numba package¶
Subpackages¶
- numba.cuda package
- Subpackages
- numba.cuda.cudadrv package
- Submodules
- numba.cuda.cudadrv.devicearray module
- numba.cuda.cudadrv.devices module
- numba.cuda.cudadrv.driver module
- numba.cuda.cudadrv.drvapi module
- numba.cuda.cudadrv.enums module
- numba.cuda.cudadrv.error module
- numba.cuda.cudadrv.libs module
- numba.cuda.cudadrv.ndarray module
- numba.cuda.cudadrv.nvvm module
- Module contents
- numba.cuda.cudadrv package
- Submodules
- numba.cuda.api module
- numba.cuda.compiler module
- numba.cuda.cudadecl module
- numba.cuda.cudaimpl module
- numba.cuda.cudamath module
- numba.cuda.decorators module
- numba.cuda.descriptor module
- numba.cuda.dispatcher module
- numba.cuda.errors module
- numba.cuda.initialize module
- numba.cuda.libdevice module
- numba.cuda.nvvmutils module
- numba.cuda.stubs module
- numba.cuda.target module
- numba.cuda.testing module
- Module contents
- Subpackages
- numba.npyufunc package
- numba.pycc package
- numba.servicelib package
- numba.targets package
- Submodules
- numba.targets.arrayobj module
- numba.targets.base module
- numba.targets.builtins module
- numba.targets.cmathimpl module
- numba.targets.cpu module
- numba.targets.descriptors module
- numba.targets.imputils module
- numba.targets.intrinsics module
- numba.targets.iterators module
- numba.targets.mathimpl module
- numba.targets.npdatetime module
- numba.targets.npyfuncs module
- numba.targets.npyimpl module
- numba.targets.operatorimpl module
- numba.targets.optional module
- numba.targets.options module
- numba.targets.printimpl module
- numba.targets.rangeobj module
- numba.targets.registry module
- numba.targets.ufunc_db module
- Module contents
- numba.typeconv package
- numba.typing package
- Submodules
- numba.typing.builtins module
- numba.typing.cffi_utils module
- numba.typing.cmathdecl module
- numba.typing.context module
- numba.typing.ctypes_utils module
- numba.typing.mathdecl module
- numba.typing.npdatetime module
- numba.typing.npydecl module
- numba.typing.operatordecl module
- numba.typing.templates module
- Module contents
Submodules¶
numba.assume module¶
A place to store all assumptions made in various part of the code base.
This allow us to do a usage analysis to discover all code that is assuming something.
Each assumption is defined as a global variable. Its value is the description of the assumption. Code that makes the assumption should assert the_assumption
numba.bytecode module¶
From NumbaPro
- class numba.bytecode.ByteCode(func)¶
Bases: numba.bytecode.ByteCodeBase
- class numba.bytecode.ByteCodeBase(func, func_qualname, argspec, filename, co_names, co_varnames, co_consts, co_freevars, table, labels)¶
Bases: object
- argspec¶
- co_consts¶
- co_freevars¶
- co_names¶
- co_varnames¶
- dump()¶
- filename¶
- func¶
- func_name¶
- func_qualname¶
- labels¶
- table¶
- class numba.bytecode.ByteCodeInst(offset, opcode, arg)¶
Bases: object
- offset:
byte offset of opcode
- opcode:
opcode integer value
- arg:
instruction arg
- lineno:
-1 means unknown
- arg¶
- block_effect¶
Effect of the block stack Returns +1 (push), 0 (none) or -1 (pop)
- classmethod get(offset, opname, arg)¶
- get_jump_target()¶
- is_jump¶
- is_terminator¶
- lineno¶
- next¶
- offset¶
- opcode¶
- opname¶
- class numba.bytecode.ByteCodeOperation(inst, args)¶
Bases: object
- exception numba.bytecode.ByteCodeSupportError¶
Bases: exceptions.Exception
- class numba.bytecode.CustomByteCode(func, func_qualname, argspec, filename, co_names, co_varnames, co_consts, co_freevars, table, labels)¶
Bases: numba.bytecode.ByteCodeBase
A simplified ByteCode class, used for hosting inner loops when loop-lifting.
- numba.bytecode.get_code_object(obj)¶
Shamelessly borrowed from llpython
- numba.bytecode.get_function_object(obj)¶
Objects that wraps function should provide a “__numba__” magic attribute that contains a name of an attribute that contains the actual python function object.
numba.callwrapper module¶
- class numba.callwrapper.PyCallWrapper(context, module, func, fndesc, exceptions)¶
Bases: object
- build()¶
- build_wrapper(api, builder, closure, args, kws)¶
- make_const_string(string)¶
- make_exception_switch(api, builder, code)¶
Handle user defined exceptions. Build a switch to check which exception class was raised.
- make_keywords(kws)¶
numba.cffi_support module¶
Alias to numba.typing.cffi_utils for backward compatibility
numba.cgutils module¶
Generic helpers for LLVM code generation.
- class numba.cgutils.IfBranchObj(builder, bbenter, bbend)¶
Bases: object
- class numba.cgutils.Structure(context, builder, value=None, ref=None, cast_ref=False)¶
Bases: object
A high-level object wrapping a alloca’ed LLVM structure, including named fields and attribute access.
- class numba.cgutils.VerboseProxy(obj)¶
Bases: object
Use to wrap llvm.core.Builder to track where segfault happens
- numba.cgutils.add_postfix(name, postfix)¶
Add postfix to string. If the postfix is already there, add a counter.
- numba.cgutils.alloca_once(builder, ty, size=None, name='')¶
Allocate stack memory at the entry block of the current function pointed by builder withe llvm type ty. The optional size arg set the number of element to allocate. The default is 1. The optional name arg set the symbol name inside the llvm IR for debugging.
- numba.cgutils.append_basic_block(builder, name='')¶
- numba.cgutils.as_bool_bit(builder, value)¶
- numba.cgutils.as_bool_byte(builder, value)¶
- numba.cgutils.cbranch_or_continue(builder, cond, bbtrue)¶
Branch conditionally or continue.
- Note: a new block is created and builder is moved to the end of the new
- block.
- numba.cgutils.divmod_by_constant(builder, val, divisor)¶
Compute the (quotient, remainder) of val divided by the constant positive divisor. The semantics reflects those of Python integer floor division, rather than C’s / LLVM’s signed division and modulo. The difference lies with a negative val.
- numba.cgutils.for_range(*args, **kwds)¶
- numba.cgutils.for_range_slice(*args, **kwds)¶
Generate LLVM IR for a for-loop based on a slice
- builder : object
- Builder object
- start : int
- The beginning value of the slice
- stop : int
- The end value of the slice
- step : int
- The step value of the slice
- intp :
- The data type
- inc : boolean, optional
- A flag to handle the step < 0 case, in which case we decrement the loop
None
- numba.cgutils.gep(builder, ptr, *inds)¶
- numba.cgutils.get_function(builder)¶
- numba.cgutils.get_item_pointer(builder, aryty, ary, inds, wraparound=False)¶
- numba.cgutils.get_item_pointer2(builder, data, shape, strides, layout, inds, wraparound=False)¶
- numba.cgutils.get_module(builder)¶
- numba.cgutils.get_null_value(ltype)¶
- numba.cgutils.get_range_from_slice(builder, slicestruct)¶
- numba.cgutils.get_record_data(builder, record)¶
- numba.cgutils.get_record_member(builder, record, offset, typ)¶
- numba.cgutils.get_strides_from_slice(builder, ndim, strides, slice, ax)¶
- numba.cgutils.global_constant(builder_or_module, name, value, linkage=LINKAGE_INTERNAL(8))¶
Get or create a (LLVM module-)global constant with name or value.
- numba.cgutils.goto_block(*args, **kwds)¶
A context manager which temporarily positions builder at the end of basic block bb (but before any terminator).
- numba.cgutils.goto_entry_block(*args, **kwds)¶
- numba.cgutils.guard_null(context, builder, value)¶
- numba.cgutils.guard_zero(context, builder, value)¶
- numba.cgutils.if_likely(*args, **kwds)¶
- numba.cgutils.if_unlikely(*args, **kwds)¶
- numba.cgutils.ifelse(*args, **kwds)¶
- numba.cgutils.ifnot(*args, **kwds)¶
- numba.cgutils.ifthen(*args, **kwds)¶
- numba.cgutils.inbound_gep(builder, ptr, *inds)¶
- numba.cgutils.init_record_by_ptr(builder, ltyp, ptr)¶
- numba.cgutils.is_false(builder, value)¶
Return a predicate representing whether value is equal to zero.
