# Copyright (c) 2020 PaddlePaddle Authors. All Rights Reserved. # # Licensed under the Apache License, Version 2.0 (the "License"); # you may not use this file except in compliance with the License. # You may obtain a copy of the License at # # http://www.apache.org/licenses/LICENSE-2.0 # # Unless required by applicable law or agreed to in writing, software # distributed under the License is distributed on an "AS IS" BASIS, # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. # See the License for the specific language governing permissions and # limitations under the License. # pylint: disable=doc-string-missing import threading import multiprocessing import Queue import os import sys import paddle_serving_server from paddle_serving_client import MultiLangClient as Client from concurrent import futures import numpy as np import grpc from .proto import general_model_config_pb2 as m_config from .proto import general_python_service_pb2 as pyservice_pb2 from .proto import pyserving_channel_pb2 as channel_pb2 from .proto import general_python_service_pb2_grpc import logging import random import time import func_timeout import enum import collections class _TimeProfiler(object): def __init__(self): self._pid = os.getpid() self._print_head = 'PROFILE\tpid:{}\t'.format(self._pid) self._time_record = Queue.Queue() self._enable = False def enable(self, enable): self._enable = enable def record(self, name_with_tag): if self._enable is False: return name_with_tag = name_with_tag.split("_") tag = name_with_tag[-1] name = '_'.join(name_with_tag[:-1]) self._time_record.put((name, tag, int(round(time.time() * 1000000)))) def print_profile(self): if self._enable is False: return sys.stderr.write(self._print_head) tmp = {} while not self._time_record.empty(): name, tag, timestamp = self._time_record.get() if name in tmp: ptag, ptimestamp = tmp.pop(name) sys.stderr.write("{}_{}:{} ".format(name, ptag, ptimestamp)) sys.stderr.write("{}_{}:{} ".format(name, tag, timestamp)) else: tmp[name] = (tag, timestamp) sys.stderr.write('\n') for name, item in tmp.items(): tag, timestamp = item self._time_record.put((name, tag, timestamp)) _profiler = _TimeProfiler() class ChannelDataEcode(enum.Enum): OK = 0 TIMEOUT = 1 class ChannelDataType(enum.Enum): CHANNEL_PBDATA = 0 CHANNEL_FUTURE = 1 class ChannelData(object): def __init__(self, future=None, pbdata=None, data_id=None, callback_func=None): self.future = future if pbdata is None: if data_id is None: raise ValueError("data_id cannot be None") pbdata = channel_pb2.ChannelData() pbdata.type = ChannelDataType.CHANNEL_FUTURE.value pbdata.ecode = ChannelDataEcode.OK.value pbdata.id = data_id self.pbdata = pbdata self.callback_func = callback_func def parse(self): # return narray feed = {} if self.pbdata.type == ChannelDataType.CHANNEL_PBDATA.value: for inst in self.pbdata.insts: feed[inst.name] = np.frombuffer(inst.data, dtype=inst.type) feed[inst.name].shape = np.frombuffer(inst.shape, dtype="int32") elif self.pbdata.type == ChannelDataType.CHANNEL_FUTURE.value: feed = self.future.result() if self.callback_func is not None: feed = self.callback_func(feed) else: raise TypeError( self._log("Error type({}) in pbdata.type.".format( self.pbdata.type))) return feed class Channel(Queue.Queue): """ The channel used for communication between Ops. 1. Support multiple different Op feed data (multiple producer) Different types of data will be packaged through the data ID 2. Support multiple different Op fetch data (multiple consumer) Only when all types of Ops get the data of the same ID, the data will be poped; The Op of the same type will not get the data of the same ID. 3. (TODO) Timeout and BatchSize are not fully supported. Note: 1. The ID of the data in the channel must be different. 2. The function add_producer() and add_consumer() are not thread safe, and can only be called during initialization. """ def __init__(self, name=None, maxsize=-1, timeout=None): Queue.Queue.