Python mo_dots.unwrap函数代码示例

    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 __init__(self, **desc):
        Domain.__init__(self, **desc)
        self.type = "range"
        self.NULL = Null

        if self.partitions:
            # IGNORE THE min, max, interval
            if not self.key:
                Log.error("Must have a key value")

            parts =listwrap(self.partitions)
            for i, p in enumerate(parts):
                self.min = MIN([self.min, p.min])
                self.max = MAX([self.max, p.max])
                if p.dataIndex != None and p.dataIndex != i:
                    Log.error("Expecting `dataIndex` to agree with the order of the parts")
                if p[self.key] == None:
                    Log.error("Expecting all parts to have {{key}} as a property", key=self.key)
                p.dataIndex = i

            # VERIFY PARTITIONS DO NOT OVERLAP, HOLES ARE FINE
            for p, q in itertools.product(parts, parts):
                if p.min <= q.min and q.min < p.max and unwrap(p) is not unwrap(q):
                    Log.error("partitions overlap!")

            self.partitions = wrap(parts)
            return
        elif any([self.min == None, self.max == None, self.interval == None]):
            Log.error("Can not handle missing parameter")

        self.key = "min"
        self.partitions = wrap([{"min": v, "max": v + self.interval, "dataIndex": i} for i, v in enumerate(frange(self.min, self.max, self.interval))])

    def _send_email(self):
        try:
            if not self.accumulation:
                return
            with Closer(connect_to_region(
                self.settings.region,
                aws_access_key_id=unwrap(self.settings.aws_access_key_id),
                aws_secret_access_key=unwrap(self.settings.aws_secret_access_key)
            )) as conn:

                # WHO ARE WE SENDING TO
                emails = Data()
                for template, params in self.accumulation:
                    content = expand_template(template, params)
                    emails[literal_field(self.settings.to_address)] += [content]
                    for c in self.cc:
                        if any(c in params.params.error for c in c.contains):
                            emails[literal_field(c.to_address)] += [content]

                # SEND TO EACH
                for to_address, content in emails.items():
                    conn.send_email(
                        source=self.settings.from_address,
                        to_addresses=listwrap(to_address),
                        subject=self.settings.subject,
                        body="\n\n".join(content),
                        format="text"
                    )

            self.next_send = Date.now() + self.settings.max_interval
            self.accumulation = []
        except Exception as e:
            self.next_send = Date.now() + self.settings.max_interval
            Log.warning("Could not send", e)

def filter(data, where):
    """
    where  - a function that accepts (record, rownum, rows) and returns boolean
    """
    if len(data) == 0 or where == None or where == TRUE:
        return data

    if isinstance(data, Container):
        return data.filter(where)

    if is_container(data):
        temp = jx_expression_to_function(where)
        dd = wrap(data)
        return wrap([unwrap(d) for i, d in enumerate(data) if temp(wrap(d), i, dd)])
    else:
        Log.error(
            "Do not know how to handle type {{type}}", type=data.__class__.__name__
        )

    try:
        return drill_filter(where, data)
    except Exception as _:
        # WOW!  THIS IS INEFFICIENT!
        return wrap(
            [unwrap(d) for d in drill_filter(where, [DataObject(d) for d in data])]
        )

def aggs_iterator(aggs, decoders, coord=True):
    """
    DIG INTO ES'S RECURSIVE aggs DATA-STRUCTURE:
    RETURN AN ITERATOR OVER THE EFFECTIVE ROWS OF THE RESULTS