- numba.cgutils.is_neg_int(builder, val)¶
- numba.cgutils.is_not_null(builder, val)¶
- numba.cgutils.is_not_scalar_zero(builder, value)¶
Return a predicate representin whether a value is not equal to zero. not exactly “not is_scalar_zero” because of nans
- numba.cgutils.is_null(builder, val)¶
- numba.cgutils.is_pointer(ltyp)¶
Whether the LLVM type typ is a struct type.
- numba.cgutils.is_scalar_neg(builder, value)¶
is _value_ negative?. Assumes _value_ is signed
- numba.cgutils.is_scalar_zero(builder, value)¶
Return a predicate representing whether value is equal to zero.
- numba.cgutils.is_scalar_zero_or_nan(builder, value)¶
Return a predicate representing whether value is equal to either zero or NaN.
- numba.cgutils.is_struct(ltyp)¶
Whether the LLVM type typ is a pointer type.
- numba.cgutils.is_struct_ptr(ltyp)¶
Whether the LLVM type typ is a pointer-to-struct type.
- numba.cgutils.is_true(builder, value)¶
Return a predicate representin whether a value is not equal to zero. not exactly “not is_scalar_zero” because of nans
- numba.cgutils.loop_nest(*args, **kwds)¶
- numba.cgutils.make_anonymous_struct(builder, values)¶
Create an anonymous struct constant containing the given LLVM values.
- numba.cgutils.normalize_slice(builder, slice, length)¶
Clip stop
- numba.cgutils.pack_array(builder, values)¶
- numba.cgutils.pointer_add(builder, ptr, offset, return_type=None)¶
Add an integral offset to pointer ptr, and return a pointer of return_type (or, if omitted, the same type as ptr).
Note the computation is done in bytes, and ignores the width of the pointed item type.
- numba.cgutils.printf(builder, format_string, *values)¶
- numba.cgutils.set_branch_weight(builder, brinst, trueweight, falseweight)¶
- numba.cgutils.set_record_data(builder, record, buf)¶
- numba.cgutils.terminate(builder, bbend)¶
- numba.cgutils.unpack_tuple(builder, tup, count)¶
numba.compiler module¶
- class numba.compiler.CompileResult¶
Bases: tuple
CompileResult(typing_context, target_context, entry_point, typing_error, type_annotation, llvm_module, llvm_func, signature, objectmode, lifted, fndesc, interpmode)
- entry_point¶
Alias for field number 2
- fndesc¶
Alias for field number 10
- interpmode¶
Alias for field number 11
- lifted¶
Alias for field number 9
- llvm_func¶
Alias for field number 6
- llvm_module¶
Alias for field number 5
- objectmode¶
Alias for field number 8
- signature¶
Alias for field number 7
- target_context¶
Alias for field number 1
- type_annotation¶
Alias for field number 4
- typing_context¶
Alias for field number 0
- typing_error¶
Alias for field number 3
- exception numba.compiler.CompilerError¶
Bases: exceptions.Exception
- class numba.compiler.Flags¶
Bases: numba.utils.ConfigOptions
- OPTIONS = frozenset(['no_wraparound', 'force_pyobject', 'no_compile', 'enable_pyobject_looplift', 'boundcheck', 'enable_looplift', 'enable_pyobject'])¶
- class numba.compiler.FunctionAttributes¶
Bases: tuple
FunctionAttributes(name, filename, lineno)
- filename¶
Alias for field number 1
- lineno¶
Alias for field number 2
- name¶
Alias for field number 0
- class numba.compiler.Pipeline(typingctx, targetctx, args, return_type, flags, locals)¶
Bases: object
Stores and manages states for the compiler pipeline
- backend_nopython_mode()¶
Native mode compilation
- backend_object_mode()¶
Object mode compilation
- compile_bytecode(bc, lifted=(), func_attr=FunctionAttributes(name='<anonymous>', filename='<unknown>', lineno=0))¶
- compile_extra(func)¶
- compile_internal(bc, func_attr=FunctionAttributes(name='<anonymous>', filename='<unknown>', lineno=0))¶
- extract_bytecode(func)¶
Extract bytecode from function
- fallback_context(*args, **kwds)¶
Wraps code that would signal a fallback to object mode
- frontend_looplift()¶
Loop lifting analysis and transformation
- giveup_context(*args, **kwds)¶
Wraps code that would signal a fallback to interpreter mode
- stage_analyze_bytecode()¶
Analyze bytecode and translating to Numba IR
- stage_annotate_type()¶
Create type annotation after type inference
- stage_compile_interp_mode()¶
Just create a compile result for interpreter mode
- stage_nopython_backend()¶
Do lowering for nopython
- stage_nopython_frontend()¶
Type inference and legalization
- stage_objectmode_backend()¶
Lowering for object mode
- stage_objectmode_frontend()¶
Front-end: Analyze bytecode, generate Numba IR, infer types
- numba.compiler.compile_bytecode(typingctx, targetctx, bc, args, return_type, flags, locals, lifted=(), func_attr=FunctionAttributes(name='<anonymous>', filename='<unknown>', lineno=0))¶
- numba.compiler.compile_extra(typingctx, targetctx, func, args, return_type, flags, locals)¶
- return_type
Use None to indicate
- numba.compiler.compile_internal(typingctx, targetctx, func, args, return_type, locals={})¶
- numba.compiler.compile_isolated(func, args, return_type=None, flags=Flags(), locals={})¶
Compile the function is an isolated environment. Good for testing.
- numba.compiler.compile_result(**kws)¶
- numba.compiler.get_function_attributes(func)¶
Extract the function attributes from a Python function or object with py_func attribute, such as CPUOverloaded.
Returns an instance of FunctionAttributes.
- numba.compiler.ir_optimize_for_py_stage(interp)¶
This passes breaks semantic for the type inferer but they reduces refct calls for object mode.
- numba.compiler.legalize_given_types(args, return_type)¶
- numba.compiler.legalize_return_type(return_type, interp, targetctx)¶
Only accept array return type iff it is passed into the function. Reject function object return types if in nopython mode.
- numba.compiler.native_lowering_stage(targetctx, interp, typemap, restype, calltypes, nocompile)¶
- numba.compiler.py_lowering_stage(targetctx, interp, nocompile)¶
- numba.compiler.translate_stage(bytecode)¶
- numba.compiler.type_inference_stage(typingctx, interp, args, return_type, locals={})¶
numba.controlflow module¶
- class numba.controlflow.CFBlock(offset)¶
Bases: object
- class numba.controlflow.CFGraph¶
Bases: object
Generic (almost) implementation of a Control Flow Graph.
- add_edge(src, dest, data=None)¶
Add an edge from node src to node dest, with optional per-edge data. If such an edge already exists, it is replaced (duplicate edges are not possible).
- add_node(node)¶
Add node to the graph. This is necessary before adding any edges from/to the node. node can be any hashable object.
- backbone()¶
Return the set of nodes constituting the graph’s backbone. (i.e. the nodes that every path starting from the entry point
must go through). By construction, it is non-empty: it contains at least the entry point.
- dead_nodes()¶
Return the set of dead nodes (eliminated from the graph).
- descendents(node)¶
Return the set of descendents of the given node, in topological order (ignoring back edges).
- dominators()¶
Return a dictionary of {node -> set(nodes)} mapping each node to the nodes dominating it.
A node D dominates a node N when any path leading to N must go through D.
- dump(file=None)¶
Dump extensive debug information.
- exit_points()¶
Return the computed set of exit nodes (may be empty).
- in_loops(node)¶
Return the list of Loop objects the node belongs to, from innermost to outermost.
- loops()¶
Return a dictionary of {node -> loop} mapping each loop header to the loop (a Loop instance) starting with it.
- nodes()¶
Return the set of live nodes.
- post_dominators()¶
Return a dictionary of {node -> set(nodes)} mapping each node to the nodes post-dominating it.
A node P post-dominates a node N when any path starting from N must go through P.
- predecessors(dest)¶
Yield (node, data) pairs representing the predecessors of node dest. (data will be None if no data was specified when adding the edge)
- process()¶
Compute various properties of the control flow graph. The graph must have been fully populated, and its entry point specified.
- set_entry_point(node)¶
Set the entry point of the graph to node.
- successors(src)¶
Yield (node, data) pairs representing the successors of node src. (data will be None if no data was specified when adding the edge)
- topo_order()¶
Return the sequence of nodes in topological order (ignoring back edges).
- topo_sort(nodes, reverse=False)¶
Iterate over the nodes in topological order (ignoring back edges). The sort isn’t guaranteed to be stable.
- class numba.controlflow.ControlFlowAnalysis(bytecode)¶
Bases: object
bytecode
blocks
blockseq
- doms: dict of set
Dominators
- backbone: set of block offsets
The set of block that is common to all possible code path.
