Python mo_dots.split_field函数代码示例

    def select(self, selectList, fromPath, varName, sourceVar):
        path = split_field(fromPath)
        is_deep = len(path) > 1
        heads = []
        list = []
        for s in selectList:
            if is_deep:
                if s.value and is_variable_name(s.value):
                    shortForm = self._translate(s.value)
                    list.append("Value2Pipe(" + shortForm + ")\n")
                else:
                    Log.error("do not know how to handle yet")
            else:
                if s.value and is_variable_name(s.value):
                    list.append("Value2Pipe(getDocValue(" + value2MVEL(s.value) + "))\n")
                elif s.value:
                    shortForm = self._translate(s.value)
                    list.append("Value2Pipe(" + shortForm + ")\n")
                else:
                    code, decode = self.Parts2Term(s.domain)
                    heads.append(code.head)
                    list.append("Value2Pipe(" + code.body + ")\n")


        if len(split_field(fromPath)) > 1:
            output = 'if (' + varName + ' != "") ' + varName + '+="|";\n' + varName + '+=' + '+"|"+'.join(["Value2Pipe("+v+")\n" for v in list]) + ';\n'
        else:
            output = varName + ' = ' + '+"|"+'.join(["Value2Pipe("+v+")\n" for v in list]) + ';\n'

        return Data(
            head="".join(heads),
            body=output
        )

def get_pull(column):
    if column.nested_path[0] == ".":
        return concat_field("fields", literal_field(column.es_column))
    else:
        depth = len(split_field(column.nested_path[0]))
        rel_name = split_field(column.es_column)[depth:]
        return join_field(["_inner"] + rel_name)

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 untype_path(encoded):
    if encoded.startswith(".."):
        remainder = encoded.lstrip(".")
        back = len(encoded) - len(remainder) - 1
        return ("." * back) + join_field(decode_property(c) for c in split_field(remainder) if not c.startswith(TYPE_PREFIX))
    else:
        return join_field(decode_property(c) for c in split_field(encoded) if not c.startswith(TYPE_PREFIX))

def _select_a_field(field):
    if is_text(field):
        return wrap({"name": field, "value": split_field(field)})
    elif is_text(wrap(field).value):
        field = wrap(field)
        return wrap({"name": field.name, "value": split_field(field.value)})
    else:
        return wrap({"name": field.name, "value": field.value})

def _select_a_field(field):
    if isinstance(field, basestring):
        return wrap({"name": field, "value": split_field(field)})
    elif isinstance(wrap(field).value, basestring):
        field = wrap(field)
        return wrap({"name": field.name, "value": split_field(field.value)})
    else:
        return wrap({"name": field.name, "value": field.value})

    def select(self, fields):
        if isinstance(fields, Mapping):
            fields=fields.value

        if isinstance(fields, text_type):
            # RETURN LIST OF VALUES
            if len(split_field(fields)) == 1:
                if self.path[0] == fields:
                    return [d[1] for d in self.data]
                else:
                    return [d[0][fields] for d in self.data]
            else:
                keys = split_field(fields)
                depth = coalesce(MIN([i for i, (k, p) in enumerate(zip(keys, self.path)) if k != p]), len(self.path))  # LENGTH OF COMMON PREFIX
                short_key = keys[depth:]

                output = FlatList()
                _select1((wrap(d[depth]) for d in self.data), short_key, 0, output)
                return output

        if isinstance(fields, list):
            output = FlatList()

            meta = []
            for f in fields:
                if hasattr(f.value, "__call__"):
                    meta.append((f.name, f.value))
                else:
                    meta.append((f.name, functools.partial(lambda v, d: d[v], f.value)))

            for row in self._values():
                agg = Data()
                for name, f in meta:
                    agg[name] = f(row)

                output.append(agg)

            return output

            # meta = []
            # for f in fields:
            #     keys = split_field(f.value)
            #     depth = coalesce(MIN([i for i, (k, p) in enumerate(zip(keys, self.path)) if k != p]), len(self.path))  # LENGTH OF COMMON PREFIX
            #     short_key = join_field(keys[depth:])
            #
            #     meta.append((f.name, depth, short_key))
            #
            # for row in self._data:
            #     agg = Data()
            #     for name, depth, short_key in meta:
            #         if short_key:
            #             agg[name] = row[depth][short_key]
            #         else:
            #             agg[name] = row[depth]
            #     output.append(agg)
            # return output

