Python mo_dots.coalesce函数代码示例

    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 _range_composer(edge, domain, es_query, to_float, schema):
    # USE RANGES
    _min = coalesce(domain.min, MIN(domain.partitions.min))
    _max = coalesce(domain.max, MAX(domain.partitions.max))

    if edge.allowNulls:
        missing_filter = set_default(
            {
                "filter": NotOp("not", AndOp("and", [
                    edge.value.exists(),
                    InequalityOp("gte", [edge.value, Literal(None, to_float(_min))]),
                    InequalityOp("lt", [edge.value, Literal(None, to_float(_max))])
                ]).partial_eval()).to_esfilter(schema)
            },
            es_query
        )
    else:
        missing_filter = None

    if isinstance(edge.value, Variable):
        calc = {"field": schema.leaves(edge.value.var)[0].es_column}
    else:
        calc = {"script": edge.value.to_es_script(schema).script(schema)}

    return wrap({"aggs": {
        "_match": set_default(
            {"range": calc},
            {"range": {"ranges": [{"from": to_float(p.min), "to": to_float(p.max)} for p in domain.partitions]}},
            es_query
        ),
        "_missing": missing_filter
    }})

    def __init__(self, value, port=None, path=None, query=None, fragment=None):
        try:
            self.scheme = None
            self.host = None
            self.port = port
            self.path = path
            self.query = query
            self.fragment = fragment

            if value == None:
                return

            if value.startswith("file://") or value.startswith("//"):
                # urlparse DOES NOT WORK IN THESE CASES
                scheme, suffix = value.split("//", 2)
                self.scheme = scheme.rstrip(":")
                parse(self, suffix, 0, 1)
                self.query = wrap(url_param2value(self.query))
            else:
                output = urlparse(value)
                self.scheme = output.scheme
                self.port = coalesce(port, output.port)
                self.host = output.netloc.split(":")[0]
                self.path = coalesce(path, output.path)
                self.query = coalesce(query, wrap(url_param2value(output.query)))
                self.fragment = coalesce(fragment, output.fragment)
        except Exception as e:
            Log.error(u"problem parsing {{value}} to URL", value=value, cause=e)

def Stats2ZeroMoment(stats):
    # MODIFIED FROM http://statsmodels.sourceforge.net/devel/_modules/statsmodels/stats/moment_helpers.html
    # ADDED count
    mc0, mc1, mc2, skew, kurt = stats.count, coalesce(stats.mean, 0), coalesce(stats.variance, 0), coalesce(stats.skew, 0), coalesce(stats.kurtosis, 0)

    mz0 = mc0
    mz1 = mc1 * mc0
    mz2 = (mc2 + mc1 * mc1) * mc0
    mc3 = coalesce(skew, 0) * (mc2 ** 1.5) # 3rd central moment
    mz3 = (mc3 + 3 * mc1 * mc2 + mc1 ** 3) * mc0  # 3rd non-central moment
    mc4 = (coalesce(kurt, 0) + 3.0) * (mc2 ** 2.0) # 4th central moment
    mz4 = (mc4 + 4 * mc1 * mc3 + 6 * mc1 * mc1 * mc2 + mc1 ** 4) * mc0

    m = ZeroMoment(mz0, mz1, mz2, mz3, mz4)
    if DEBUG:
        from mo_testing.fuzzytestcase import assertAlmostEqualValue

        globals()["DEBUG"] = False
        try:
            v = ZeroMoment2Stats(m)
            assertAlmostEqualValue(v.count, stats.count, places=10)
            assertAlmostEqualValue(v.mean, stats.mean, places=10)
            assertAlmostEqualValue(v.variance, stats.variance, places=10)
            assertAlmostEqualValue(v.skew, stats.skew, places=10)
            assertAlmostEqualValue(v.kurtosis, stats.kurtosis, places=10)
        except Exception as e:
            v = ZeroMoment2Stats(m)
            Log.error("programmer error")
        globals()["DEBUG"] = True
    return m