- dump(file=None)¶
- incoming_blocks(block)¶
Yield (incoming block, number of stack pops) pairs for block.
- iterblocks()¶
Return all blocks in sequence of occurrence
- iterliveblocks()¶
Return all live blocks in sequence of occurrence
- jump(target, pops=0)¶
Register a jump (conditional or not) to target offset. pops is the number of stack pops implied by the jump (default 0).
- op_BREAK_LOOP(inst)¶
- op_FOR_ITER(inst)¶
- op_JUMP_ABSOLUTE(inst)¶
- op_JUMP_FORWARD(inst)¶
- op_JUMP_IF_FALSE(inst)¶
- op_JUMP_IF_FALSE_OR_POP(inst)¶
- op_JUMP_IF_TRUE(inst)¶
- op_JUMP_IF_TRUE_OR_POP(inst)¶
- op_POP_BLOCK(inst)¶
- op_POP_JUMP_IF_FALSE(inst)¶
- op_POP_JUMP_IF_TRUE(inst)¶
- op_RAISE_VARARGS(inst)¶
- op_RETURN_VALUE(inst)¶
- op_SETUP_LOOP(inst)¶
- run()¶
- class numba.controlflow.Loop¶
Bases: numba.controlflow.Loop
A control flow loop, as detected by a CFGraph object.
numba.ctypes_support module¶
This file fixes portability issues for ctypes
numba.ctypes_utils module¶
Alias to numba.typing.ctypes_utils for backward compatibility
numba.dataflow module¶
- class numba.dataflow.BlockInfo(offset, incoming_blocks)¶
Bases: object
- append(inst, **kws)¶
- dump()¶
- make_incoming()¶
Create an incoming variable (due to not enough values being available on our stack) and request its assignment from our incoming blocks’ own stacks.
- make_temp(prefix='')¶
- pop(discard=False)¶
Pop a variable from the stack, or request it from incoming blocks if the stack is empty. If discard is true, the variable isn’t meant to be used anymore, which allows reducing the number of temporaries created.
- push(val)¶
- request_outgoing(outgoing_block, phiname, stack_index)¶
Request the assignment of the next available stack variable for block outgoing_block with target name phiname.
- terminator¶
- tos¶
- class numba.dataflow.DataFlowAnalysis(cfa)¶
Bases: object
Perform stack2reg
This is necessary to resolve blocks that propagates stack value. This would allow the use of and and or and python2.6 jumps.
- add_syntax_block(info, block)¶
Add an inner syntax block.
- dispatch(info, inst)¶
- dump()¶
- dup_topx(info, inst, count)¶
- op_BINARY_ADD(info, inst)¶
- op_BINARY_AND(info, inst)¶
- op_BINARY_DIVIDE(info, inst)¶
- op_BINARY_FLOOR_DIVIDE(info, inst)¶
- op_BINARY_LSHIFT(info, inst)¶
- op_BINARY_MODULO(info, inst)¶
- op_BINARY_MULTIPLY(info, inst)¶
- op_BINARY_OR(info, inst)¶
- op_BINARY_POWER(info, inst)¶
- op_BINARY_RSHIFT(info, inst)¶
- op_BINARY_SUBSCR(info, inst)¶
- op_BINARY_SUBTRACT(info, inst)¶
- op_BINARY_TRUE_DIVIDE(info, inst)¶
- op_BINARY_XOR(info, inst)¶
- op_BREAK_LOOP(info, inst)¶
- op_BUILD_LIST(info, inst)¶
- op_BUILD_MAP(info, inst)¶
- op_BUILD_SET(info, inst)¶
- op_BUILD_SLICE(info, inst)¶
slice(TOS1, TOS) or slice(TOS2, TOS1, TOS)
- op_BUILD_TUPLE(info, inst)¶
- op_CALL_FUNCTION(info, inst)¶
- op_COMPARE_OP(info, inst)¶
- op_DELETE_ATTR(info, inst)¶
- op_DUP_TOP(info, inst)¶
- op_DUP_TOPX(info, inst)¶
- op_DUP_TOP_TWO(info, inst)¶
- op_FOR_ITER(info, inst)¶
- op_GET_ITER(info, inst)¶
- op_INPLACE_ADD(info, inst)¶
- op_INPLACE_AND(info, inst)¶
- op_INPLACE_DIVIDE(info, inst)¶
- op_INPLACE_FLOOR_DIVIDE(info, inst)¶
- op_INPLACE_LSHIFT(info, inst)¶
- op_INPLACE_MODULO(info, inst)¶
- op_INPLACE_MULTIPLY(info, inst)¶
- op_INPLACE_OR(info, inst)¶
- op_INPLACE_POWER(info, inst)¶
- op_INPLACE_RSHIFT(info, inst)¶
- op_INPLACE_SUBTRACT(info, inst)¶
- op_INPLACE_TRUE_DIVIDE(info, inst)¶
- op_INPLACE_XOR(info, inst)¶
- op_JUMP_ABSOLUTE(info, inst)¶
- op_JUMP_FORWARD(info, inst)¶
- op_JUMP_IF_FALSE(info, inst)¶
- op_JUMP_IF_FALSE_OR_POP(info, inst)¶
- op_JUMP_IF_TRUE(info, inst)¶
- op_JUMP_IF_TRUE_OR_POP(info, inst)¶
- op_LOAD_ATTR(info, inst)¶
- op_LOAD_CONST(info, inst)¶
- op_LOAD_DEREF(info, inst)¶
- op_LOAD_FAST(info, inst)¶
- op_LOAD_GLOBAL(info, inst)¶
- op_POP_BLOCK(info, inst)¶
- op_POP_JUMP_IF_FALSE(info, inst)¶
- op_POP_JUMP_IF_TRUE(info, inst)¶
- op_POP_TOP(info, inst)¶
- op_PRINT_ITEM(info, inst)¶
- op_PRINT_NEWLINE(info, inst)¶
- op_RAISE_VARARGS(info, inst)¶
- op_RETURN_VALUE(info, inst)¶
- op_ROT_FOUR(info, inst)¶
- op_ROT_THREE(info, inst)¶
- op_ROT_TWO(info, inst)¶
- op_SETUP_LOOP(info, inst)¶
- op_SLICE_0(info, inst)¶
TOS = TOS[:]
- op_SLICE_1(info, inst)¶
TOS = TOS1[TOS:]
- op_SLICE_2(info, inst)¶
TOS = TOS1[:TOS]
- op_SLICE_3(info, inst)¶
TOS = TOS2[TOS1:TOS]
- op_STORE_ATTR(info, inst)¶
- op_STORE_FAST(info, inst)¶
- op_STORE_MAP(info, inst)¶
- op_STORE_SLICE_0(info, inst)¶
TOS[:] = TOS1
- op_STORE_SLICE_1(info, inst)¶
TOS1[TOS:] = TOS2
- op_STORE_SLICE_2(info, inst)¶
TOS1[:TOS] = TOS2
- op_STORE_SLICE_3(info, inst)¶
TOS2[TOS1:TOS] = TOS3
- op_STORE_SUBSCR(info, inst)¶
- op_UNARY_INVERT(info, inst)¶
- op_UNARY_NEGATIVE(info, inst)¶
- op_UNARY_NOT(info, inst)¶
- op_UNARY_POSITIVE(info, inst)¶
- op_UNPACK_SEQUENCE(info, inst)¶
- pop_syntax_block(info)¶
Pop the innermost syntax block and revert its stack effect.
- run()¶
- run_on_block(blk)¶
numba.decorators module¶
Contains function decorators and target_registry
- numba.decorators.autojit(*args, **kws)¶
Deprecated.
Use jit instead. Calls to jit internally.
- numba.decorators.jit([signature_or_function[, locals={}[, target='cpu'[, **targetoptions]]]])¶
This function is used to compile a Python function into native code. It is designed to be used as a decorator for the function to be compiled, but it can also be called as a regular function.
- signature_or_function: function or str
- This argument takes either the function to be compiled, or the signature of the function to be compiled. If this function is used as a decorator, the function to be compiled is the decorated function. In that case, this argument should only be used to optionally specify the function signature. If this function is called like a regular function, and this argument is used to specify the function signature, this function will return another jit function object which can be called again with the function to be compiled as this argument.
argtypes: deprecated
restype: deprecated
- locals: dict
- Mapping of local variable names to Numba types. Used to override the types deduced by Numba’s type inference engine.
- targets: str
- Specifies the target platform to compile for. Valid targets are cpu, gpu, npyufunc, and cuda. Defaults to cpu.
- targetoptions:
- For a cpu target, valid options are:
- nopython: bool
- Set to True to disable the use of PyObjects and Python API calls. The default behavior is to allow the use of PyObjects and Python API. Default value is False.
