From d9d1ff699da33d2fb86c80784919a8ffce30b153 Mon Sep 17 00:00:00 2001 From: "P. Sai Vinay" Date: Thu, 3 Feb 2022 20:48:01 +0530 Subject: [PATCH] Support Scripted aggregations and Add type hints to field_mappings.py and refactor/improve capability_matrix usage --- eland/dataframe.py | 10 +- eland/field_mappings.py | 303 ++++---- eland/operations.py | 26 +- eland/query.py | 49 +- eland/query_compiler.py | 39 +- eland/series.py | 10 +- eland/tasks.py | 4 +- noxfile.py | 1 + tests/notebook/test_demo_notebook.ipynb | 968 ++---------------------- tests/series/test_arithmetics_pytest.py | 144 ++-- 10 files changed, 392 insertions(+), 1162 deletions(-) diff --git a/eland/dataframe.py b/eland/dataframe.py index 86f3cc98..9c5a85b4 100644 --- a/eland/dataframe.py +++ b/eland/dataframe.py @@ -617,11 +617,11 @@ def es_info(self): is_source_field: False Mappings: capabilities: - es_field_name is_source es_dtype es_date_format pd_dtype is_searchable is_aggregatable is_scripted aggregatable_es_field_name - timestamp timestamp True date strict_date_hour_minute_second datetime64[ns] True True False timestamp - OriginAirportID OriginAirportID True keyword None object True True False OriginAirportID - DestAirportID DestAirportID True keyword None object True True False DestAirportID - FlightDelayMin FlightDelayMin True integer None int64 True True False FlightDelayMin + display_name es_field_name is_source es_dtype es_date_format pd_dtype is_searchable is_aggregatable is_scripted aggregatable_es_field_name + 0 timestamp timestamp True date strict_date_hour_minute_second datetime64[ns] True True False timestamp + 1 OriginAirportID OriginAirportID True keyword None object True True False OriginAirportID + 2 DestAirportID DestAirportID True keyword None object True True False DestAirportID + 3 FlightDelayMin FlightDelayMin True integer None int64 True True False FlightDelayMin Operations: tasks: [('boolean_filter': ('boolean_filter': {'bool': {'must': [{'term': {'OriginAirportID': 'AMS'}}, {'range': {'FlightDelayMin': {'gt': 60}}}]}})), ('tail': ('sort_field': '_doc', 'count': 5))] size: 5 diff --git a/eland/field_mappings.py b/eland/field_mappings.py index 92ac7f72..919fe6ab 100644 --- a/eland/field_mappings.py +++ b/eland/field_mappings.py @@ -24,6 +24,7 @@ Mapping, NamedTuple, Optional, + Sequence, Set, TextIO, Tuple, @@ -31,8 +32,8 @@ ) import numpy as np -import pandas as pd # type: ignore -from pandas.core.dtypes.common import ( # type: ignore +import pandas as pd # type: ignore[import] +from pandas.core.dtypes.common import ( # type: ignore[import] is_bool_dtype, is_datetime64_any_dtype, is_datetime_or_timedelta_dtype, @@ -40,7 +41,7 @@ is_integer_dtype, is_string_dtype, ) -from pandas.core.dtypes.inference import is_list_like +from pandas.core.dtypes.inference import is_list_like # type: ignore[import] if TYPE_CHECKING: from elasticsearch import Elasticsearch @@ -68,6 +69,7 @@ class Field(NamedTuple): """Holds all information on a particular field in the mapping""" column: str + display_name: str es_field_name: str is_source: bool es_dtype: str @@ -80,21 +82,23 @@ class Field(NamedTuple): @property def is_numeric(self) -> bool: - return is_integer_dtype(self.pd_dtype) or is_float_dtype(self.pd_dtype) + return is_integer_dtype(self.pd_dtype) or is_float_dtype(self.pd_dtype) # type: ignore[no-any-return] @property def is_timestamp(self) -> bool: - return is_datetime_or_timedelta_dtype(self.pd_dtype) + return is_datetime_or_timedelta_dtype(self.pd_dtype) # type: ignore[no-any-return] @property def is_bool(self) -> bool: - return is_bool_dtype(self.pd_dtype) + return is_bool_dtype(self.pd_dtype) # type: ignore[no-any-return] @property - def np_dtype(self): + def np_dtype(self) -> Any: return np.dtype(self.pd_dtype) - def is_es_agg_compatible(self, es_agg) -> bool: + def is_es_agg_compatible( + self, es_agg: Union[Tuple[str, List[float]], str, Sequence[object]] + ) -> bool: # Unpack the actual aggregation if this is 'extended_stats/percentiles' if isinstance(es_agg, tuple): if es_agg[0] == "extended_stats": @@ -221,8 +225,10 @@ def __init__( ) # Populate capability matrix of fields - self._mappings_capabilities = FieldMappings._create_capability_matrix( - all_fields, source_fields, all_fields_caps + self._mappings_capabilities: pd.DataFrame = ( + FieldMappings._create_capability_matrix( + all_fields, source_fields, all_fields_caps + ) ) if display_names is not None: @@ -231,7 +237,7 @@ def __init__( @staticmethod def _extract_fields_from_mapping( mappings: Dict[str, Any], source_only: bool = False - ) -> Dict[str, str]: + ) -> Dict[str, Tuple[str, Any]]: """ Extract all field names and types from a mapping. ``` @@ -294,15 +300,15 @@ def _extract_fields_from_mapping( Returns ------- - fields, dates_format: tuple(dict, dict) + fields, dates_format: Dict[str, Tuple[str, Any]] where: fields: dict of field names and types dates_format: Dict of date field names and format """ - fields = {} + fields: Dict[str, Tuple[str, Any]] = {} # Recurse until we get a 'type: xxx' - def flatten(x, name=""): + def flatten(x: Union[str, Dict[str, str]], name: str = "") -> None: if isinstance(x, dict): for a in x: if a == "type" and isinstance( @@ -352,7 +358,11 @@ def flatten(x, name=""): return fields @staticmethod - def _create_capability_matrix(all_fields, source_fields, all_fields_caps): + def _create_capability_matrix( + all_fields: Dict[str, Tuple[str, Any]], + source_fields: Dict[str, Tuple[str, Any]], + all_fields_caps: Dict[str, Dict[str, Any]], + ) -> pd.DataFrame: """ { "fields": { @@ -380,56 +390,56 @@ def _create_capability_matrix(all_fields, source_fields, all_fields_caps): } } """ - all_fields_caps_fields = all_fields_caps["fields"] + # Filter the required fields + all_fields_caps_fields = { + key: value + for key, value in all_fields_caps["fields"].items() + if key in all_fields + } - capability_matrix = {} + capability_matrix: List[pd.Series] = [] for field, field_caps in all_fields_caps_fields.items(): - if field in all_fields: - # v = {'long': {'type': 'long', 'searchable': True, 'aggregatable': True}} - for kk, vv in field_caps.items(): - _source = field in source_fields - es_field_name = field - es_dtype = vv["type"] - es_date_format = all_fields[field][1] - pd_dtype = FieldMappings._es_dtype_to_pd_dtype(vv["type"]) - is_searchable = vv["searchable"] - is_aggregatable = vv["aggregatable"] - scripted = False - aggregatable_es_field_name = None # this is populated later - - caps = [ - es_field_name, - _source, - es_dtype, - es_date_format, - pd_dtype, - is_searchable, - is_aggregatable, - scripted, - aggregatable_es_field_name, - ] - - capability_matrix[field] = caps - - if "non_aggregatable_indices" in vv: - warnings.warn( - f"Field {field} has conflicting aggregatable fields across indexes " - f"{str(vv['non_aggregatable_indices'])}", - UserWarning, - ) - if "non_searchable_indices" in vv: - warnings.warn( - f"Field {field} has conflicting searchable fields across indexes " - f"{str(vv['non_searchable_indices'])}", - UserWarning, - ) - - capability_matrix_df = pd.DataFrame.from_dict( - capability_matrix, orient="index", columns=FieldMappings.column_labels + # v = {'long': {'type': 'long', 'searchable': True, 'aggregatable': True}} + for _, vv in field_caps.items(): + capability_row = pd.Series( + { + "display_name": field, + "es_field_name": field, # field name in ES + "is_source": field in source_fields, + "es_dtype": vv["type"], + "es_date_format": all_fields[field][1], + "pd_dtype": FieldMappings._es_dtype_to_pd_dtype(vv["type"]), + "is_searchable": vv["searchable"], + "is_aggregatable": vv["aggregatable"], + "is_scripted": False, + "aggregatable_es_field_name": None, # this is populated later + }, + ) + + if "non_aggregatable_indices" in vv: + warnings.warn( + f"Field {field} has conflicting aggregatable fields across indexes " + f"{str(vv['non_aggregatable_indices'])}", + UserWarning, + ) + if "non_searchable_indices" in vv: + warnings.warn( + f"Field {field} has conflicting searchable fields across indexes " + f"{str(vv['non_searchable_indices'])}", + UserWarning, + ) + + capability_matrix.append(capability_row) + + # concatenating List[pd.Series] is efficient than appending each one + capability_matrix_df: pd.DataFrame = ( + pd.concat(capability_matrix, axis=1) + .transpose() + .sort_values("display_name", ignore_index=True) ) - def find_aggregatable(row, df): + def find_aggregatable(row: pd.Series, df: pd.DataFrame) -> pd.Series: # convert series to dict so we can add 'aggregatable_es_field_name' row_as_dict = row.to_dict() if not row_as_dict["is_aggregatable"]: @@ -456,10 +466,10 @@ def find_aggregatable(row, df): ) # return just source fields (as these are the only ones we display) - return capability_matrix_df[capability_matrix_df.is_source].sort_index() + return capability_matrix_df[capability_matrix_df.is_source] @classmethod - def _es_dtype_to_pd_dtype(cls, es_dtype): + def _es_dtype_to_pd_dtype(cls, es_dtype: str) -> str: """ Mapping Elasticsearch types to pandas dtypes -------------------------------------------- @@ -477,7 +487,7 @@ def _es_dtype_to_pd_dtype(cls, es_dtype): return cls.ES_DTYPE_TO_PD_DTYPE.get(es_dtype, "object") @staticmethod - def _pd_dtype_to_es_dtype(pd_dtype) -> Optional[str]: + def _pd_dtype_to_es_dtype(pd_dtype: str) -> Optional[str]: """ Mapping pandas dtypes to Elasticsearch dtype -------------------------------------------- @@ -533,7 +543,7 @@ def _generate_es_mappings( ------- mapping : str """ - es_dtype: Union[str, Dict[str, Any]] + es_dtype: Union[Optional[str], Dict[str, Any]] mapping_props: Dict[str, Any] = {} @@ -587,16 +597,16 @@ def aggregatable_field_name(self, display_name: str) -> Optional[str]: raise KeyError if the field_name doesn't exist in the mapping, or isn't aggregatable """ - mapping: Optional[pd.Series] = None + mapping: pd.Series = self._mappings_capabilities.loc[ + self._mappings_capabilities["display_name"] == display_name + ].squeeze() - try: - mapping = self._mappings_capabilities.loc[display_name] - except KeyError: + if mapping.empty: raise KeyError( f"Can not get aggregatable field name for invalid display name {display_name}" - ) from None + ) - if mapping is not None and mapping.aggregatable_es_field_name is None: + if mapping.aggregatable_es_field_name is None: warnings.warn(f"Aggregations not supported for '{display_name}'") return mapping.aggregatable_es_field_name @@ -649,7 +659,7 @@ def date_field_format(self, es_field_name: str) -> str: str A string (for date fields) containing the date format for the field """ - return self._mappings_capabilities.loc[ + return self._mappings_capabilities.loc[ # type: ignore[no-any-return] self._mappings_capabilities.es_field_name == es_field_name ].es_date_format.squeeze() @@ -669,46 +679,47 @@ def field_name_pd_dtype(self, es_field_name: str) -> str: KeyError If es_field_name does not exist in mapping """ - if es_field_name not in self._mappings_capabilities.es_field_name: + capability_row: pd.Series = self._mappings_capabilities.loc[ + self._mappings_capabilities["es_field_name"] == es_field_name + ].squeeze() + + if capability_row.empty or capability_row["is_scripted"] == True: raise KeyError(f"es_field_name {es_field_name} does not exist") - pd_dtype = self._mappings_capabilities.loc[ - self._mappings_capabilities.es_field_name == es_field_name - ].pd_dtype.squeeze() - return pd_dtype + return capability_row["pd_dtype"] # type: ignore[no-any-return] def add_scripted_field( self, scripted_field_name: str, display_name: str, pd_dtype: str ) -> None: # if this display name is used somewhere else, drop it - if display_name in self._mappings_capabilities.index: - self._mappings_capabilities = self._mappings_capabilities.drop( - index=[display_name] - ) - - # ['es_field_name', 'is_source', 'es_dtype', 'es_date_format', 'pd_dtype', 'is_searchable', - # 'is_aggregatable', 'is_scripted', 'aggregatable_es_field_name'] - - capabilities = { - display_name: [ - scripted_field_name, - False, - self._pd_dtype_to_es_dtype(pd_dtype), - None, - pd_dtype, - True, - True, - True, - scripted_field_name, - ] - } - - capability_matrix_row = pd.DataFrame.from_dict( - capabilities, orient="index", columns=FieldMappings.column_labels + try: + # display_name can be None + index = self._mappings_capabilities[ + (self._mappings_capabilities.display_name == display_name) + | (self._mappings_capabilities.display_name.isna()) + ].index + if not index.empty: + self._mappings_capabilities.drop(labels=index, inplace=True) + except KeyError: + pass + + scripted_field_mapping: pd.Series = pd.Series( + { + "display_name": display_name, + "es_field_name": scripted_field_name, + "is_source": False, + "es_dtype": self._pd_dtype_to_es_dtype(pd_dtype), + "es_date_format": None, + "pd_dtype": pd_dtype, + "is_searchable": True, + "is_aggregatable": True, + "is_scripted": True, + "aggregatable_es_field_name": scripted_field_name, + }, ) self._mappings_capabilities = self._mappings_capabilities.append( - capability_matrix_row + scripted_field_mapping, ignore_index=True ) def numeric_source_fields(self) -> List[str]: @@ -725,9 +736,9 @@ def all_source_fields(self) -> List[Field]: """ source_fields: List[Field] = [] - for column, row in self._mappings_capabilities.iterrows(): - row = row.to_dict() - row["column"] = column + for row in self._mappings_capabilities.itertuples(index=False): + row = row._asdict() + row["column"] = row["display_name"] source_fields.append(Field(**row)) return source_fields @@ -747,8 +758,9 @@ def groupby_source_fields(self, by: List[str]) -> Tuple[List[Field], List[Field] """ groupby_fields: Dict[str, Field] = {} aggregatable_fields: List[Field] = [] - for column, row in self._mappings_capabilities.iterrows(): - row = row.to_dict() + for row in self._mappings_capabilities.itertuples(index=False): + row = row._asdict() + column = row["display_name"] row["column"] = column if column not in by: aggregatable_fields.append(Field(**row)) @@ -795,31 +807,36 @@ def metric_source_fields( es_date_formats.append(es_date_format) # return in display_name order - return pd_dtypes, es_field_names, es_date_formats + return pd_dtypes, es_field_names, es_date_formats # type: ignore def get_field_names(self, include_scripted_fields: bool = True) -> List[str]: if include_scripted_fields: - return self._mappings_capabilities.es_field_name.to_list() + return self._mappings_capabilities.es_field_name.to_list() # type: ignore[no-any-return] - return self._mappings_capabilities[ # noqa: E712 + return self._mappings_capabilities[ # type: ignore[no-any-return] self._mappings_capabilities.is_scripted == False ].es_field_name.to_list() - def _get_display_names(self): - return self._mappings_capabilities.index.to_list() + def _get_display_names(self) -> List[str]: + return self._mappings_capabilities.display_name.to_list() # type: ignore[no-any-return] - def _set_display_names(self, display_names: List[str]): + def _set_display_names(self, display_names: List[str]) -> None: if not is_list_like(display_names): raise ValueError(f"'{display_names}' is not list like") if list(set(display_names) - set(self.display_names)): raise KeyError(f"{display_names} not in display names {self.display_names}") - self._mappings_capabilities = self._mappings_capabilities.reindex(display_names) + # Now filter and maintain order by display_names + self._mappings_capabilities = self._mappings_capabilities.iloc[ + pd.Index(self._mappings_capabilities["display_name"]).get_indexer( + display_names + ) + ].reset_index(drop=True) display_names = property(_get_display_names, _set_display_names) - def dtypes(self): + def dtypes(self) -> pd.Series: """ Returns ------- @@ -827,15 +844,21 @@ def dtypes(self): Index: Display name Values: pd_dtype as np.dtype """ - pd_dtypes = self._mappings_capabilities["pd_dtype"] + pd_dtypes = self._mappings_capabilities[["display_name", "pd_dtype"]].set_index( + "display_name" + ) + + if isinstance(pd_dtypes, pd.DataFrame): + pd_dtypes = pd_dtypes.squeeze(axis=1) # Set name of the returned series as None pd_dtypes.name = None + pd_dtypes.index.name = None # Convert return from 'str' to 'np.dtype' return pd_dtypes.apply(lambda x: np.dtype(x)) - def