const_string(c.builder.module, "pandas")
pd_class_obj = c.pyapi.import_module_noblock(mod_name)
dt_index = numba.cgutils.create_struct_proxy(
typ)(c.context, c.builder, val)
arr = box_array(_dt_index_data_typ, dt_index.data, c)
# TODO: support name boxing
# if typ.is_named:
# name = c.pyapi.from_native_value(string_type, series.name)
# else:
# name = c.pyapi.make_none()
res = c.pyapi.call_method(pd_class_obj, "DatetimeIndex", (arr,))
c.pyapi.decref(pd_class_obj)
return res