- forceobj: bool
- Set to True to force the use of PyObjects for every value. Default value is False.
- looplift: bool
- Set to True to enable jitting loops in nopython mode while leaving surrounding code in object mode. This allows functions to allocate NumPy arrays and use Python objects, while the tight loops in the function can still be compiled in nopython mode. Any arrays that the tight loop uses should be created before the loop is entered. Default value is True.
- wraparound: bool
- Set to True to enable array indexing wraparound for negative indices, for a small performance penalty. Default value is True.
compiled function
The function can be used in the following ways:
jit(signature, [target=’cpu’, [**targetoptions]]) -> jit(function)
Equivalent to:
d = dispatcher(function, targetoptions) d.compile(signature)
Create a dispatcher object for a python function and default target-options. Then, compile the funciton with the given signature.
Example:
@jit(“void(int32, float32)”) def foo(x, y):
return x + y
jit(function) -> dispatcher
Same as old autojit. Create a dispatcher function object that specialize at call site.
Example:
@jit def foo(x, y):
return x + y
jit([target=’cpu’, [**targetoptions]]) -> configured_jit(function)
Same as old autojit and 2). But configure with target and default target-options.
Example:
@jit(target=’cpu’, nopython=True) def foo(x, y):
return x + y
- numba.decorators.njit(*args, **kws)¶
Equivalent to jit(nopython=True)
See documentation for jit function/decorator for full description.
numba.dispatcher module¶
- class numba.dispatcher.LiftedLoop(bytecode, typingctx, targetctx, locals, flags)¶
Bases: numba.dispatcher._OverloadedBase
Implementation of the hidden dispatcher objects used for lifted loop (a lifted loop is really compiled as a separate function).
- compile(sig)¶
- class numba.dispatcher.Overloaded(py_func, locals={}, targetoptions={})¶
Bases: numba.dispatcher._OverloadedBase
Implementation of user-facing dispatcher objects (i.e. created using the @jit decorator). This is an abstract base class. Subclasses should define the targetdescr class attribute.
- compile(sig, locals={}, **targetoptions)¶
numba.dummyarray module¶
- class numba.dummyarray.Array(dims, itemsize)¶
Bases: object
A dummy numpy array-like object. Consider it an array without the actual data, but offset from the base data pointer.
- dims: tuple of Dim
- describing each dimension of the array
- ndim: int
- number of dimension
- shape: tuple of int
- size of each dimension
- strides: tuple of int
- stride of each dimension
- itemsize: int
- itemsize
- extent: (start, end)
- start and end offset containing the memory region
- classmethod from_desc(offset, shape, strides, itemsize)¶
- is_array = True¶
- is_c_contig¶
- is_f_contig¶
- iter_contiguous_extent()¶
Generates extents
- ravel(order='C')¶
- reshape(*newshape, **kws)¶
- class numba.dummyarray.Dim(start, stop, size, stride, single)¶
Bases: object
A single dimension of the array
- start:
- start offset
- stop:
- stop offset
- size:
- number of items
- stride:
- item stride
- copy(start=None, stop=None, size=None, stride=None, single=None)¶
- get_offset(idx)¶
- is_contiguous(itemsize)¶
- normalize(base)¶
- single¶
- size¶
- start¶
- stop¶
- stride¶
- class numba.dummyarray.Extent¶
Bases: tuple
Extent(begin, end)
- begin¶
Alias for field number 0
- end¶
Alias for field number 1
- numba.dummyarray.compute_index(indices, dims)¶
- numba.dummyarray.is_element_indexing(item, ndim)¶
- numba.dummyarray.iter_strides_c_contig(arr, shape=None)¶
yields the c-contigous strides
- numba.dummyarray.iter_strides_f_contig(arr, shape=None)¶
yields the f-contigous strides
numba.findlib module¶
- numba.findlib.find_file(pat, libdir=None)¶
- numba.findlib.find_lib(libname, libdir=None, platform=None)¶
- numba.findlib.get_lib_dir()¶
Anaconda specific
numba.interpreter module¶
- class numba.interpreter.Assigner¶
Bases: object
This object keeps track of potential assignment simplifications inside a code block. For example $O.1 = x followed by y = $0.1 can be simplified into y = x, but it’s not possible anymore if we have x = z in-between those two instructions.
NOTE: this is not only an optimization, but is actually necessary due to certain limitations of Numba - such as only accepting the returning of an array passed as function argument.
- assign(srcvar, destvar)¶
Assign srcvar to destvar. Return either srcvar or a possible simplified assignment source (earlier assigned to srcvar).
- get_assignment_source(destname)¶
Get a possible assignment source (a ir.Var instance) to replace destname, otherwise None.
- class numba.interpreter.Interpreter(bytecode)¶
Bases: object
A bytecode interpreter that builds up the IR.
- block_constains_opname(offset, opname)¶
- code_consts¶
- code_freevars¶
- code_locals¶
- code_names¶
- current_scope¶
- dump(file=None)¶
- get(name)¶
Get the variable (a Var instance) with the given name.
- get_closure_value(index)¶
Get a value from the cell contained in this function’s closure.
- get_global_value(name)¶
Get a global value from the func_global (first) or as a builtins (second). If both failed, return a ir.UNDEFINED.
- get_used_globals()¶
Return a dictionary of global variables used by the bytecode.
- init_first_block()¶
- insert_block(offset, scope=None, loc=None)¶
- interpret()¶
- op_BINARY_ADD(inst, lhs, rhs, res)¶
- op_BINARY_AND(inst, lhs, rhs, res)¶
- op_BINARY_DIVIDE(inst, lhs, rhs, res)¶
- op_BINARY_FLOOR_DIVIDE(inst, lhs, rhs, res)¶
- op_BINARY_LSHIFT(inst, lhs, rhs, res)¶
- op_BINARY_MODULO(inst, lhs, rhs, res)¶
- op_BINARY_MULTIPLY(inst, lhs, rhs, res)¶
- op_BINARY_OR(inst, lhs, rhs, res)¶
- op_BINARY_POWER(inst, lhs, rhs, res)¶
- op_BINARY_RSHIFT(inst, lhs, rhs, res)¶
- op_BINARY_SUBSCR(inst, target, index, res)¶
- op_BINARY_SUBTRACT(inst, lhs, rhs, res)¶
- op_BINARY_TRUE_DIVIDE(inst, lhs, rhs, res)¶
- op_BINARY_XOR(inst, lhs, rhs, res)¶
- op_BREAK_LOOP(inst)¶
- op_BUILD_LIST(inst, items, res)¶
- op_BUILD_MAP(inst, size, res)¶
- op_BUILD_SET(inst, items, res)¶
- op_BUILD_SLICE(inst, start, stop, step, res, slicevar)¶
- op_BUILD_TUPLE(inst, items, res)¶
- op_CALL_FUNCTION(inst, func, args, kws, res)¶
- op_COMPARE_OP(inst, lhs, rhs, res)¶
- op_DELETE_ATTR(inst, target)¶
- op_DUP_TOP(inst, orig, duped)¶
- op_DUP_TOPX(inst, orig, duped)¶
- op_DUP_TOP_TWO(inst, orig, duped)¶
- op_FOR_ITER(inst, iterator, pair, indval, pred)¶
Assign new block other this instruction.