    def __init__(self, instance_manager, disable_prices=False, kwargs=None):
        self.settings = kwargs
        self.instance_manager = instance_manager
        aws_args = dict(
            region_name=kwargs.aws.region,
            aws_access_key_id=unwrap(kwargs.aws.aws_access_key_id),
            aws_secret_access_key=unwrap(kwargs.aws.aws_secret_access_key)
        )
        self.ec2_conn = boto.ec2.connect_to_region(**aws_args)
        self.vpc_conn = boto.vpc.connect_to_region(**aws_args)
        self.price_locker = Lock()
        self.prices = None
        self.price_lookup = None
        self.done_spot_requests = Signal()
        self.net_new_locker = Lock()
        self.net_new_spot_requests = UniqueIndex(("id",))  # SPOT REQUESTS FOR THIS SESSION
        self.watcher = None
        self.active = None

        self.settings.uptime.bid_percentile = coalesce(self.settings.uptime.bid_percentile, self.settings.bid_percentile)
        self.settings.uptime.history = coalesce(Date(self.settings.uptime.history), DAY)
        self.settings.uptime.duration = coalesce(Duration(self.settings.uptime.duration), Date("5minute"))
        self.settings.max_percent_per_type = coalesce(self.settings.max_percent_per_type, 1)

        if ENABLE_SIDE_EFFECTS and instance_manager and instance_manager.setup_required():
            self._start_life_cycle_watcher()
        if not disable_prices:
            self.pricing()

 def _open(self):
     """ DO NOT USE THIS UNLESS YOU close() FIRST"""
     try:
         self.db = connect(
             host=self.settings.host,
             port=self.settings.port,
             user=coalesce(self.settings.username, self.settings.user),
             passwd=coalesce(self.settings.password, self.settings.passwd),
             db=coalesce(self.settings.schema, self.settings.db),
             read_timeout=coalesce(self.settings.read_timeout, (EXECUTE_TIMEOUT / 1000) - 10 if EXECUTE_TIMEOUT else None, 5*60),
             charset=u"utf8",
             use_unicode=True,
             ssl=coalesce(self.settings.ssl, None),
             cursorclass=cursors.SSCursor
         )
     except Exception as e:
         if self.settings.host.find("://") == -1:
             Log.error(
                 u"Failure to connect to {{host}}:{{port}}",
                 host=self.settings.host,
                 port=self.settings.port,
                 cause=e
             )
         else:
             Log.error(u"Failure to connect.  PROTOCOL PREFIX IS PROBABLY BAD", e)
     self.cursor = None
     self.partial_rollback = False
     self.transaction_level = 0
     self.backlog = []  # accumulate the write commands so they are sent at once
     if self.readonly:
         self.begin()

    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, **desc):
     desc = wrap(desc)
     self._set_slots_to_null(self.__class__)
     set_default(self, desc)
     self.name = coalesce(desc.name, desc.type)
     self.isFacet = coalesce(desc.isFacet, False)
     self.dimension = Null
     self.limit = desc.limit

def get_decoders_by_depth(query):
    """
    RETURN A LIST OF DECODER ARRAYS, ONE ARRAY FOR EACH NESTED DEPTH
    """
    schema = query.frum.schema
    output = FlatList()

    if query.edges:
        if query.sort and query.format != "cube":
            # REORDER EDGES/GROUPBY TO MATCH THE SORT
            query.edges = sort_edges(query, "edges")
    elif query.groupby:
        if query.sort and query.format != "cube":
            query.groupby = sort_edges(query, "groupby")

    for edge in wrap(coalesce(query.edges, query.groupby, [])):
        limit = coalesce(edge.domain.limit, query.limit, DEFAULT_LIMIT)
        if edge.value != None and not isinstance(edge.value, NullOp):
            edge = edge.copy()
            vars_ = edge.value.vars()
            for v in vars_:
                if not schema.leaves(v.var):
                    Log.error("{{var}} does not exist in schema", var=v)
        elif edge.range:
            vars_ = edge.range.min.vars() | edge.range.max.vars()
            for v in vars_:
                if not schema[v.var]:
                    Log.error("{{var}} does not exist in schema", var=v)
        elif edge.domain.dimension:
            vars_ = edge.domain.dimension.fields
            edge.domain.dimension = edge.domain.dimension.copy()
            edge.domain.dimension.fields = [schema[v].es_column for v in vars_]
        elif all(edge.domain.partitions.where):
            vars_ = set()
            for p in edge.domain.partitions:
                vars_ |= p.where.vars()