    :param aggs: ES AGGREGATE OBJECT
    :param decoders:
    :param coord: TURN ON LOCAL COORDINATE LOOKUP
    """
    depth = max(d.start + d.num_columns for d in decoders)

    def _aggs_iterator(agg, d):
        agg = drill(agg)

        if d > 0:
            for k, v in agg.items():
                if k == "_match":
                    v = drill(v)
                    for i, b in enumerate(v.get("buckets", EMPTY_LIST)):
                        b["_index"] = i
                        for a, parts in _aggs_iterator(b, d - 1):
                            yield a, parts + (b,)
                elif k == "_other":
                    for b in v.get("buckets", EMPTY_LIST):
                        for a, parts in _aggs_iterator(b, d - 1):
                            yield a, parts + (Null,)
                elif k == "_missing":
                    b = drill(v)
                    for a, parts in _aggs_iterator(b, d - 1):
                        yield a, parts + (b,)
                elif k.startswith("_join_"):
                    v["key"] = int(k[6:])
                    for a, parts in _aggs_iterator(v, d - 1):
                        yield a, parts + (v,)
        else:
            for k, v in agg.items():
                if k == "_match":
                    v = drill(v)
                    for i, b in enumerate(v.get("buckets", EMPTY_LIST)):
                        b["_index"] = i
                        yield b, (b,)
                elif k == "_other":
                    for b in v.get("buckets", EMPTY_LIST):
                        yield b, (Null,)
                elif k == "_missing":
                    b = drill(v,)
                    yield b, (v,)
                elif k.startswith("_join_"):
                    v["_index"] = int(k[6:])
                    yield v, (v,)

    if coord:
        for a, parts in _aggs_iterator(unwrap(aggs), depth - 1):
            coord = tuple(d.get_index(parts) for d in decoders)
            if any(c is None for c in coord):
                continue
            yield parts, coord, a
    else:
        for a, parts in _aggs_iterator(unwrap(aggs), depth - 1):
            yield parts, None, a

    def extend(self, records):
        """
        JUST SO WE MODEL A Queue
        """
        records = {
            v["id"]: v["value"] if "value" in v else mo_json.json2value(v['json'])
            for v in records
        }

        unwrap(self.data).update(records)
        self.refresh()
        Log.note("{{num}} documents added", num=len(records))

def _where_terms(master, where, schema):
    """
    USE THE SCHEMA TO CONVERT DIMENSION NAMES TO ES FILTERS
    master - TOP LEVEL WHERE (FOR PLACING NESTED FILTERS)
    """
    if isinstance(where, Mapping):
        if where.term:
            # MAP TERM
            try:
                output = _map_term_using_schema(master, [], where.term, schema.edges)
                return output
            except Exception as e:
                Log.error("programmer problem?", e)
        elif where.terms:
            # MAP TERM
            output = FlatList()
            for k, v in where.terms.items():
                if not isinstance(v, (list, set)):
                    Log.error("terms filter expects list of values")
                edge = schema.edges[k]
                if not edge:
                    output.append({"terms": {k: v}})
                else:
                    if isinstance(edge, text_type):
                        # DIRECT FIELD REFERENCE
                        return {"terms": {edge: v}}
                    try:
                        domain = edge.getDomain()
                    except Exception as e:
                        Log.error("programmer error", e)
                    fields = domain.dimension.fields
                    if isinstance(fields, Mapping):
                        or_agg = []
                        for vv in v:
                            and_agg = []
                            for local_field, es_field in fields.items():
                                vvv = vv[local_field]
                                if vvv != None:
                                    and_agg.append({"term": {es_field: vvv}})
                            or_agg.append({"and": and_agg})
                        output.append({"or": or_agg})
                    elif isinstance(fields, list) and len(fields) == 1 and is_variable_name(fields[0]):
                        output.append({"terms": {fields[0]: v}})
                    elif domain.partitions:
                        output.append({"or": [domain.getPartByKey(vv).esfilter for vv in v]})
            return {"and": output}
        elif where["or"]:
            return {"or": [unwrap(_where_terms(master, vv, schema)) for vv in where["or"]]}
        elif where["and"]:
            return {"and": [unwrap(_where_terms(master, vv, schema)) for vv in where["and"]]}
        elif where["not"]:
            return {"not": unwrap(_where_terms(master, where["not"], schema))}
    return where

    def extend(self, records):
        """
        JUST SO WE MODEL A Queue
        """
        records = {v["id"]: v["value"] for v in records}

        unwrap(self.data).update(records)

        data_as_json = convert.value2json(self.data, pretty=True)