- op_GET_ITER(inst, value, res)¶
- op_INPLACE_ADD(inst, lhs, rhs, res)¶
- op_INPLACE_AND(inst, lhs, rhs, res)¶
- op_INPLACE_DIVIDE(inst, lhs, rhs, res)¶
- op_INPLACE_FLOOR_DIVIDE(inst, lhs, rhs, res)¶
- op_INPLACE_LSHIFT(inst, lhs, rhs, res)¶
- op_INPLACE_MODULO(inst, lhs, rhs, res)¶
- op_INPLACE_MULTIPLY(inst, lhs, rhs, res)¶
- op_INPLACE_OR(inst, lhs, rhs, res)¶
- op_INPLACE_POWER(inst, lhs, rhs, res)¶
- op_INPLACE_RSHIFT(inst, lhs, rhs, res)¶
- op_INPLACE_SUBTRACT(inst, lhs, rhs, res)¶
- op_INPLACE_TRUE_DIVIDE(inst, lhs, rhs, res)¶
- op_INPLACE_XOR(inst, lhs, rhs, res)¶
- op_JUMP_ABSOLUTE(inst)¶
- op_JUMP_FORWARD(inst)¶
- op_JUMP_IF_FALSE(inst, pred)¶
- op_JUMP_IF_FALSE_OR_POP(inst, pred)¶
- op_JUMP_IF_TRUE(inst, pred)¶
- op_JUMP_IF_TRUE_OR_POP(inst, pred)¶
- op_LOAD_ATTR(inst, item, res)¶
- op_LOAD_CONST(inst, res)¶
- op_LOAD_DEREF(inst, res)¶
- op_LOAD_FAST(inst, res)¶
- op_LOAD_GLOBAL(inst, res)¶
- op_POP_BLOCK(inst)¶
- op_POP_JUMP_IF_FALSE(inst, pred)¶
- op_POP_JUMP_IF_TRUE(inst, pred)¶
- op_PRINT_ITEM(inst, item, printvar, res)¶
- op_PRINT_NEWLINE(inst, printvar, res)¶
- op_RAISE_VARARGS(inst, exc)¶
- op_RETURN_VALUE(inst, retval, castval)¶
- op_SETUP_LOOP(inst)¶
- op_SLICE_0(inst, base, res, slicevar, indexvar, nonevar)¶
- op_SLICE_1(inst, base, start, nonevar, res, slicevar, indexvar)¶
- op_SLICE_2(inst, base, nonevar, stop, res, slicevar, indexvar)¶
- op_SLICE_3(inst, base, start, stop, res, slicevar, indexvar)¶
- op_STORE_ATTR(inst, target, value)¶
- op_STORE_FAST(inst, value)¶
- op_STORE_MAP(inst, dct, key, value)¶
- op_STORE_SLICE_0(inst, base, value, slicevar, indexvar, nonevar)¶
- op_STORE_SLICE_1(inst, base, start, nonevar, value, slicevar, indexvar)¶
- op_STORE_SLICE_2(inst, base, nonevar, stop, value, slicevar, indexvar)¶
- op_STORE_SLICE_3(inst, base, start, stop, value, slicevar, indexvar)¶
- op_STORE_SUBSCR(inst, target, index, value)¶
- op_UNARY_INVERT(inst, value, res)¶
- op_UNARY_NEGATIVE(inst, value, res)¶
- op_UNARY_NOT(inst, value, res)¶
- op_UNARY_POSITIVE(inst, value, res)¶
- op_UNPACK_SEQUENCE(inst, iterable, stores, tupleobj)¶
- store(value, name, redefine=False)¶
Store value (a Var instance) into the variable named name (a str object).
numba.intrinsics module¶
Numba only functions
- numba.intrinsics.array_ravel(arr)¶
Flatten a C/F array into a 1D array without enforcing the ordering of the each element.
arr: array
A flattened 1D array
numba.io_support module¶
numba.ir module¶
- class numba.ir.Assign(value, target, loc)¶
Bases: numba.ir.Stmt
- class numba.ir.Block(scope, loc)¶
Bases: object
A code block
- append(inst)¶
- dump(file=<open file '<stdout>', mode 'w' at 0x10028e150>)¶
- insert_after(stmt, other)¶
Insert stmt after other.
- insert_before_terminator(stmt)¶
- is_terminated¶
- prepend(inst)¶
- remove(inst)¶
- terminator¶
- verify()¶
- class numba.ir.Branch(cond, truebr, falsebr, loc)¶
Bases: numba.ir.Stmt
- is_terminator = True¶
- class numba.ir.Const(value, loc)¶
Bases: object
- class numba.ir.Del(value, loc)¶
Bases: numba.ir.Stmt
- class numba.ir.DelAttr(target, attr, loc)¶
Bases: numba.ir.Stmt
- class numba.ir.Expr(op, loc, **kws)¶
Bases: numba.ir.Inst
An IR expression (an instruction which can only be part of a larger statement).
- classmethod binop(fn, lhs, rhs, loc)¶
- classmethod build_list(items, loc)¶
- classmethod build_map(size, loc)¶
- classmethod build_set(items, loc)¶
- classmethod build_tuple(items, loc)¶
- classmethod call(func, args, kws, loc)¶
- classmethod cast(value, loc)¶
A node for implicit casting at the return statement
- classmethod exhaust_iter(value, count, loc)¶
- classmethod getattr(value, attr, loc)¶
- classmethod getitem(value, index, loc)¶
- classmethod getiter(value, loc)¶
- classmethod inplace_binop(fn, lhs, rhs, loc)¶
- classmethod iternext(value, loc)¶
- list_vars()¶
- classmethod pair_first(value, loc)¶
- classmethod pair_second(value, loc)¶
- classmethod static_getitem(value, index, loc)¶
- classmethod unary(fn, value, loc)¶
- class numba.ir.FreeVar(index, name, value, loc)¶
Bases: object
A freevar, as loaded by LOAD_DECREF. (i.e. a variable defined in an enclosing non-global scope)
- class numba.ir.Global(name, value, loc)¶
Bases: object
- class numba.ir.Inst¶
Bases: object
Base class for all IR instructions.
- list_vars()¶
List the variables used (read or written) by the instruction.
- class numba.ir.Intrinsic(name, type, args)¶
Bases: object
For inserting intrinsic node into the IR
- class numba.ir.Jump(target, loc)¶
Bases: numba.ir.Stmt
- is_terminator = True¶
- exception numba.ir.NotDefinedError¶
Bases: exceptions.NameError
- class numba.ir.Raise(exception, loc)¶
Bases: numba.ir.Stmt
- is_terminator = True¶
- exception numba.ir.RedefinedError¶
Bases: exceptions.NameError
- class numba.ir.Return(value, loc)¶
Bases: numba.ir.Stmt
- is_terminator = True¶
- class numba.ir.Scope(parent, loc)¶
Bases: object
- parent: Scope
Parent scope
- localvars: VarMap
Scope-local variable map
- loc: Loc
Start of scope location
- define(name, loc)¶
Define a variable
- get(name)¶
Refer to a variable
- get_or_define(name, loc)¶
- has_parent¶
- make_temp(loc)¶
- redefine(name, loc)¶
Redefine if the name is already defined
- class numba.ir.SetAttr(target, attr, value, loc)¶
Bases: numba.ir.Stmt
- class numba.ir.SetItem(target, index, value, loc)¶
Bases: numba.ir.Stmt
- class numba.ir.Stmt¶
Bases: numba.ir.Inst
Base class for IR statements (instructions which can appear on their own in a Block).
- is_terminator = False¶
- list_vars()¶
- class numba.ir.StoreMap(dct, key, value, loc)¶
Bases: numba.ir.Stmt
- class numba.ir.Var(scope, name, loc)¶
Bases: object
scope: Scope
name: str
- loc: Loc
Definition location
- is_temp¶
- exception numba.ir.VerificationError¶
Bases: exceptions.Exception
numba.irpasses module¶
Contains optimization passes for the IR.
numba.looplifting module¶
- class numba.looplifting.SubOffset(val, sub=1)¶
Bases: object
The loop-jitting may insert bytecode between two bytecode but we cannot guarantee that there is enough integral space between two offsets. This class workaround the problem by introducing a fractional part to the offset.
- next()¶
Helper method to get the next suboffset by incrementing the fractional part only
- numba.looplifting.discover_args_and_returns(bytecode, insts, outer_rds, outer_wrs)¶
Basic analysis for args and returns This completely ignores the ordering or the read-writes.
- numba.looplifting.find_previous_inst(insts, offset)¶
- numba.looplifting.find_varnames_uses(bytecode, insts)¶
- numba.looplifting.insert_instruction(insts, item)¶
- numba.looplifting.insert_loop_call(bytecode, loop, args, outer, outerlabels, returns, dispatcher_factory)¶
- numba.looplifting.lift_loop(bytecode, dispatcher_factory)¶
Lift the top-level loops.
- outer: ByteCode of a copy of the loop-less function.
- loops: a list of ByteCode of the loops.
- numba.looplifting.make_loop_bytecode(bytecode, loop, args, returns)¶
- numba.looplifting.separate_loops(bytecode, outer, loops)¶
Separate top-level loops from the function
Stores loopless instructions from the original function into outer. Stores list of loop instructions into loops. Both outer and loops are list-like (append(item) defined).
- numba.looplifting.stitch_instructions(outer, loop)¶
numba.lowering module¶
- class numba.lowering.BaseLower(context, fndesc, interp)¶
Bases: object
Lower IR to LLVM
- init()¶
- init_argument(arg)¶
- lower()¶
- lower_block(block)¶
- post_lower()¶
Called after all blocks are lowered
- pre_lower()¶
Called before lowering all blocks.
- typeof(varname)¶
- class numba.lowering.ExternalFunctionDescriptor(name, restype, argtypes)¶
Bases: numba.lowering.FunctionDescriptor
A FunctionDescriptor subclass for opaque external functions (e.g. raw C functions).