        try:
            vars_ |= edge.value.vars()
            depths = set(len(c.nested_path) - 1 for v in vars_ for c in schema.leaves(v.var))
            if -1 in depths:
                Log.error(
                    "Do not know of column {{column}}",
                    column=unwraplist([v for v in vars_ if schema[v] == None])
                )
            if len(depths) > 1:
                Log.error("expression {{expr|quote}} spans tables, can not handle", expr=edge.value)
            max_depth = MAX(depths)
            while len(output) <= max_depth:
                output.append([])
        except Exception as e:
            # USUALLY THE SCHEMA IS EMPTY, SO WE ASSUME THIS IS A SIMPLE QUERY
            max_depth = 0
            output.append([])

        output[max_depth].append(AggsDecoder(edge, query, limit))
    return output

 def single(col, r):
     min = coalesce(r["gte"], r[">="])
     max = coalesce(r["lte"], r["<="])
     if min and max:
         # SPECIAL CASE (BETWEEN)
         return db.quote_column(col) + SQL(" BETWEEN ") + db.quote_value(min) + SQL(" AND ") + db.quote_value(max)
     else:
         return " AND ".join(
             db.quote_column(col) + name2sign[sign] + db.quote_value(value)
             for sign, value in r.items()
         )

 def single(col, r):
     min = coalesce(r["gte"], r[">="])
     max = coalesce(r["lte"], r["<="])
     if min != None and max != None:
         # SPECIAL CASE (BETWEEN)
         sql = quote_column(col) + SQL(" BETWEEN ") + quote_value(min) + SQL_AND + quote_value(max)
     else:
         sql = SQL_AND.join(
             quote_column(col) + name2sign[sign] + quote_value(value)
             for sign, value in r.items()
         )
     return sql

    def _get_from_elasticsearch(self, revision, locale=None, get_diff=False, get_moves=True):
        rev = revision.changeset.id
        if self.es.cluster.version.startswith("1.7."):
            query = {
                "query": {"filtered": {
                    "query": {"match_all": {}},
                    "filter": {"and": [
                        {"term": {"changeset.id12": rev[0:12]}},
                        {"term": {"branch.name": revision.branch.name}},
                        {"term": {"branch.locale": coalesce(locale, revision.branch.locale, DEFAULT_LOCALE)}},
                        {"range": {"etl.timestamp": {"gt": MIN_ETL_AGE}}}
                    ]}
                }},
                "size": 20
            }
        else:
            query = {
                "query": {"bool": {"must": [
                    {"term": {"changeset.id12": rev[0:12]}},
                    {"term": {"branch.name": revision.branch.name}},
                    {"term": {"branch.locale": coalesce(locale, revision.branch.locale, DEFAULT_LOCALE)}},
                    {"range": {"etl.timestamp": {"gt": MIN_ETL_AGE}}}
                ]}},
                "size": 20
            }

        for attempt in range(3):
            try:
                with self.es_locker:
                    docs = self.es.search(query).hits.hits
                if len(docs) == 0:
                    return None
                best = docs[0]._source
                if len(docs) > 1:
                    for d in docs:
                        if d._id.endswith(d._source.branch.locale):
                            best = d._source
                    Log.warning("expecting no more than one document")
                return best
            except Exception as e:
                e = Except.wrap(e)
                if "EsRejectedExecutionException[rejected execution (queue capacity" in e:
                    (Till(seconds=Random.int(30))).wait()
                    continue
                else:
                    Log.warning("Bad ES call, waiting for {{num}} seconds", num=WAIT_AFTER_NODE_FAILURE, cause=e)
                    Till(seconds=WAIT_AFTER_NODE_FAILURE).wait()
                    continue