__init__(self, maxsize=maxsize) self._maxsize = maxsize self._timeout = timeout self.name = name self._stop = False self._cv = threading.Condition() self._producers = [] self._producer_res_count = {} # {data_id: count} self._push_res = {} # {data_id: {op_name: data}} self._consumers = {} # {op_name: idx} self._idx_consumer_num = {} # {idx: num} self._consumer_base_idx = 0 self._front_res = [] def get_producers(self): return self._producers def get_consumers(self): return self._consumers.keys() def _log(self, info_str): return "[{}] {}".format(self.name, info_str) def debug(self): return self._log("p: {}, c: {}".format(self.get_producers(), self.get_consumers())) def add_producer(self, op_name): """ not thread safe, and can only be called during initialization. """ if op_name in self._producers: raise ValueError( self._log("producer({}) is already in channel".format(op_name))) self._producers.append(op_name) def add_consumer(self, op_name): """ not thread safe, and can only be called during initialization. """ if op_name in self._consumers: raise ValueError( self._log("consumer({}) is already in channel".format(op_name))) self._consumers[op_name] = 0 if self._idx_consumer_num.get(0) is None: self._idx_consumer_num[0] = 0 self._idx_consumer_num[0] += 1 def push(self, channeldata, op_name=None): logging.debug( self._log("{} try to push data: {}".format(op_name, channeldata.pbdata))) if len(self._producers) == 0: raise Exception( self._log( "expected number of producers to be greater than 0, but the it is 0." )) elif len(self._producers) == 1: with self._cv: while self._stop is False: try: self.put(channeldata, timeout=0) break except Queue.Empty: self._cv.wait() self._cv.notify_all() logging.debug(self._log("{} push data succ!".format(op_name))) return True elif op_name is None: raise Exception( self._log( "There are multiple producers, so op_name cannot be None.")) producer_num = len(self._producers) data_id = channeldata.pbdata.id put_data = None with self._cv: logging.debug(self._log("{} get lock".format(op_name))) if data_id not in self._push_res: self._push_res[data_id] = { name: None for name in self._producers } self._producer_res_count[data_id] = 0 self._push_res[data_id][op_name] = channeldata if self._producer_res_count[data_id] + 1 == producer_num: put_data = self._push_res[data_id] self._push_res.pop(data_id) self._producer_res_count.pop(data_id) else: self._producer_res_count[data_id] += 1 if put_data is None: logging.debug( self._log("{} push data succ, but not push to queue.". format(op_name))) else: while self._stop is False: try: self.put(put_data, timeout=0) break except Queue.Empty: self._cv.wait() logging.debug( self._log("multi | {} push data succ!".format(op_name))) self._cv.notify_all() return True def front(self, op_name=None): logging.debug(self._log("{} try to get data".format(op_name))) if len(self._consumers) == 0: raise Exception( self._log( "expected number of consumers to be greater than 0, but the it is 0." )) elif len(self._consumers) == 1: resp = None with self._cv: while self._stop is False and resp is None: try: resp = self.get(timeout=0) break except Queue.Empty: self._cv.wait() logging.debug(self._log("{} get data succ!".format(op_name))) return resp elif op_name is None: raise Exception( self._log( "There are multiple consumers, so op_name cannot be None.")) with self._cv: # data_idx = consumer_idx - base_idx while self._stop is False and self._consumers[ op_name] - self._consumer_base_idx >= len(self._front_res): try: channeldata = self.get(timeout=0) self._front_res.append(channeldata) break except Queue.Empty: self._cv.wait() consumer_idx = self._consumers[op_name] base_idx = self._consumer_base_idx data_idx = consumer_idx - base_idx resp = self._front_res[data_idx] logging.debug(self._log("{} get data: {}".format(op_name, resp))) self._idx_consumer_num[consumer_idx] -= 1 if consumer_idx == base_idx and self._idx_consumer_num[ consumer_idx] == 0: self._idx_consumer_num.pop(consumer_idx) self._front_res.pop(0) self._consumer_base_idx += 1 self._consumers[op_name] += 1 new_consumer_idx = self._consumers[op_name] if self._idx_consumer_num.get(new_consumer_idx) is