        Log.error("multiselect over FlatList not supported")

                def defParent(name):
                    # DO NOT MAKE THE SAME PARENT TWICE
                    if name in parentVarNames:
                        return
                    parentVarNames.add(name)

                    if len(split_field(name)) == 1:
                        contextVariables.append("Map " + name + " = new HashMap();\n")
                    else:
                        defParent(join_field(split_field(name)[0:-1]))
                        contextVariables.append(name + " = new HashMap();\n")

    def getFrameVariables(self, body):
        contextVariables = []
        columns = self.fromData.columns

        parentVarNames = set()    # ALL PARENTS OF VARIABLES WITH "." IN NAME
        body = body.replace(".?", ".")

        for i, c in enumerate(columns):
            j = body.find(c.name, 0)
            while j >= 0:
                s = j
                j = body.find(c.name, s + 1)

                test0 = body[s - 1: s + len(c.name) + 1:]
                test3 = body[s - 8: s + len(c.name):]

                if test0[:-1] == "\"" + c.name:
                    continue
                if test3 == "_source." + c.name:
                    continue

                def defParent(name):
                    # DO NOT MAKE THE SAME PARENT TWICE
                    if name in parentVarNames:
                        return
                    parentVarNames.add(name)

                    if len(split_field(name)) == 1:
                        contextVariables.append("Map " + name + " = new HashMap();\n")
                    else:
                        defParent(join_field(split_field(name)[0:-1]))
                        contextVariables.append(name + " = new HashMap();\n")

                body = body.replace(c.name, "-"*len(c.name))

                if self.isLean or c.useSource:
                    if len(split_field(c.name)) > 1:
                        defParent(join_field(split_field(c.name)[0:-1]))
                        contextVariables.append(c.name + " = getSourceValue(\"" + c.name + "\");\n")
                    else:
                        contextVariables.append(c.name + " = _source[\"" + c.name + "\"];\n")
                else:
                    if len(split_field(c.name)) > 1:
                        defParent(join_field(split_field(c.name)[0:-1]))
                        contextVariables.append(c.name + " = getDocValue(\"" + c.name + "\");\n")
                    else:
                        contextVariables.append(c.name + " = getDocValue(\"" + c.name + "\");\n")
                break

        return "".join(contextVariables)

def to_python(self, not_null=False, boolean=False, many=False):
    path = split_field(self.var)
    agg = "row"
    if not path:
        return agg
    elif path[0] in ["row", "rownum"]:
        # MAGIC VARIABLES
        agg = path[0]
        path = path[1:]
        if len(path) == 0:
            return agg
    elif path[0] == "rows":
        if len(path) == 1:
            return "rows"
        elif path[1] in ["first", "last"]:
            agg = "rows." + path[1] + "()"
            path = path[2:]
        else:
            Log.error("do not know what {{var}} of `rows` is", var=path[1])

    for p in path[:-1]:
        if not_null:
            agg = agg + ".get(" + convert.value2quote(p) + ")"
        else:
            agg = agg + ".get(" + convert.value2quote(p) + ", EMPTY_DICT)"
    output = agg + ".get(" + convert.value2quote(path[-1]) + ")"
    if many:
        output = "listwrap(" + output + ")"
    return output

    def not_monitor(self, please_stop):
        Log.alert("metadata scan has been disabled")
        please_stop.on_go(lambda: self.todo.add(THREAD_STOP))
        while not please_stop:
            column = self.todo.pop()
            if column == THREAD_STOP:
                break

