Python mo_threads.Thread类代码示例

    def __init__(self, name):
        self.name = name
        self.lock = Lock("rate locker")
        self.request_rate = 0.0
        self.last_request = Date.now()

        Thread.run("rate logger", self._daemon)

    def __init__(self, rate=None, amortization_period=None, source=None, database=None, kwargs=None):
        self.amortization_period = coalesce(amortization_period, AMORTIZATION_PERIOD)
        self.rate = coalesce(rate, HG_REQUEST_PER_SECOND)
        self.cache_locker = Lock()
        self.cache = {}  # MAP FROM url TO (ready, headers, response, timestamp) PAIR
        self.no_cache = {}  # VERY SHORT TERM CACHE
        self.workers = []
        self.todo = Queue(APP_NAME+" todo")
        self.requests = Queue(APP_NAME + " requests", max=int(self.rate * self.amortization_period.seconds))
        self.url = URL(source.url)
        self.db = Sqlite(database)
        self.inbound_rate = RateLogger("Inbound")
        self.outbound_rate = RateLogger("hg.mo")

        if not self.db.query("SELECT name FROM sqlite_master WHERE type='table'").data:
            with self.db.transaction() as t:
                t.execute(
                    "CREATE TABLE cache ("
                    "   path TEXT PRIMARY KEY, "
                    "   headers TEXT, "
                    "   response TEXT, "
                    "   timestamp REAL "
                    ")"
                )

        self.threads = [
            Thread.run(APP_NAME+" worker" + text_type(i), self._worker)
            for i in range(CONCURRENCY)
        ]
        self.limiter = Thread.run(APP_NAME+" limiter", self._rate_limiter)
        self.cleaner = Thread.run(APP_NAME+" cleaner", self._cache_cleaner)

    def __init__(self, host, index, alias=None, name=None, port=9200, kwargs=None):
        global _elasticsearch
        if hasattr(self, "settings"):
            return

        from pyLibrary.queries.containers.list_usingPythonList import ListContainer
        from pyLibrary.env import elasticsearch as _elasticsearch

        self.settings = kwargs
        self.default_name = coalesce(name, alias, index)
        self.default_es = _elasticsearch.Cluster(kwargs=kwargs)
        self.todo = Queue("refresh metadata", max=100000, unique=True)

        self.es_metadata = Null
        self.last_es_metadata = Date.now()-OLD_METADATA

        self.meta=Data()
        table_columns = metadata_tables()
        column_columns = metadata_columns()
        self.meta.tables = ListContainer("meta.tables", [], wrap({c.names["."]: c for c in table_columns}))
        self.meta.columns = ColumnList()
        self.meta.columns.insert(column_columns)
        self.meta.columns.insert(table_columns)
        # TODO: fix monitor so it does not bring down ES
        if ENABLE_META_SCAN:
            self.worker = Thread.run("refresh metadata", self.monitor)
        else:
            self.worker = Thread.run("refresh metadata", self.not_monitor)
        return

def capture_termination_signal(please_stop):
    """
    WILL SIGNAL please_stop WHEN THIS AWS INSTANCE IS DUE FOR SHUTDOWN
    """
    def worker(please_stop):
        seen_problem = False
        while not please_stop:
            request_time = (time.time() - timer.START)/60  # MINUTES
            try:
                response = requests.get("http://169.254.169.254/latest/meta-data/spot/termination-time")
                seen_problem = False
                if response.status_code not in [400, 404]:
                    Log.alert("Shutdown AWS Spot Node {{name}} {{type}}", name=machine_metadata.name, type=machine_metadata.aws_instance_type)
                    please_stop.go()
            except Exception as e:
                e = Except.wrap(e)
                if "Failed to establish a new connection: [Errno 10060]" in e or "A socket operation was attempted to an unreachable network" in e:
                    Log.note("AWS Spot Detection has shutdown, probably not a spot node, (http://169.254.169.254 is unreachable)")
                    return
                elif seen_problem:
                    # IGNORE THE FIRST PROBLEM
                    Log.warning("AWS shutdown detection has more than one consecutive problem: (last request {{time|round(1)}} minutes since startup)", time=request_time, cause=e)
                seen_problem = True

                (Till(seconds=61) | please_stop).wait()
            (Till(seconds=11) | please_stop).wait()

    Thread.run("listen for termination", worker)

def queue_consumer(pull_queue, please_stop=None):
    queue = aws.Queue(pull_queue)
    time_offset = None
    request_count = 0

    while not please_stop:
        request = queue.pop(till=please_stop)
        if please_stop:
            break
        if not request:
            Log.note("Nothing in queue, pausing for 5 seconds...")
            (please_stop | Till(seconds=5)).wait()
            continue

        if SKIP_TRY_REQUESTS and 'try' in request.where['and'].eq.branch:
            Log.note("Skipping try revision.")
            queue.commit()
            continue

        now = Date.now().unix
        if time_offset is None:
            time_offset = now - request.meta.request_time

        next_request = request.meta.request_time + time_offset
        if next_request > now:
            Log.note("Next request in {{wait_time}}", wait_time=Duration(seconds=next_request - now))
            Till(till=next_request).wait()

