Summary: 101 instances, 87 unique Text Count # TODO use _column_copy, but for now this works... 1 # TODO: Perhaps this should be part of dispatch backend registration 1 # TODO should add reduce here as well... 1 # TODO Add type translation 1 # TODO: Support native Velox execution for boolean ops 1 # # TODO fix nullability 1 # TODO: Support SizeProxy 1 # TODO: round(-2.5) returns -2.0 in Numpy/PyTorch but returns -3.0 in Velox 1 # TODO: using a dict for field type lookup 1 // FIXME This is fragile as it assumes velox::TypeKind numbers numeric types 1 # TODO: we need a type annotation for Columns with statically accessible dtype 1 # FIXME: https://github.com/facebookresearch/torcharrow/issues/62 to_arrow doesn't work as expected for map 1 // TODO: Move the Koksi binding of Velox type to OSS 5 # TODO: more efficient/straightfowrad interop 2 // TODO: deprecate this method and migrate to OperatorHandle::fromGenericUDF 1 # TODO: infer the type from the whole list 1 # TODO: should we propagate python list if both keys and vals are lists of strings? 1 # TODO only flat columns supported... 1 # TODO implement 1 // TODO: Model binary functions as UDF. 1 # TODO: use pc.max once upgrade to later version of PyArrow 1 # TODO: automtically populate it 1 # TODO Make GroupedDatFrame also an IGroupedDataframe to make it truly compositional 1 # TODO check that all non-masked items are strings 1 # TODO: should we just move _py_arithmetic_op to INumericColumn since it only works for boolean/numeric types 1 TODO: rename this method to _EmptyColumnBuilder to emphasize it returns a builder 1 # TODO: docstring 1 # TODO: figure out zero copy from Velox vector 1 # TODO check that all non-masked items are legal numbers (i.e not nan) 1 // TODO: Use Column once ConstantColumn is supported 1 # TODO: this actually puts the type annotations on the tuple wrong. 1 # TODO Check why teh following assert isn't the case 1 packages=find_packages() # TODO ??? 1 # TODO: implement this 1 # TODO fix me 1 # TODO if not, adopt the type or? 1 # TODO Add type translation. 1 // TODO: move this method as static... 1 # TODO should we hijack this (or have yet another doc decorator) 1 # TODO Make this abstract and add all the abstract methods here ... 1 # TODO refine this test 1 # TODO handle duplicate columns, if ever... 1 # TODO: this implementation is very inefficient, we should wrap the column directly instead of round-tripping through python 5 # TODO: rename the dispatch key to be "_from_pylist" 1 # TODO extend that this works inside struct frames as well, 1 cols = data.values() # TODO: also allow data to be a single Velox RowColumn 1 # TODO this must be weakened (to deal with nulls, etc)... 1 # TODO: Fix this! 1 # TODO: decide on nulls, here we assume all defined (mask = False) for new parent... 2 # TODO: More efficient interop, such as leveraging arrow 1 # TODO Use error handling to mke this more efficient.. 1 // TODO: return SimpleColumn instead? 1 // TODO: use row type instead of string 1 # TODO improve perf by looking at lists instead of first building a map 1 # TODO: if func is annotated, check whether its input parameter is IColumn when format="column" 1 // TODO: we may only need to register UDFs that TorchArrow required? 1 # TODO: check that fields of named tuples match? 1 # FIXME: https://github.com/facebookresearch/torcharrow/issues/62 1 # TODO: calculate without materializing all the occurrences 1 # TODO: Deprecate _finalized since Velox Column doesn't have "Builder" mode 1 // TODO: Decide on how we want to handle column-scalar type promotion. 1 # TODO Add native kernel support 1 // TODO: Open source some part of utility codes in Koski (PyVelox?) 1 # TODO Decide on tracing level: If we trace 'min' om a 1 # TODO: perhaps this should be a private method 1 // TODO: refactor into some type utility class 1 # TODO: use pc.min once upgrade to later version of PyArrow 1 assert mask is None # TODO: remove mask parameter in _FullColumn 1 # TODO decide on whether we nat to have arbitrarily nested wheres... 1 // TODO: remove these method binding and migrate to the 1 // TODO: Model binary functions as UDF. 1 // TODO: Move uses of static variables into .cpp. Static variables are local to 1 "TypeKind", // TODO: Move the Koksi binding of Velox type to OSS 1 # TODO: Rename this class to IColumnVelox or IColumnCpu 1 # TODO Add functionality for first and last 1 # TODO: this might cause name duplicates, do disambiguation 1 # TODO: refactor these static methods out of scope.py 1 # TODO: support non-primitive types 1 # TODO: is there any better approach? 1 # TODO: pad_sequence also works for nest numeric list 1 # TODO fix slow implementation by vectorization, 1 # TODO Current short cut has 'everything', excpet for columns as a DataFrameDemo 1 # # DESIGN BUG: TODO needs actually both sides for symmetric promotion rules ... 1 // TODO: support type promotion in TorchArrow-Velox backend so registering less 1 # TODO: check type annotations of inputs too in order to infer the input format 2 # TODO: clarify int32 vs int64 1 // TODO: This causes an extra type-based dispatch. 4