- exception numba.lowering.ForbiddenConstruct(msg, loc)¶
Bases: numba.lowering.LoweringError
- class numba.lowering.FunctionDescriptor(native, modname, qualname, unique_name, doc, typemap, restype, calltypes, args, kws, mangler=None, argtypes=None)¶
Bases: object
- args¶
- argtypes¶
- calltypes¶
- doc¶
- kws¶
- lookup_module()¶
Return the module in which this function is supposed to exist. This may be a dummy module if the function was dynamically generated.
- mangled_name¶
- modname¶
- native¶
- qualname¶
- restype¶
- typemap¶
- unique_name¶
- class numba.lowering.Lower(context, fndesc, interp)¶
Bases: numba.lowering.BaseLower
- alloca(name, type)¶
- alloca_lltype(name, lltype)¶
- getvar(name)¶
- loadvar(name)¶
- lower_assign(ty, inst)¶
- lower_binop(resty, expr)¶
- lower_expr(resty, expr)¶
- lower_inst(inst)¶
- storevar(value, name)¶
- exception numba.lowering.LoweringError(msg, loc)¶
Bases: exceptions.Exception
- class numba.lowering.PythonFunctionDescriptor(native, modname, qualname, unique_name, doc, typemap, restype, calltypes, args, kws, mangler=None, argtypes=None)¶
Bases: numba.lowering.FunctionDescriptor
- classmethod from_object_mode_function(interp)¶
Build a FunctionDescriptor for a Python function to be compiled and executed in object mode.
- classmethod from_specialized_function(interp, typemap, restype, calltypes, mangler)¶
Build a FunctionDescriptor for a specialized Python function.
- numba.lowering.default_mangler(name, argtypes)¶
numba.macro module¶
Macro handling passes
Macros are expanded on block-by-block
- class numba.macro.Macro(name, func, callable=False, argnames=None)¶
Bases: object
A macro object is expanded to a function call
- argnames¶
- callable¶
- func¶
- name¶
- exception numba.macro.MacroError¶
Bases: exceptions.Exception
- numba.macro.expand_macros(blocks)¶
- numba.macro.expand_macros_in_block(constants, block)¶
- numba.macro.module_getattr_folding(constants, block)¶
numba.npdatetime module¶
Helper functions for numpy.timedelta64 and numpy.datetime64. For now, multiples-of-units (for example timedeltas expressed in tens of seconds) are not supported.
- numba.npdatetime.can_cast_timedelta_units(src, dest)¶
- numba.npdatetime.combine_datetime_timedelta_units(datetime_unit, timedelta_unit)¶
Return the unit result of combining datetime_unit with timedelta_unit (e.g. by adding or subtracting). None is returned if combining those units is forbidden.
- numba.npdatetime.get_best_unit(unit_a, unit_b)¶
Get the best (i.e. finer-grained) of two units.
- numba.npdatetime.get_datetime_timedelta_conversion(datetime_unit, timedelta_unit)¶
Compute a possible conversion for combining datetime_unit and timedelta_unit (presumably for adding or subtracting). Return (result unit, integer datetime multiplier, integer timedelta multiplier). RuntimeError is raised if the combination is impossible.
- numba.npdatetime.get_timedelta_conversion_factor(src_unit, dest_unit)¶
Return an integer multiplier allowing to convert from timedeltas of src_unit to dest_unit.
numba.numpy_support module¶
- class numba.numpy_support.UFuncLoopSpec¶
Bases: numba.numpy_support._UFuncLoopSpec
An object describing a ufunc loop’s inner types. Properties: - inputs: the inputs’ Numba types - outputs: the outputs’ Numba types - numpy_inputs: the inputs’ Numpy dtypes - numpy_outputs: the outputs’ Numpy dtypes - ufunc_sig: the string representing the ufunc’s type signature, in
Numpy format (e.g. “ii->i”)- numpy_inputs¶
- numpy_outputs¶
- numba.numpy_support.as_dtype(nbtype)¶
Return a numpy dtype instance corresponding to the given Numba type. NotImplementedError is if no correspondence is known.
- numba.numpy_support.from_dtype(dtype)¶
Return a Numba Type instance corresponding to the given Numpy dtype. NotImplementedError is raised on unsupported Numpy dtypes.
- numba.numpy_support.from_struct_dtype(dtype)¶
- numba.numpy_support.is_array(val)¶
- numba.numpy_support.is_arrayscalar(val)¶
- numba.numpy_support.map_arrayscalar_type(val)¶
- numba.numpy_support.map_layout(val)¶
- numba.numpy_support.supported_ufunc_loop(ufunc, loop)¶
Return whether the loop for the ufunc is supported -in nopython-.
loop should be a UFuncLoopSpec instance, and ufunc a numpy ufunc.
For ufuncs implemented using the ufunc_db, it is supported if the ufunc_db contains a lowering definition for ‘loop’ in the ‘ufunc’ entry.
For other ufuncs, it is type based. The loop will be considered valid if it only contains the following letter types: ‘?bBhHiIlLqQfd’. Note this is legacy and when implementing new ufuncs the ufunc_db should be preferred, as it allows for a more fine-grained incremental support.
- numba.numpy_support.ufunc_find_matching_loop(ufunc, arg_types)¶
Find the appropriate loop to be used for a ufunc based on the types of the operands
ufunc - The ufunc we want to check arg_types - The tuple of arguments to the ufunc, including any
explicit output(s).- return value - A UFuncLoopSpec identifying the loop, or None
- if no matching loop is found.
numba.objmode module¶
Lowering implementation for object mode.
- class numba.objmode.PyLower(context, fndesc, interp)¶
Bases: numba.lowering.BaseLower
- alloca(name, ltype=None)¶
Allocate a stack slot and initialize it to NULL. The default is to allocate a pyobject pointer. Use ltype to override.
- builtin_lookup(mod, name)¶
- mod:
- The __builtins__ dictionary or module, as looked up in a module’s globals.
- name: str
- The object to lookup
- check_error(obj)¶
- check_int_status(num, ok_value=0)¶
Raise an exception if num is smaller than ok_value.
- check_occurred()¶
- cleanup()¶
- decref(value)¶
This is allow to be called on non pyobject pointer, in which case no code is inserted.
- delvar(name)¶
Delete the variable slot with the given name. This will decref the corresponding Python object.
- get_builtin_obj(name)¶
- get_env_const(index)¶
Look up constant number index inside the environment body. A borrowed reference is returned.
- get_module_dict()¶
- incref(value)¶
- init()¶
- init_argument(arg)¶
- is_null(obj)¶
- loadvar(name)¶
Load the llvm value of the variable named name.
- lower_assign(inst)¶
The returned object must have a new reference
- lower_binop(expr, inplace=False)¶
- lower_const(const)¶
- lower_expr(expr)¶
- lower_global(name, value)¶
Check global scope dictionary.
- Check __builtins__.
2a) is it a dictionary (for non __main__ module) 2b) is it a module (for __main__ module)
- lower_inst(inst)¶
- post_lower()¶
- pre_lower()¶
- return_error_occurred()¶
- return_exception_raised()¶
- storevar(value, name)¶
Stores a llvm value and allocate stack slot if necessary. The llvm value can be of arbitrary type.
numba.pythonapi module¶
- exception numba.pythonapi.NativeError¶
Bases: exceptions.RuntimeError
- class numba.pythonapi.PythonAPI(context, builder)¶
Bases: object
Code generation facilities to call into the CPython C API (and related helpers).
- alloca_obj()¶
- bool_from_bool(bval)¶
Get a Python bool from a LLVM boolean.
- bool_from_long(ival)¶
- borrow_none()¶
- bytes_from_string_and_size(string, size)¶
- call(callee, args, kws)¶
- call_function_objargs(callee, objargs)¶
- complex_adaptor(cobj, cmplx)¶
- complex_from_doubles(realval, imagval)¶
- complex_imag_as_double(cobj)¶
- complex_real_as_double(cobj)¶
- create_np_datetime(val, unit_code)¶
- create_np_timedelta(val, unit_code)¶
- decref(obj)¶
- dict_getitem_string(dic, name)¶
Returns a borrowed reference
- dict_new(presize=0)¶
- dict_pack(keyvalues)¶
keyvalues: iterable of (str, llvm.Value of PyObject*)
- dict_setitem(dictobj, nameobj, valobj)¶
- dict_setitem_string(dictobj, name, valobj)¶
- err_clear()¶
- err_occurred()¶
- err_set_object(exctype, excval)¶
- err_set_string(exctype, msg)¶
- extract_np_datetime(obj)¶
- extract_np_timedelta(obj)¶
- extract_record_data(obj, pbuf)¶
- float_as_double(fobj)¶
- float_from_double(fval)¶
- from_native_array(typ, ary)¶
- from_native_return(val, typ)¶
- from_native_value(val, typ)¶
- from_tuple(typ, val)¶
- get_c_object(name)¶
Get a Python object through its C-accessible name. (e.g. “PyExc_ValueError”).