es_dtypes(self): + def es_dtypes(self) -> pd.Series: """ Returns ------- @@ -843,17 +866,23 @@ def es_dtypes(self): Index: Display name Values: es_dtype as a string """ - es_dtypes = self._mappings_capabilities["es_dtype"] + es_dtypes = self._mappings_capabilities[["display_name", "es_dtype"]].set_index( + "display_name" + ) + + if isinstance(es_dtypes, pd.DataFrame): + es_dtypes = es_dtypes.squeeze(axis=1) # Set name of the returned series as None es_dtypes.name = None + es_dtypes.index.name = None return es_dtypes def es_info(self, buf: TextIO) -> None: buf.write("Mappings:\n") buf.write(f" capabilities:\n{self._mappings_capabilities.to_string()}\n") - def rename(self, old_name_new_name_dict): + def rename(self, old_name_new_name_dict: Dict[str, str]) -> None: """ Renames display names in-place @@ -869,32 +898,34 @@ def rename(self, old_name_new_name_dict): ----- For the names that do not exist this is a no op """ - self._mappings_capabilities = self._mappings_capabilities.rename( - index=old_name_new_name_dict + self._mappings_capabilities["display_name"].replace( + old_name_new_name_dict, inplace=True ) - def get_renames(self): - # return dict of renames { old_name: new_name, ... } (inefficient) - renames = {} - - for display_name in self.display_names: - field_name = self._mappings_capabilities.loc[display_name].es_field_name - if field_name != display_name: - renames[field_name] = display_name - - return renames + def get_renames(self) -> Dict[str, str]: + """ + This method returns the differences between `display_name` and `es_field_name` + by querying directly + """ + renames_df: pd.DataFrame = self._mappings_capabilities[ + self._mappings_capabilities.apply( + lambda x: x["display_name"] != x["es_field_name"], axis=1 + ) + ][["display_name", "es_field_name"]] + # {'products.manufacturer': 'manufacturer', 'products.base_unit_price': base_unit_price} + return renames_df.set_index("es_field_name")["display_name"].to_dict() # type: ignore[no-any-return] def verify_mapping_compatibility( - ed_mapping: Mapping[str, Mapping[str, Mapping[str, Mapping[str, str]]]], - es_mapping: Mapping[str, Mapping[str, Mapping[str, Mapping[str, str]]]], + ed_mapping: Mapping[str, Mapping[str, Mapping[str, Any]]], + es_mapping: Mapping[str, Mapping[str, Mapping[str, Any]]], es_type_overrides: Optional[Mapping[str, str]] = None, ) -> None: """Given a mapping generated by Eland and an existing ES index mapping attempt to see if the two are compatible. If not compatible raise ValueError with a list of problems between the two to be reported to the user. """ - problems = [] + problems: List[str] = [] es_type_overrides = es_type_overrides or {} ed_props = ed_mapping["mappings"]["properties"] @@ -909,10 +940,10 @@ def verify_mapping_compatibility( problems.append(f"- {key!r} is missing from ES index mapping") continue - key_type = es_type_overrides.get(key, key_def["type"]) - es_key_type = es_props[key]["type"] + key_type: Any = es_type_overrides.get(key, key_def["type"]) + es_key_type: Any = es_mapping[key]["type"] if key_type != es_key_type and es_key_type not in ES_COMPATIBLE_TYPES.get( - key_type, () + key_type, set() ): problems.append( f"- {key!r} column type ({key_type!r}) not compatible with " diff --git a/eland/operations.py b/eland/operations.py index 15c0174d..7d096e93 100644 --- a/eland/operations.py +++ b/eland/operations.py @@ -166,7 +166,9 @@ def count(self, query_compiler: "QueryCompiler") -> pd.Series: counts = {} for field in fields: - body = Query(query_params.query) + body = Query( + query=query_params.query, script_fields=query_params.script_fields + ) body.exists(field, must=True) field_exists_count = query_compiler._client.count( @@ -225,7 +227,7 @@ def idx( # Consider only Numeric fields fields = [field for field in fields if (field.is_numeric)] - body = Query(query_params.query) + body = Query(query=query_params.query, script_fields=query_params.script_fields) for field in fields: body.top_hits_agg( @@ -358,7 +360,7 @@ def _metric_aggs( # Consider if field is Int/Float/Bool