None: self._idx_consumer_num[new_consumer_idx] = 0 self._idx_consumer_num[new_consumer_idx] += 1 self._cv.notify_all() logging.debug(self._log("multi | {} get data succ!".format(op_name))) return resp # reference, read only def stop(self): #TODO self.close() self._stop = True class Op(object): def __init__(self, name, inputs, server_model=None, server_port=None, device=None, client_config=None, server_name=None, fetch_names=None, concurrency=1, timeout=-1, retry=2): self._run = False self.name = name # to identify the type of OP, it must be globally unique self._concurrency = concurrency # amount of concurrency self.set_input_ops(inputs) self.set_client(client_config, server_name, fetch_names) self._server_model = server_model self._server_port = server_port self._device = device self._timeout = timeout self._retry = retry self._input = None self._outputs = [] def set_client(self, client_config, server_name, fetch_names): self._client = None if client_config is None or \ server_name is None or \ fetch_names is None: return self._client = Client() self._client.load_client_config(client_config) self._client.connect([server_name]) self._fetch_names = fetch_names def with_serving(self): return self._client is not None def get_input_channel(self): return self._input def get_input_ops(self): return self._input_ops def set_input_ops(self, ops): if not isinstance(ops, list): ops = [] if ops is None else [ops] self._input_ops = [] for op in ops: if not isinstance(op, Op): raise TypeError( self._log('input op must be Op type, not {}'.format( type(op)))) self._input_ops.append(op) def add_input_channel(self, channel): if not isinstance(channel, Channel): raise TypeError( self._log('input channel must be Channel type, not {}'.format( type(channel)))) channel.add_consumer(self.name) self._input = channel def get_output_channels(self): return self._outputs def add_output_channel(self, channel): if not isinstance(channel, Channel): raise TypeError( self._log('output channel must be Channel type, not {}'.format( type(channel)))) channel.add_producer(self.name) self._outputs.append(channel) def preprocess(self, channeldata): if isinstance(channeldata, dict): raise Exception( self._log( 'this Op has multiple previous inputs. Please override this method' )) feed = channeldata.parse() return feed def midprocess(self, data): if not isinstance(data, dict): raise Exception( self._log( 'data must be dict type(the output of preprocess()), but get {}'. format(type(data)))) logging.debug(self._log('data: {}'.format(data))) logging.debug(self._log('fetch: {}'.format(self._fetch_names))) call_future = self._client.predict( feed=data, fetch=self._fetch_names, asyn=True) logging.debug(self._log("get call_future")) return call_future def postprocess(self, output_data): return output_data def errorprocess(self, error_info, data_id): data = channel_pb2.ChannelData() data.ecode = 1 data.id = data_id data.error_info = error_info return data def stop(self): self._input.stop() for channel in self._outputs: channel.stop() self._run = False def _parse_channeldata(self, channeldata): data_id, error_data = None, None if isinstance(channeldata, dict): parsed_data = {} key = channeldata.keys()[0] data_id = channeldata[key].pbdata.id for _, data in channeldata.items(): if data.pbdata.ecode != 0: error_data = data break else: data_id = channeldata.pbdata.id if channeldata.pbdata.ecode != 0: error_data = channeldata.pbdata return data_id, error_data def start(self, concurrency_idx): self._run = True while self._run: _profiler.record("{}{}-get_0".format(self.name, concurrency_idx)) input_data = self._input.front(self.name) _profiler.record("{}{}-get_1".format(self.name, concurrency_idx)) logging.debug(self._log("input_data: {}".format(input_data))) data_id, error_data = self._parse_channeldata(input_data) output_data = None if error_data is None: _profiler.record("{}{}-prep_0".format(self.name, concurrency_idx)) data = self.preprocess(input_data) _profiler.record("{}{}-prep_1".format(self.name, concurrency_idx)) call_future = None error_info = None if self.with_serving(): _profiler.record("{}{}-midp_0".format(self.name, concurrency_idx)) if self._timeout > 0: for i in range(self._retry): try: call_future = func_timeout.func_timeout( self._timeout, self.midprocess, args=(data, )) except func_timeout.FunctionTimedOut: logging.error("error: timeout") error_info = "{}({}): timeout".format( self.name, concurrency_idx) if i + 1 < self._retry: error_info = None logging.warn( self._log("warn: timeout, retry({})". format(i + 1))) except Exception as e: logging.error("error: {}".format(e)) error_info = "{}({}): {}".format( self.name, concurrency_idx, e) logging.warn(self._log(e)) # TODO break else: break else: call_future = self.midprocess(data) _profiler.record("{}{}-midp_1".format(self.name, concurrency_idx)) _profiler.record("{}{}-postp_0".format(self.name, concurrency_idx)) if error_info is not None: error_data = self.errorprocess(error_info, data_id) output_data = ChannelData(pbdata=error_data) else: if self.with_serving(): # use call_future output_data = ChannelData( future=call_future, data_id=data_id, callback_func=self.postprocess) else: post_data = self.postprocess(data) if not isinstance(post_data, dict): raise TypeError( self._log( 'output_data must be dict type, but get {}'. format(type(output_data)))) pbdata = channel_pb2.ChannelData() for name, value in post_data.items(): inst = channel_pb2.Inst() inst.data = value.tobytes() inst.name = name inst.shape = np.array( value.shape, dtype="int32").tobytes() inst.type = str(value.dtype) pbdata.insts.append(inst) pbdata.ecode = 0 pbdata.id = data_id output_data = ChannelData(pbdata=pbdata) _profiler.record("{}{}-postp_1".format(self.name, concurrency_idx)) else: output_data = ChannelData(pbdata=error_data) _profiler.record("{}{}-push_0".format(self.name, concurrency_idx)) for channel in self._outputs: channel.push(output_data, self.name) _profiler.record("{}{}-push_1".format(self.name, concurrency_idx)) def _log(self, info_str): return "[{}] {}".format(self.name, info_str) def get_concurrency(self): return self._concurrency class GeneralPythonService( general_python_service_pb2_grpc.GeneralPythonService): def __init__(self, in_channel, out_channel, retry=2): super(GeneralPythonService, self).__init__() self.name = "#G" self.set_in_channel(in_channel) self.set_out_channel(out_channel) logging.debug(self._log(in_channel.debug())) logging.debug(self._log(out_channel.debug())) #TODO: # multi-lock for different clients # diffenert lock for server and client self._id_lock = threading.Lock() self._cv = threading.Condition() self._globel_resp_dict = {} self._id_counter = 0 self._retry = retry self._recive_func = threading.Thread( target=GeneralPythonService._recive_out_channel_func, args=(self, )) self._recive_func.start() def _log(self, info_str): return "[{}] {}".format(self.name, info_str) def set_in_channel(self, in_channel): if not isinstance(in_channel, Channel): raise TypeError( self._log('in_channel must be Channel type, but get {}'.format( type(in_channel)))) in_channel.add_producer(self.name) self._in_channel = in_channel def set_out_channel(self, out_channel): if not isinstance(out_channel, Channel): raise TypeError( self._log('out_channel must be Channel type, but get {}'.format( type(out_channel)))) out_channel.add_consumer(self.name) self._out_channel = out_channel def _recive_out_channel_func(self): while True: channeldata = self._out_channel.front(self.name) if not isinstance(channeldata, ChannelData): raise TypeError( self._log('data must be ChannelData type, but get {}'. format(type(channeldata)))) with self._cv: data_id = channeldata.pbdata.id self._globel_resp_dict[data_id] = channeldata self._cv.notify_all() def _get_next_id(self): with self._id_lock: self._id_counter += 1 return self._id_counter - 1 def _get_data_in_globel_resp_dict(self, data_id): resp = None with self._cv: while data_id not in self._globel_resp_dict: self._cv.wait() resp = self._globel_resp_dict.pop(data_id) self._cv.notify_all() return resp def _pack_data_for_infer(self, request): logging.debug(self._log('start