        Thread.run("request "+text_type(request_count), one_request, request)
        request_count += 1
        queue.commit()

 def inners():
     for t in data.hits.hits:
         for i in t.inner_hits[literal_field(query_path)].hits.hits:
             t._inner = i._source
             for k, e in post_expressions.items():
                 t[k] = e(t)
             yield t
     if more_filter:
         Thread.join(need_more)
         for t in more[0].hits.hits:
             yield t

    def __exit__(self, exc_type, exc_val, exc_tb):
        Log.note("clean pulse exit")
        self.please_stop.go()
        with suppress_exception:
            self.target_queue.close()
            Log.note("stop put into queue")

        try:
            self.pulse.disconnect()
        except Exception as e:
            Log.warning("Can not disconnect during pulse exit, ignoring", e)
        Thread.__exit__(self, exc_type, exc_val, exc_tb)

class StructuredLogger_usingThreadedStream(StructuredLogger):
    # stream CAN BE AN OBJCET WITH write() METHOD, OR A STRING
    # WHICH WILL eval() TO ONE
    def __init__(self, stream):
        assert stream

        if isinstance(stream, text_type):
            name = stream
            stream = self.stream = eval(stream)
            if name.startswith("sys.") and PY3:
                self.stream = Data(write=lambda d: stream.write(d.decode('utf8')))
        else:
            name = "stream"
            self.stream = stream

        # WRITE TO STREAMS CAN BE *REALLY* SLOW, WE WILL USE A THREAD
        from mo_threads import Queue

        def utf8_appender(value):
            if isinstance(value, text_type):
                value = value.encode('utf8')
            self.stream.write(value)

        appender = utf8_appender

        self.queue = Queue("queue for " + self.__class__.__name__ + "(" + name + ")", max=10000, silent=True)
        self.thread = Thread("log to " + self.__class__.__name__ + "(" + name + ")", time_delta_pusher, appender=appender, queue=self.queue, interval=0.3)
        self.thread.parent.remove_child(self.thread)  # LOGGING WILL BE RESPONSIBLE FOR THREAD stop()
        self.thread.start()

    def write(self, template, params):
        try:
            self.queue.add({"template": template, "params": params})
            return self
        except Exception as e:
            raise e  # OH NO!

    def stop(self):
        try:
            self.queue.add(THREAD_STOP)  # BE PATIENT, LET REST OF MESSAGE BE SENT
            self.thread.join()
        except Exception as e:
            if DEBUG_LOGGING:
                raise e

        try:
            self.queue.close()
        except Exception as f:
            if DEBUG_LOGGING:
                raise f

    def __init__(self):
        self.out_of_memory_restart = False

        self.total_locker = Lock()
        self.total_files_requested = 0
        self.total_tuids_mapped = 0

        self.threads_locker = Lock()
        self.waiting = 0
        self.threads_waiting = 0

        self.requests_locker = Lock()
        self.requests_total = 0
        self.requests_complete = 0
        self.requests_incomplete = 0
        self.requests_passed = 0
        self.requests_failed = 0

        self.prev_mem = 0
        self.curr_mem = 0
        self.initial_growth = {}

        Thread.run("pc-daemon", self.run_pc_daemon)
        Thread.run("threads-daemon", self.run_threads_daemon)
        Thread.run("memory-daemon", self.run_memory_daemon)
        Thread.run("requests-daemon", self.run_requests_daemon)

class StructuredLogger_usingThread(StructuredLogger):

    def __init__(self, logger):
        if not isinstance(logger, StructuredLogger):
            Log.error("Expecting a StructuredLogger")

        self.queue = Queue("Queue for " + self.__class__.__name__, max=10000, silent=True, allow_add_after_close=True)
        self.logger = logger

        def worker(logger, please_stop):
            try:
                while not please_stop:
                    logs = self.queue.pop_all()
                    if not logs:
                        (Till(seconds=1) | please_stop).wait()
                        continue
                    for log in logs:
                        if log is THREAD_STOP:
                            please_stop.go()
                        else:
                            logger.write(**log)
            except Exception as e:
                print("problem in " + StructuredLogger_usingThread.__name__ + ": " + str(e))
            finally:
                Log.note("stop the child")
                logger.stop()

        self.thread = Thread("Thread for " + self.__class__.__name__, worker, logger)
        self.thread.parent.remove_child(self.thread)  # LOGGING WILL BE RESPONSIBLE FOR THREAD stop()
        self.thread.start()

    def write(self, template, params):
        try:
            self.queue.add({"template": template, "params": params})
            return self
        except Exception as e:
            e = Except.wrap(e)
            raise e  # OH NO!