        Log.warning("ES did not deliver, fall back to HG")
        return None

    def send_email(self,
            from_address=None,
            to_address=None,
            subject=None,
            text_data=None,
            html_data=None
    ):
        """Sends an email.

        from_addr is an email address; to_addrs is a list of email adresses.
        Addresses can be plain (e.g. "[email protected]") or with real names
        (e.g. "John Smith <[email protected]>").

        text_data and html_data are both strings.  You can specify one or both.
        If you specify both, the email will be sent as a MIME multipart
        alternative, i.e., the recipient will see the HTML content if his
        viewer supports it; otherwise he'll see the text content.
        """

        settings = self.settings

        from_address = coalesce(from_address, settings["from"], settings.from_address)
        to_address = listwrap(coalesce(to_address, settings.to_address, settings.to_addrs))

        if not from_address or not to_address:
            raise Exception("Both from_addr and to_addrs must be specified")
        if not text_data and not html_data:
            raise Exception("Must specify either text_data or html_data")

        if not html_data:
            msg = MIMEText(text_data)
        elif not text_data:
            msg = MIMEText(html_data, 'html')
        else:
            msg = MIMEMultipart('alternative')
            msg.attach(MIMEText(text_data, 'plain'))
            msg.attach(MIMEText(html_data, 'html'))

        msg['Subject'] = coalesce(subject, settings.subject)
        msg['From'] = from_address
        msg['To'] = ', '.join(to_address)

        if self.server:
            # CALL AS PART OF A SMTP SESSION
            self.server.sendmail(from_address, to_address, msg.as_string())
        else:
            # CALL AS STAND-ALONE
            with self:
                self.server.sendmail(from_address, to_address, msg.as_string())

def compileDuration2Term(edge):
    if edge.esscript:
        Log.error("edge script not supported yet")

    # IS THERE A LIMIT ON THE DOMAIN?
    numPartitions = len(edge.domain.partitions)
    value = edge.value
    if is_variable_name(value):
        value = "doc[\"" + value + "\"].value"

    ref = coalesce(edge.domain.min, edge.domain.max, durations.ZERO)
    nullTest = compileNullTest(edge)

    ms = edge.domain.interval.milli
    if edge.domain.interval.month > 0:
        ms = durations.YEAR.milli / 12 * edge.domain.interval.month

    partition2int = "Math.floor((" + value + "-" + value2MVEL(ref) + ")/" + ms + ")"
    partition2int = "((" + nullTest + ") ? " + numPartitions + " : " + partition2int + ")"

    def int2Partition(value):
        if Math.round(value) == numPartitions:
            return edge.domain.NULL
        return edge.domain.getPartByKey(ref.add(edge.domain.interval.multiply(value)))

    return Data(toTerm={"head": "", "body": partition2int}, fromTerm=int2Partition)

 def convert(self, expr):
     """
     EXPAND INSTANCES OF name TO value
     """
     if expr is True or expr == None or expr is False:
         return expr
     elif Math.is_number(expr):
         return expr
     elif expr == ".":
         return "."
     elif is_variable_name(expr):
         return coalesce(self.dimensions[expr], expr)
     elif isinstance(expr, text_type):
         Log.error("{{name|quote}} is not a valid variable name", name=expr)
     elif isinstance(expr, Date):
         return expr
     elif isinstance(expr, QueryOp):
         return self._convert_query(expr)
     elif isinstance(expr, Mapping):
         if expr["from"]:
             return self._convert_query(expr)
         elif len(expr) >= 2:
             #ASSUME WE HAVE A NAMED STRUCTURE, NOT AN EXPRESSION
             return wrap({name: self.convert(value) for name, value in expr.leaves()})
         else:
             # ASSUME SINGLE-CLAUSE EXPRESSION
             k, v = expr.items()[0]
             return converter_map.get(k, self._convert_bop)(self, k, v)
     elif isinstance(expr, (list, set, tuple)):
         return wrap([self.convert(value) for value in expr])
     else:
         return expr