inferce')) pbdata = channel_pb2.ChannelData() data_id = self._get_next_id() pbdata.id = data_id for idx, name in enumerate(request.feed_var_names): logging.debug( self._log('name: {}'.format(request.feed_var_names[idx]))) logging.debug(self._log('data: {}'.format(request.feed_insts[idx]))) inst = channel_pb2.Inst() inst.data = request.feed_insts[idx] inst.shape = request.shape[idx] inst.name = name inst.type = request.type[idx] pbdata.insts.append(inst) pbdata.ecode = 0 #TODO: parse request error return ChannelData(pbdata=pbdata), data_id def _pack_data_for_resp(self, channeldata): logging.debug(self._log('get channeldata')) logging.debug(self._log('gen resp')) resp = pyservice_pb2.Response() resp.ecode = channeldata.pbdata.ecode if resp.ecode == 0: if channeldata.pbdata.type == ChannelDataType.CHANNEL_PBDATA.value: for inst in channeldata.pbdata.insts: logging.debug(self._log('append data')) resp.fetch_insts.append(inst.data) logging.debug(self._log('append name')) resp.fetch_var_names.append(inst.name) logging.debug(self._log('append shape')) resp.shape.append(inst.shape) logging.debug(self._log('append type')) resp.type.append(inst.type) elif channeldata.pbdata.type == ChannelDataType.CHANNEL_FUTURE.value: feed = channeldata.futures.result() if channeldata.callback_func is not None: feed = channeldata.callback_func(feed) for name, var in feed: logging.debug(self._log('append data')) resp.fetch_insts.append(var.tobytes()) logging.debug(self._log('append name')) resp.fetch_var_names.append(name) logging.debug(self._log('append shape')) resp.shape.append( np.array( var.shape, dtype="int32").tobytes()) resp.type.append(str(var.dtype)) else: raise TypeError( self._log("Error type({}) in pbdata.type.".format( self.pbdata.type))) else: resp.error_info = channeldata.pbdata.error_info return resp def inference(self, request, context): _profiler.record("{}-prepack_0".format(self.name)) data, data_id = self._pack_data_for_infer(request) _profiler.record("{}-prepack_1".format(self.name)) resp_channeldata = None for i in range(self._retry): logging.debug(self._log('push data')) _profiler.record("{}-push_0".format(self.name)) self._in_channel.push(data, self.name) _profiler.record("{}-push_1".format(self.name)) logging.debug(self._log('wait for infer')) _profiler.record("{}-fetch_0".format(self.name)) resp_channeldata = self._get_data_in_globel_resp_dict(data_id) _profiler.record("{}-fetch_1".format(self.name)) if resp_channeldata.pbdata.ecode == 0: break logging.warn("retry({}): {}".format( i + 1, resp_channeldata.pbdata.error_info)) _profiler.record("{}-postpack_0".format(self.name)) resp = self._pack_data_for_resp(resp_channeldata) _profiler.record("{}-postpack_1".format(self.name)) _profiler.print_profile() return resp class PyServer(object): def __init__(self, retry=2, profile=False): self._channels = [] self._user_ops = [] self._total_ops = [] self._op_threads = [] self._port = None self._worker_num = None self._in_channel = None self._out_channel = None self._retry = retry _profiler.enable(profile) def add_channel(self, channel): self._channels.append(channel) def add_op(self, op): self._user_ops.append(op) def add_ops(self, ops): self._user_ops.extend(ops) def gen_desc(self): logging.info('here will generate desc for PAAS') pass def _topo_sort(self): indeg_num = {} outdegs = {op.name: [] for op in self._user_ops} que_idx = 0 # scroll queue ques = [Queue.Queue() for _ in range(2)] for idx, op in enumerate(self._user_ops): # check the name of op is globally unique if op.name in indeg_num: raise Exception("the name of Op must be unique") indeg_num[op.name] = len(op.get_input_ops()) if indeg_num[op.name] == 0: ques[que_idx].put(op) for pred_op in op.get_input_ops(): outdegs[pred_op.name].append(op) # get dag_views dag_views = [] sorted_op_num = 0 while True: que = ques[que_idx] next_que = ques[(que_idx + 1) % 2] dag_view = [] while que.qsize() != 0: op = que.get() dag_view.append(op) op_name = op.name sorted_op_num += 1 for succ_op in outdegs[op_name]: indeg_num[succ_op.name] -= 1 if indeg_num[succ_op.name] == 0: next_que.put(succ_op) dag_views.append(dag_view) if next_que.qsize() == 0: break que_idx = (que_idx + 1) % 2 if sorted_op_num < len(self._user_ops): raise Exception("not legal DAG") if len(dag_views[0]) != 1: raise Exception("DAG contains multiple input Ops") if len(dag_views[-1]) != 1: raise Exception("DAG contains multiple output Ops") # create channels and virtual ops virtual_op_idx = 0 channel_idx = 0 virtual_ops = [] channels = [] input_channel = None for v_idx, view in enumerate(dag_views): if v_idx + 1 >= len(dag_views): break next_view = dag_views[v_idx + 1] actual_next_view = [] pred_op_of_next_view_op = {} for op in view: # create virtual op for succ_op in outdegs[op.name]: if succ_op in next_view: actual_next_view.append(succ_op) if succ_op.name not in pred_op_of_next_view_op: pred_op_of_next_view_op[succ_op.name] = [] pred_op_of_next_view_op[succ_op.name].append(op) else: vop = Op(name="vir{}".format(virtual_op_idx), inputs=[]) virtual_op_idx += 1 virtual_ops.append(virtual_op) outdegs[vop.name] = [succ_op] actual_next_view.append(vop) # TODO: combine vop pred_op_of_next_view_op[vop.name] = [op] # create channel processed_op = set() for o_idx, op in enumerate(actual_next_view): op_name = op.name if op_name in processed_op: continue channel = Channel(name="chl{}".format(channel_idx)) channel_idx += 1 channels.append(channel) op.add_input_channel(channel) pred_ops = pred_op_of_next_view_op[op_name] if v_idx == 0: input_channel = channel else: for pred_op in pred_ops: pred_op.add_output_channel(channel) processed_op.add(op_name) # combine channel for other_op in actual_next_view[o_idx:]: if other_op.name in processed_op: continue other_pred_ops = pred_op_of_next_view_op[other_op.name] if len(other_pred_ops) != len(pred_ops): continue same_flag = True for pred_op in pred_ops: if pred_op not in other_pred_ops: same_flag = False break if same_flag: other_op.add_input_channel(channel) processed_op.add(other_op.name) output_channel = Channel(name="Ochl") channels.append(output_channel) last_op = dag_views[-1][0] last_op.add_output_channel(output_channel) self._ops = virtual_ops for op in self._user_ops: if len(op.get_input_ops()) == 0: continue self._ops.append(op) self._channels = channels return input_channel, output_channel def prepare_server(self, port, worker_num): self._port = port self._worker_num = worker_num input_channel, output_channel = self._topo_sort() self._in_channel = input_channel self._out_channel = output_channel for op in self._ops: if op.with_serving(): self.prepare_serving(op) self.gen_desc() def _op_start_wrapper(self, op, concurrency_idx): return op.start(concurrency_idx) def _run_ops(self): for op in self._ops: op_concurrency = op.get_concurrency() logging.debug("run op: {}, op_concurrency: {}".format( op.name, op_concurrency)) for c in range(op_concurrency): th = threading.Thread( target=self._op_start_wrapper, args=(op, c)) th.start() self._op_threads.append(th) def _stop_ops(self): for op in self._ops: op.stop() def run_server(self): self._run_ops() server = grpc.server( futures.ThreadPoolExecutor(max_workers=self._worker_num)) general_python_service_pb2_grpc.add_GeneralPythonServiceServicer_to_server( GeneralPythonService(self._in_channel, self._out_channel, self._retry), server) server.add_insecure_port('[::]:{}'.format(self._port)) server.start() server.wait_for_termination() self._stop_ops() # TODO for th in self._op_threads: th.join() def prepare_serving(self, op): model_path = op._server_model port = op._server_port device = op._device if device == "cpu": cmd = "(Use MultiLangServer) python -m paddle_serving_server.serve" \ " --model {} --thread 4 --port {} --use_multilang &>/dev/null &".format(model_path, port) else: cmd = "(Use MultiLangServer) python -m paddle_serving_server_gpu.serve" \ " --model {} --thread 4 --port {} --use_multilang &>/dev/null &".format(model_path, port) # run a server (not in PyServing) logging.info("run a server (not in PyServing): {}".format(cmd)) return # os.system(cmd)