- get_null_object()¶
- gil_ensure()¶
Ensure the GIL is acquired. The returned value must be consumed by gil_release().
- gil_release(gil)¶
Release the acquired GIL by gil_ensure(). Must be pair with a gil_ensure().
- import_module_noblock(modname)¶
- incref(obj)¶
- iter_next(iterobj)¶
- list_getitem(lst, idx)¶
Returns a borrowed reference.
- list_new(szval)¶
- list_pack(items)¶
- list_setitem(seq, idx, val)¶
Warning: Steals reference to val
- long_as_longlong(numobj)¶
- long_as_ulonglong(numobj)¶
- long_from_long(ival)¶
- long_from_longlong(ival)¶
- long_from_ssize_t(ival)¶
- long_from_ulonglong(ival)¶
- make_none()¶
- native_error_type¶
- numba_array_adaptor(ary, ptr)¶
- number_add(lhs, rhs, inplace=False)¶
- number_and(lhs, rhs, inplace=False)¶
- number_as_ssize_t(numobj)¶
- number_divide(lhs, rhs, inplace=False)¶
- number_float(val)¶
- number_floordivide(lhs, rhs, inplace=False)¶
- number_invert(obj)¶
- number_long(numobj)¶
- number_lshift(lhs, rhs, inplace=False)¶
- number_multiply(lhs, rhs, inplace=False)¶
- number_negative(obj)¶
- number_or(lhs, rhs, inplace=False)¶
- number_positive(obj)¶
- number_power(lhs, rhs, inplace=False)¶
- number_remainder(lhs, rhs, inplace=False)¶
- number_rshift(lhs, rhs, inplace=False)¶
- number_subtract(lhs, rhs, inplace=False)¶
- number_truedivide(lhs, rhs, inplace=False)¶
- number_xor(lhs, rhs, inplace=False)¶
- object_delattr_string(obj, attr)¶
- object_dump(obj)¶
Dump a Python object on C stderr. For debugging purposes.
- object_getattr_string(obj, attr)¶
- object_getitem(obj, key)¶
- object_getiter(obj)¶
- object_istrue(obj)¶
- object_not(obj)¶
- object_richcompare(lhs, rhs, opstr)¶
Refer to Python source Include/object.h for macros definition of the opid.
- object_setattr_string(obj, attr, val)¶
- object_setitem(obj, key, val)¶
- object_str(obj)¶
- parse_tuple(args, fmt, *objs)¶
- parse_tuple_and_keywords(args, kws, fmt, keywords, *objs)¶
- print_object(obj)¶
- print_string(text)¶
- raise_exception(exctype, excval)¶
- raise_missing_global_error(name)¶
- raise_missing_name_error(name)¶
- raise_native_error(msg)¶
- recreate_record(pdata, size, dtypeaddr)¶
- release_record_buffer(pbuf)¶
- return_none()¶
- sequence_getslice(obj, start, stop)¶
- sequence_tuple(obj)¶
- set_add(set, value)¶
- set_new(iterable=None)¶
- string_as_string(strobj)¶
- string_from_constant_string(string)¶
- string_from_string_and_size(string, size)¶
- sys_write_stdout(fmt, *args)¶
- to_native_arg(obj, typ)¶
- to_native_array(typ, ary)¶
- to_native_value(obj, typ)¶
- tuple_getitem(tup, idx)¶
Borrow reference
- tuple_new(count)¶
- tuple_pack(items)¶
- tuple_setitem(tuple_val, index, item)¶
Steals a reference to item.
- tuple_size(tup)¶
- numba.pythonapi.fix_python_api()¶
Execute once to install special symbols into the LLVM symbol table
numba.serialize module¶
Serialization support for compiled functions.
numba.sigutils module¶
- numba.sigutils.is_signature(sig)¶
- numba.sigutils.normalize_signature(sig)¶
- numba.sigutils.parse_signature(signature_str)¶
numba.six module¶
Utilities for writing code that runs on Python 2 and 3
- class numba.six.Module_six_moves_urllib¶
Bases: module
Create a six.moves.urllib namespace that resembles the Python 3 namespace
- error = <module 'numba.six.moves.urllib.error' (built-in)>¶
- parse = <module 'numba.six.moves.urllib_parse' (built-in)>¶
- request = <module 'numba.six.moves.urllib.request' (built-in)>¶
- response = <module 'numba.six.moves.urllib.response' (built-in)>¶
- robotparser = <module 'numba.six.moves.urllib.robotparser' (built-in)>¶
- class numba.six.Module_six_moves_urllib_error(name)¶
Bases: numba.six._LazyModule
Lazy loading of moved objects in six.moves.urllib_error
- class numba.six.Module_six_moves_urllib_parse(name)¶
Bases: numba.six._LazyModule
Lazy loading of moved objects in six.moves.urllib_parse
- class numba.six.Module_six_moves_urllib_request(name)¶
Bases: numba.six._LazyModule
Lazy loading of moved objects in six.moves.urllib_request
- class numba.six.Module_six_moves_urllib_response(name)¶
Bases: numba.six._LazyModule
Lazy loading of moved objects in six.moves.urllib_response
- class numba.six.Module_six_moves_urllib_robotparser(name)¶
Bases: numba.six._LazyModule
Lazy loading of moved objects in six.moves.urllib_robotparser
- class numba.six.MovedAttribute(name, old_mod, new_mod, old_attr=None, new_attr=None)¶
Bases: numba.six._LazyDescr
- class numba.six.MovedModule(name, old, new=None)¶
Bases: numba.six._LazyDescr
- numba.six.add_metaclass(metaclass)¶
Class decorator for creating a class with a metaclass.
- numba.six.add_move(move)¶
Add an item to six.moves.
- numba.six.b(s)¶
Byte literal
- numba.six.byte2int(bs)¶
- numba.six.create_bound_method(func, obj)¶
- numba.six.exec_(_code_, _globs_=None, _locs_=None)¶
Execute code in a namespace.
- numba.six.get_unbound_function(unbound)¶
Get the function out of a possibly unbound function
- numba.six.indexbytes(buf, i)¶
- numba.six.iterbytes(buf)¶
- numba.six.iteritems(d, **kw)¶
Return an iterator over the (key, value) pairs of a dictionary.
- numba.six.iterkeys(d, **kw)¶
Return an iterator over the keys of a dictionary.
- numba.six.iterlists(d, **kw)¶
Return an iterator over the (key, [values]) pairs of a dictionary.
- numba.six.itervalues(d, **kw)¶
Return an iterator over the values of a dictionary.
- numba.six.remove_move(name)¶
Remove item from six.moves.
- numba.six.reraise(tp, value, tb=None)¶
Reraise an exception.
- numba.six.u(s)¶
Text literal
- numba.six.with_metaclass(meta, *bases)¶
Create a base class with a metaclass.
- numba.six.wraps(wrapped, assigned=('__module__', '__name__', '__doc__'), updated=('__dict__', ))¶
numba.special module¶
- numba.special.typeof(val)¶
Get the type of a variable or value.
Used outside of Numba code, infers the type for the object.
numba.testing module¶
- numba.testing.allow_interpreter_mode(fn)¶
Temporarily re-enable intepreter mode
- numba.testing.discover_tests(startdir)¶
Discover test under a directory
- numba.testing.run_tests(suite, xmloutput=None, verbosity=1)¶
- suite [TestSuite]
A suite of all tests to run
- xmloutput [str or None]
Path of XML output directory (optional)
- verbosity [int]
Verbosity level of tests output
Returns the TestResult object after running the test suite.
- numba.testing.test(**kwargs)¶
Run all tests under numba.tests.
descriptions
verbosity
buffer
failfast
- xmloutput [str]
Path of XML output directory
numba.type_annotations module¶
numba.typeinfer module¶
Type inference base on CPA. The algorithm guarantees monotonic growth of type-sets for each variable.
- Steps:
- seed initial types
- build constrains
- propagate constrains
- unify types
Constrain propagation is precise and does not regret (no backtracing). Constrains push types forward following the dataflow.
- class numba.typeinfer.BuildTupleConstrain(target, items, loc)¶
Bases: object
- class numba.typeinfer.CallConstrain(target, func, args, kws, loc)¶
Bases: object
Constrain for calling functions. Perform case analysis foreach combinations of argument types.
- resolve(context, typevars, fnty)¶
- class numba.typeinfer.ConstrainNetwork¶
Bases: object
- TODO: It is possible to optimize constrain propagation to consider only
- dirty type variables.