        File(self.filename).write(data_as_json)
        Log.note("{{num}} documents added",  num= len(records))

def es_fieldop(es, query):
    FromES = es09.util.build_es_query(query)
    select = listwrap(query.select)
    FromES.query = {
        "filtered": {
            "query": {
                "match_all": {}
            },
            "filter": simplify_esfilter(jx_expression(query.where).to_esfilter())
        }
    }
    FromES.size = coalesce(query.limit, 200000)
    FromES.fields = FlatList()
    for s in select.value:
        if s == "*":
            FromES.fields = None
        elif isinstance(s, list):
            FromES.fields.extend(s)
        elif isinstance(s, Mapping):
            FromES.fields.extend(s.values())
        else:
            FromES.fields.append(s)
    FromES.sort = [{s.field: "asc" if s.sort >= 0 else "desc"} for s in query.sort]

    data = es09.util.post(es, FromES, query.limit)

    T = data.hits.hits
    matricies = {}
    for s in select:
        if s.value == "*":
            matricies[s.name] = Matrix.wrap([t._source for t in T])
        elif isinstance(s.value, Mapping):
            # for k, v in s.value.items():
            #     matricies[join_field(split_field(s.name)+[k])] = Matrix.wrap([unwrap(t.fields)[v] for t in T])
            matricies[s.name] = Matrix.wrap([{k: unwrap(t.fields).get(v, None) for k, v in s.value.items()}for t in T])
        elif isinstance(s.value, list):
            matricies[s.name] = Matrix.wrap([tuple(unwrap(t.fields).get(ss, None) for ss in s.value) for t in T])
        elif not s.value:
            matricies[s.name] = Matrix.wrap([unwrap(t.fields).get(s.value, None) for t in T])
        else:
            try:
                matricies[s.name] = Matrix.wrap([unwrap(t.fields).get(s.value, None) for t in T])
            except Exception as e:
                Log.error("", e)

    cube = Cube(query.select, query.edges, matricies, frum=query)
    cube.frum = query
    return cube

    def __setitem__(self, key, value):
        if key == "":
            get_logger().error("key is empty string.  Probably a bad idea")
        if isinstance(key, str):
            key = key.decode("utf8")
        d=self
        try:
            value = unwrap(value)
            if key.find(".") == -1:
                if value is None:
                    dict.pop(d, key, None)
                else:
                    dict.__setitem__(d, key, value)
                return self

            seq = _split_field(key)
            for k in seq[:-1]:
                d = _getdefault(d, k)
            if value == None:
                dict.pop(d, seq[-1], None)
            else:
                dict.__setitem__(d, seq[-1], value)
            return self
        except Exception as e:
            raise e

def make_log_from_settings(settings):
    assert settings["class"]

    # IMPORT MODULE FOR HANDLER
    path = settings["class"].split(".")
    class_name = path[-1]
    path = ".".join(path[:-1])
    constructor = None
    try:
        temp = __import__(path, globals(), locals(), [class_name], 0)
        constructor = object.__getattribute__(temp, class_name)
    except Exception as e:
        if settings.stream and not constructor:
            # PROVIDE A DEFAULT STREAM HANLDER
            constructor = StructuredLogger_usingThreadedStream
        else:
            Log.error("Can not find class {{class}}",  {"class": path}, cause=e)

    # IF WE NEED A FILE, MAKE SURE DIRECTORY EXISTS
    if settings.filename != None:
        from mo_files import File

        f = File(settings.filename)
        if not f.parent.exists:
            f.parent.create()

    settings['class'] = None
    params = unwrap(settings)
    log_instance = constructor(**params)
    return log_instance

def tuple(data, field_name):
    """
    RETURN LIST  OF TUPLES
    """
    if isinstance(data, Cube):
        Log.error("not supported yet")

    if isinstance(data, FlatList):
        Log.error("not supported yet")

    if is_data(field_name) and "value" in field_name:
        # SIMPLIFY {"value":value} AS STRING
        field_name = field_name["value"]