    def stop(self):
        try:
            self.queue.add(THREAD_STOP)  # BE PATIENT, LET REST OF MESSAGE BE SENT
            self.thread.join()
            Log.note("joined on thread")
        except Exception as e:
            Log.note("problem in threaded logger" + str(e))

        with suppress_exception:
            self.queue.close()

 def __init__(self, name):
     Table.__init__(self, "meta.columns")
     self.db_file = File("metadata." + name + ".sqlite")
     self.data = {}  # MAP FROM ES_INDEX TO (abs_column_name to COLUMNS)
     self.locker = Lock()
     self._schema = None
     self.db = sqlite3.connect(
         database=self.db_file.abspath, check_same_thread=False, isolation_level=None
     )
     self.last_load = Null
     self.todo = Queue(
         "update columns to db"
     )  # HOLD (action, column) PAIR, WHERE action in ['insert', 'update']
     self._db_load()
     Thread.run("update " + name, self._db_worker)

    def __init__(self, name, config):
        config = wrap(config)
        if config.debug.logs:
            Log.error("not allowed to configure logging on other process")

        self.process = Process(name, [PYTHON, "mo_threads" + os.sep + "python_worker.py"], shell=True)
        self.process.stdin.add(value2json(set_default({"debug": {"trace": True}}, config)))

        self.lock = Lock("wait for response from "+name)
        self.current_task = None
        self.current_response = None
        self.current_error = None

        self.daemon = Thread.run("", self._daemon)
        self.errors = Thread.run("", self._stderr)

    def __init__(self, stream):
        assert stream

        if isinstance(stream, text_type):
            name = stream
            stream = self.stream = eval(stream)
            if name.startswith("sys.") and PY3:
                self.stream = Data(write=lambda d: stream.write(d.decode('utf8')))
        else:
            name = "stream"
            self.stream = stream

        # WRITE TO STREAMS CAN BE *REALLY* SLOW, WE WILL USE A THREAD
        from mo_threads import Queue

        def utf8_appender(value):
            if isinstance(value, text_type):
                value = value.encode('utf8')
            self.stream.write(value)

        appender = utf8_appender

        self.queue = Queue("queue for " + self.__class__.__name__ + "(" + name + ")", max=10000, silent=True)
        self.thread = Thread("log to " + self.__class__.__name__ + "(" + name + ")", time_delta_pusher, appender=appender, queue=self.queue, interval=0.3)
        self.thread.parent.remove_child(self.thread)  # LOGGING WILL BE RESPONSIBLE FOR THREAD stop()
        self.thread.start()

    def __init__(
        self,
        host,
        index,
        port=9200,
        type="log",
        queue_size=1000,
        batch_size=100,
        kwargs=None,
    ):
        """
        settings ARE FOR THE ELASTICSEARCH INDEX
        """
        kwargs.timeout = Duration(coalesce(kwargs.timeout, "30second")).seconds
        kwargs.retry.times = coalesce(kwargs.retry.times, 3)
        kwargs.retry.sleep = Duration(coalesce(kwargs.retry.sleep, MINUTE)).seconds
        kwargs.host = Random.sample(listwrap(host), 1)[0]

        schema = json2value(value2json(SCHEMA), leaves=True)
        schema.mappings[type].properties["~N~"].type = "nested"
        self.es = Cluster(kwargs).get_or_create_index(
            schema=schema,
            limit_replicas=True,
            typed=True,
            kwargs=kwargs,
        )
        self.batch_size = batch_size
        self.es.add_alias(coalesce(kwargs.alias, kwargs.index))
        self.queue = Queue("debug logs to es", max=queue_size, silent=True)

        self.worker = Thread.run("add debug logs to es", self._insert_loop)

 def __init__(self, host, index, port=9200, type="log", max_size=1000, batch_size=100, kwargs=None):
     """
     settings ARE FOR THE ELASTICSEARCH INDEX
     """
     self.es = Cluster(kwargs).get_or_create_index(
         schema=mo_json.json2value(value2json(SCHEMA), leaves=True),
         limit_replicas=True,
         tjson=True,
         kwargs=kwargs
     )
     self.batch_size = batch_size
     self.es.add_alias(coalesce(kwargs.alias, kwargs.index))
     self.queue = Queue("debug logs to es", max=max_size, silent=True)
     self.es.settings.retry.times = coalesce(self.es.settings.retry.times, 3)
     self.es.settings.retry.sleep = Duration(coalesce(self.es.settings.retry.sleep, MINUTE))
     Thread.run("add debug logs to es", self._insert_loop)