- append(constrain)¶
- propagate(context, typevars)¶
- class numba.typeinfer.ExhaustIterConstrain(target, count, iterator, loc)¶
Bases: object
- class numba.typeinfer.GetAttrConstrain(target, attr, value, loc, inst)¶
Bases: object
- class numba.typeinfer.IntrinsicCallConstrain(target, func, args, kws, loc)¶
- class numba.typeinfer.PairFirstConstrain(target, pair, loc)¶
Bases: object
- class numba.typeinfer.PairSecondConstrain(target, pair, loc)¶
Bases: object
- class numba.typeinfer.Propagate(dst, src, loc)¶
Bases: object
A simple constrain for direct propagation of types for assignments.
- class numba.typeinfer.SetAttrConstrain(target, attr, value, loc)¶
Bases: object
- class numba.typeinfer.SetItemConstrain(target, index, value, loc)¶
Bases: object
- class numba.typeinfer.StaticGetItemConstrain(target, value, index, loc)¶
Bases: object
- class numba.typeinfer.TypeInferer(context, blocks)¶
Bases: object
Operates on block that shares the same ir.Scope.
- build_constrain()¶
- constrain_statement(inst)¶
- dump()¶
- get_function_types(typemap)¶
- get_return_type(typemap)¶
- get_state_token()¶
The algorithm is monotonic. It can only grow the typesets. The sum of all lengths of type sets is a cheap and accurate description of our progress.
- propagate()¶
- resolve_value_type(inst, val)¶
Resolve the type of a simple Python value, such as can be represented by literals.
- seed_return(typ)¶
Seeding of return value is optional.
- seed_type(name, typ)¶
All arguments should be seeded.
- sentry_modified_builtin(inst, gvar)¶
Ensure that builtins are modified.
- typeof_assign(inst)¶
- typeof_call(inst, target, call)¶
- typeof_const(inst, target, const)¶
- typeof_expr(inst, target, expr)¶
- typeof_global(inst, target, gvar)¶
- typeof_intrinsic_call(inst, target, func, *args)¶
- typeof_setattr(inst)¶
- typeof_setitem(inst)¶
- unify()¶
- class numba.typeinfer.TypeVar(context, var)¶
Bases: object
- add_types(*types)¶
- get()¶
- getone()¶
- lock(typ)¶
- union(other)¶
- exception numba.typeinfer.TypingError(msg, loc=None)¶
Bases: exceptions.Exception
numba.types module¶
These type objects do not have a fixed machine representation. It is up to the targets to choose their representation.
- numba.types.optional¶
alias of Optional
numba.unittest_support module¶
This file fixes portability issues for unittest
numba.utils module¶
- class numba.utils.BenchmarkResult(func, records, loop)¶
Bases: object
- class numba.utils.SortedMap(seq)¶
Bases: _abcoll.Mapping
Immutable
- class numba.utils.SortedSet(seq)¶
Bases: _abcoll.Set
- class numba.utils.UniqueDict¶
Bases: dict
- numba.utils.benchmark(func, maxsec=1)¶
- numba.utils.bit_length(intval)¶
Return the number of bits necessary to represent integer intval.
- class numba.utils.cached_property(func, name=None)¶
Bases: object
Decorator that converts a method with a single self argument into a property cached on the instance.
Optional name argument allows you to make cached properties of other methods. (e.g. url = cached_property(get_absolute_url, name=’url’) )
- class numba.utils.finalize(obj, func, *args, **kwargs)¶
Class for finalization of weakrefable objects
finalize(obj, func, *args, **kwargs) returns a callable finalizer object which will be called when obj is garbage collected. The first time the finalizer is called it evaluates func(*arg, **kwargs) and returns the result. After this the finalizer is dead, and calling it just returns None.
When the program exits any remaining finalizers for which the atexit attribute is true will be run in reverse order of creation. By default atexit is true.
- alive¶
Whether finalizer is alive
- atexit¶
Whether finalizer should be called at exit
- detach()¶
If alive then mark as dead and return (obj, func, args, kwargs); otherwise return None
- peek()¶
If alive then return (obj, func, args, kwargs); otherwise return None
- numba.utils.format_time(tm)¶
- numba.utils.runonce(fn)¶
- numba.utils.shutting_down(globals=<built-in function globals>)¶
Whether the interpreter is currently shutting down. For use in finalizers, __del__ methods, and similar; it is advised to early bind this function rather than look it up when calling it, since at shutdown module globals may be cleared.
- numba.utils.total_ordering(cls)¶
Class decorator that fills in missing ordering methods
Module contents¶
Expose top-level symbols that are safe for import *
- numba.jit([signature_or_function[, locals={}[, target='cpu'[, **targetoptions]]]])¶
This function is used to compile a Python function into native code. It is designed to be used as a decorator for the function to be compiled, but it can also be called as a regular function.
- signature_or_function: function or str
- This argument takes either the function to be compiled, or the signature of the function to be compiled. If this function is used as a decorator, the function to be compiled is the decorated function. In that case, this argument should only be used to optionally specify the function signature. If this function is called like a regular function, and this argument is used to specify the function signature, this function will return another jit function object which can be called again with the function to be compiled as this argument.
argtypes: deprecated
restype: deprecated
- locals: dict
- Mapping of local variable names to Numba types. Used to override the types deduced by Numba’s type inference engine.
- targets: str
- Specifies the target platform to compile for. Valid targets are cpu, gpu, npyufunc, and cuda. Defaults to cpu.
- targetoptions:
- For a cpu target, valid options are:
- nopython: bool
- Set to True to disable the use of PyObjects and Python API calls. The default behavior is to allow the use of PyObjects and Python API. Default value is False.
- forceobj: bool
- Set to True to force the use of PyObjects for every value. Default value is False.
- looplift: bool
- Set to True to enable jitting loops in nopython mode while leaving surrounding code in object mode. This allows functions to allocate NumPy arrays and use Python objects, while the tight loops in the function can still be compiled in nopython mode. Any arrays that the tight loop uses should be created before the loop is entered. Default value is True.
- wraparound: bool
- Set to True to enable array indexing wraparound for negative indices, for a small performance penalty. Default value is True.
compiled function
The function can be used in the following ways:
jit(signature, [target=’cpu’, [**targetoptions]]) -> jit(function)
Equivalent to:
d = dispatcher(function, targetoptions) d.compile(signature)
Create a dispatcher object for a python function and default target-options. Then, compile the funciton with the given signature.
Example:
@jit(“void(int32, float32)”) def foo(x, y):
return x + y
jit(function) -> dispatcher
Same as old autojit. Create a dispatcher function object that specialize at call site.
Example:
@jit def foo(x, y):
return x + y
jit([target=’cpu’, [**targetoptions]]) -> configured_jit(function)
Same as old autojit and 2). But configure with target and default target-options.
Example:
@jit(target=’cpu’, nopython=True) def foo(x, y):
return x + y
- numba.autojit(*args, **kws)¶
Deprecated.
Use jit instead. Calls to jit internally.
- numba.njit(*args, **kws)¶
Equivalent to jit(nopython=True)
See documentation for jit function/decorator for full description.
- numba.vectorize(ftylist[, target='cpu'[, **kws]])¶
A decorator to create numpy ufunc object from Numba compiled code.
- ftylist: iterable
- An iterable of type signatures, which are either function type object or a string describing the function type.
- target: str
- A string for code generation target. Default to “cpu”.
A NumPy universal function
- @vectorize([‘float32(float32, float32)’,
- ‘float64(float64, float64)’])
- def sum(a, b):
- return a + b
- numba.guvectorize(ftylist, signature[, target='cpu'[, **kws]])¶
A decorator to create numpy generialized-ufunc object from Numba compiled code.
- ftylist: iterable
- An iterable of type signatures, which are either function type object or a string describing the function type.
- signature: str
- A NumPy generialized-ufunc signature. e.g. “(m, n), (n, p)->(m, p)”
- target: str
- A string for code generation target. Defaults to “cpu”.
A NumPy generialized universal-function
- @guvectorize([‘void(int32[:,:], int32[:,:], int32[:,:])’,
- ‘void(float32[:,:], float32[:,:], float32[:,:])’], ‘(x, y),(x, y)->(x, y)’)
- def add_2d_array(a, b):
- for i in range(c.shape[0]):
- for j in range(c.shape[1]):
- c[i, j] = a[i, j] + b[i, j]
- numba.export(prototype)¶
- numba.exportmany(prototypes)¶
- numba.from_dtype(dtype)¶
Return a Numba Type instance corresponding to the given Numpy dtype. NotImplementedError is raised on unsupported Numpy dtypes.
- numba.optional¶
alias of Optional
- numba.typeof(val)¶
Get the type of a variable or value.
Used outside of Numba code, infers the type for the object.