    # SIMPLE PYTHON ITERABLE ASSUMED
    if is_text(field_name):
        if len(split_field(field_name)) == 1:
            return [(d[field_name],) for d in data]
        else:
            path = split_field(field_name)
            output = []
            flat_list._tuple1(data, path, 0, output)
            return output
    elif is_list(field_name):
        paths = [_select_a_field(f) for f in field_name]
        output = FlatList()
        _tuple((), unwrap(data), paths, 0, output)
        return output
    else:
        paths = [_select_a_field(field_name)]
        output = FlatList()
        _tuple((), data, paths, 0, output)
        return output

        def post(sql):
            # FIND OUT THE default DOMAIN SIZES
            result = self.db.column_query(sql)
            num_edges = len(edges)
            for e, edge in enumerate(edges):
                domain = edge.domain
                if domain.type == "default":
                    domain.type = "set"
                    parts = set(result[e])
                    domain.partitions = [{"index": i, "value": p} for i, p in enumerate(parts)]
                    domain.map = {p: i for i, p in enumerate(parts)}
                else:
                    Log.error("Do not know what to do here, yet")

            # FILL THE DATA CUBE
            maps = [(unwrap(e.domain.map), result[i]) for i, e in enumerate(edges)]
            cubes = FlatList()
            for c, s in enumerate(select):
                data = Matrix(*[len(e.domain.partitions) + (1 if e.allow_nulls else 0) for e in edges])
                for rownum, value in enumerate(result[c + num_edges]):
                    coord = [m[r[rownum]] for m, r in maps]
                    data[coord] = value
                cubes.append(data)

            if isinstance(query.select, list):
                return cubes
            else:
                return cubes[0]

    def __init__(self, name, params, cwd=None, env=None, debug=False, shell=False, bufsize=-1):
        self.name = name
        self.service_stopped = Signal("stopped signal for " + strings.quote(name))
        self.stdin = Queue("stdin for process " + strings.quote(name), silent=True)
        self.stdout = Queue("stdout for process " + strings.quote(name), silent=True)
        self.stderr = Queue("stderr for process " + strings.quote(name), silent=True)

        try:
            self.debug = debug or DEBUG
            self.service = service = subprocess.Popen(
                params,
                stdin=subprocess.PIPE,
                stdout=subprocess.PIPE,
                stderr=subprocess.PIPE,
                bufsize=bufsize,
                cwd=cwd if isinstance(cwd, (basestring, NullType, NoneType)) else cwd.abspath,
                env=unwrap(set_default(env, os.environ)),
                shell=shell
            )

            self.please_stop = Signal()
            self.please_stop.on_go(self._kill)
            self.thread_locker = Lock()
            self.children = [
                Thread.run(self.name + " stdin", self._writer, service.stdin, self.stdin, please_stop=self.service_stopped, parent_thread=self),
                Thread.run(self.name + " stdout", self._reader, "stdout", service.stdout, self.stdout, please_stop=self.service_stopped, parent_thread=self),
                Thread.run(self.name + " stderr", self._reader, "stderr", service.stderr, self.stderr, please_stop=self.service_stopped, parent_thread=self),
                Thread.run(self.name + " waiter", self._monitor, parent_thread=self),
            ]
        except Exception as e:
            Log.error("Can not call", e)

        if self.debug:
            Log.note("{{process}} START: {{command}}", process=self.name, command=" ".join(map(strings.quote, params)))

def dict2Multiset(dic):
    if dic == None:
        return None
    from mo_collections.multiset import Multiset
    output = Multiset()
    output.dic = unwrap(dic).copy()
    return output