            if column.jx_type in STRUCT or split_field(column.es_column)[-1] == EXISTS_TYPE:
                DEBUG and Log.note("{{column.es_column}} is a struct", column=column)
                column.last_updated = Date.now()
                continue
            elif column.last_updated > Date.now() - TOO_OLD and column.cardinality is not None:
                # DO NOT UPDATE FRESH COLUMN METADATA
                DEBUG and Log.note("{{column.es_column}} is still fresh ({{ago}} ago)", column=column, ago=(Date.now()-Date(column.last_updated)).seconds)
                continue

            with Timer("Update {{col.es_index}}.{{col.es_column}}", param={"col": column}, silent=not DEBUG, too_long=0.05):
                if untype_path(column.name) in ["build.type", "run.type"]:
                    try:
                        self._update_cardinality(column)
                    except Exception as e:
                        Log.warning("problem getting cardinality for {{column.name}}", column=column, cause=e)
                else:
                    column.last_updated = Date.now()

def unnest_path(encoded):
    if encoded.startswith(".."):
        encoded = encoded.lstrip(".")
        if not encoded:
            encoded = "."

    return join_field(decode_property(c) for c in split_field(encoded) if c != NESTED_TYPE)

def setValues(expression, constants):
    if not constants:
        return expression

    constants = constants.copy()

    # EXPAND ALL CONSTANTS TO PRIMITIVE VALUES (MVEL CAN ONLY ACCEPT PRIMITIVE VALUES)
    for c in constants:
        value = c.value
        n = c.name
        if len(split_field(n)) >= 3:
            continue    # DO NOT GO TOO DEEP
        if isinstance(value, list):
            continue  # DO NOT MESS WITH ARRAYS

        if isinstance(value, Mapping):
            for k, v in value.items():
                constants.append({"name": n + "." + k, "value": v})

    for c in reverse(constants):# REVERSE ORDER, SO LONGER NAMES ARE TESTED FIRST
        s = 0
        while True:
            s = expression.find(c.name, s)
            if s == -1:
                break
            if re.match(r"\w", expression[s - 1]):
                break
            if re.match(r"\w", expression[s + len(c.name)]):
                break

            v = value2MVEL(c.value)
            expression = expression[:s:] + "" + v + expression[:s + len(c.name):]

    return expression

    def parse_field(fieldname, data, depth):
        """
        RETURN (first, rest) OF fieldname
        """
        col = split_field(fieldname)
        d = data
        for i, c in enumerate(col):
            try:
                d = d[c]
            except Exception as e:
                Log.error("{{name}} does not exist", name=fieldname)
            if is_list(d) and len(col) > 1:
                if len(primary_column) <= depth + i:
                    primary_nested.append(True)
                    primary_column.append(c)
                    primary_branch.append(d)
                elif primary_nested[depth] and primary_column[depth + i] != c:
                    Log.error("only one branch of tree allowed")
                else:
                    primary_nested[depth + i] = True
                    primary_column[depth + i] = c
                    primary_branch[depth + i] = d

                return c, join_field(col[i + 1 :])
            else:
                if len(primary_column) <= depth + i:
                    primary_nested.append(False)
                    primary_column.append(c)
                    primary_branch.append([d])
        return fieldname, None

    def get_column(self, name):
        sub_schema = self.schema

        while '.' in sub_schema.more:
            sub_schema = sub_schema.more.get('.')

        for n in split_field(name):
            m = sub_schema.more
            if n in m:
                sub_schema = m.get(n)
            else:
                Log.error("{{name}} not found in schema", name=name)

            while '.' in sub_schema.more:
                sub_schema = sub_schema.more.get('.')

        return Column(
            name,
            self.values[name],
            self.reps[name],
            self.defs[name],
            self.num_rows,
            sub_schema,
            self.schema.max_repetition_level(name),
            self.schema.max_definition_level(name)
        )