fields = [field for field in fields if (field.is_numeric or field.is_bool)] - body = Query(query_params.query) + body = Query(query=query_params.query, script_fields=query_params.script_fields) # Convert pandas aggs to ES equivalent es_aggs = self._map_pd_aggs_to_es_aggs(pd_aggs, percentiles) @@ -447,7 +449,7 @@ def _terms_aggs( # Get just aggregatable field_names aggregatable_field_names = query_compiler._mappings.aggregatable_field_names() - body = Query(query_params.query) + body = Query(query=query_params.query, script_fields=query_params.script_fields) for field in aggregatable_field_names.keys(): body.terms_aggs(field, func, field, es_size=es_size) @@ -486,7 +488,7 @@ def _hist_aggs( numeric_source_fields = query_compiler._mappings.numeric_source_fields() - body = Query(query_params.query) + body = Query(query=query_params.query, script_fields=query_params.script_fields) results = self._metric_aggs(query_compiler, ["min", "max"], numeric_only=True) min_aggs = {} @@ -853,7 +855,7 @@ def aggs_groupby( field for field in agg_fields if (field.is_numeric or field.is_bool) ] - body = Query(query_params.query) + body = Query(query=query_params.query, script_fields=query_params.script_fields) # To return for creating multi-index on columns headers = [agg_field.column for agg_field in agg_fields] @@ -1237,7 +1239,9 @@ def search_yield_pandas_dataframes( ) script_fields = query_params.script_fields - query = Query(query_params.query) + query = Query( + query=query_params.query, script_fields=query_params.script_fields + ) body = query.to_search_body() if script_fields is not None: @@ -1268,7 +1272,7 @@ def index_count(self, query_compiler: "QueryCompiler", field: str) -> int: # TODO - this is not necessarily valid as the field may not exist in ALL these docs return size - body = Query(query_params.query) + body = Query(query=query_params.query, script_fields=query_params.script_fields) body.exists(field, must=True) count: int = query_compiler._client.count( @@ -1302,7 +1306,7 @@ def index_matches_count( query_compiler, items ) - body = Query(query_params.query) + body = Query(query=query_params.query, script_fields=query_params.script_fields) if field == Index.ID_INDEX_FIELD: body.ids(items, must=True) @@ -1427,7 +1431,9 @@ def es_info(self, query_compiler: "QueryCompiler", buf: TextIO) -> None: _source = query_compiler._mappings.get_field_names() script_fields = query_params.script_fields - query = Query(query_params.query) + query = Query( + query=query_params.query, script_fields=query_params.script_fields + ) body = query.to_search_body() if script_fields is not None: body["script_fields"] = script_fields diff --git a/eland/query.py b/eland/query.py index be20d1f2..e2df100d 100644 --- a/eland/query.py +++ b/eland/query.py @@ -27,16 +27,17 @@ class Query: Simple class to manage building Elasticsearch queries. """ - def __init__(self, query: Optional["Query"] = None): - # type defs - self._query: BooleanFilter - self._aggs: Dict[str, Any] - self._composite_aggs: Dict[str, Any] + def __init__( + self, + query: Optional["Query"] = None, + script_fields: Optional[Dict[str, Dict[str, Any]]] = None, + ): + self._script_fields = script_fields or {} if query is None: - self._query = BooleanFilter() - self._aggs = {} - self._composite_aggs = {} + self._query: BooleanFilter = BooleanFilter() + self._aggs: Dict[str, Any] = {} + self._composite_aggs: Dict[str, Any] = {} else: # Deep copy the incoming query so we can change it self._query = deepcopy(query._query) @@ -122,7 +123,12 @@ def terms_aggs( } } """ - agg = {func: {"field": field}} + if field in self._script_fields: + _field = self._script_fields[field] + else: + _field = {"field": field} + agg = {func: _field} + if es_size: agg[func]["size"] = str(es_size) @@ -142,8 +148,11 @@ def metric_aggs(self, name: str, func: str, field: str) -> None: } } """ - agg = {func: {"field": field}} - self._aggs[name] = agg + if field in self._script_fields: + _field = self._script_fields[field] + else: + _field = {"field": field} + self._aggs[name] = {func: _field} def percentile_agg(self, name: str, field: str, percents: List[float]) -> None: """ @@ -160,7 +169,11 @@ def percentile_agg(self, name: str, field: str, percents: List[float]) -> None: } """ - agg = {"percentiles": {"field": field, "percents": percents}} + if field in self._script_fields: + _field = self._script_fields[field] + else: + _field = {"field": field} + agg = {"percentiles": {**_field, "percents": percents}} self._aggs[name] = agg def top_hits_agg( @@ -191,7 +204,11 @@ def composite_agg_bucket_terms(self, name: str, field: str) -> None: } } """ - self._composite_aggs[name] = {"terms": {"field": field}} + if field in self._script_fields: + _field = self._script_fields[field] + else: + _field = {"field": field} + self._composite_aggs[name] = {"terms": _field} def composite_agg_bucket_date_histogram( self, @@ -204,7 +221,11 @@ def composite_agg_bucket_date_histogram( raise ValueError( "calendar_interval and fixed_interval parmaeters are mutually exclusive" ) - agg = {"field": field} + if field in self._script_fields: + _field = self._script_fields[field] + else: + _field = {"field": field} + agg: Dict[str, Any] = {"field": _field} if calendar_interval is not None: agg["calendar_interval"] = calendar_interval elif fixed_interval is not None: diff --git a/eland/query_compiler.py b/eland/query_compiler.py index 8c8b5622..82b5a26f 100644 --- a/eland/query_compiler.py +++ b/eland/query_compiler.py @@ -247,10 +247,7 @@ def _es_results_to_pandas( i = 0 for i, hit in enumerate(results, 1): - if "_source" in hit: - row = hit["_source"] - else: - row = {} + row = hit["_source"] if "_source" in hit else {} # script_fields appear in 'fields' if "fields" in hit: @@ -291,7 +288,9 @@ def _es_results_to_pandas( return df - def _flatten_dict(self, y, field_mapping_cache: "FieldMappingCache"): + def _flatten_dict( + self, y: Dict[str, Any], field_mapping_cache: "FieldMappingCache" + ): out = {} def flatten(x, name=""): @@ -519,16 +518,24 @@ def search_yield_pandas_dataframes(self) -> Generator["pd.DataFrame", None, None return self._operations.search_yield_pandas_dataframes(self) # __getitem__ methods - def getitem_column_array(self, key, numeric=False): - """Get column data for target labels. - - Args: - key: Target labels by which to retrieve data. - numeric: A boolean representing whether or not the key passed in represents + def getitem_column_array( + self, key: Union[str, List[str]], numeric: bool = False + ) -> "QueryCompiler": + """ + Get column data for target labels. + + Paremeters + ----------- + key: Union[str, List[str]] + A list of target labels by which to retrieve data. + numeric: bool + A boolean representing whether or not the key passed in represents the numeric index or the named index. - Returns: - A new QueryCompiler. + Returns + -------- + A new QueryCompiler. + """ result = self.copy() @@ -751,12 +758,14 @@ def check_arithmetics(self, right: "QueryCompiler") -> None: ) def arithmetic_op_fields( - self, display_name: str, arithmetic_object: "ArithmeticSeries" + self, + display_name: str, + arithmetic_object: "ArithmeticSeries", ) -> "QueryCompiler": result = self.copy() # create a new field name for this display name - scripted_field_name = f"script_field_{display_name}" + scripted_field_name: str = f"script_field_{display_name}" # add scripted field result._mappings.add_scripted_field( diff --git a/eland/series.py b/eland/series.py index 2d7f6b47..da7f5bb0 100644 --- a/eland/series.py +++ b/eland/series.py @@ -1382,7 +1382,10 @@ def _numeric_op(self, right: Any, method_name: str) -> "Series": Naming is consistent for rops """ - # print("_numeric_op", self, right, method_name) + # We are assigning a display name internally + # TODO: Override this if new name is given explicitly + # Example: ed_df['new_field'] = ed_df['DestCountry'] + ed_df['OriginCountry'] + display_name: Optional[str] = self.name if isinstance(right, Series): # Check we can the 2 Series are compatible (raises on error): self._query_compiler.check_arithmetics(right._query_compiler) @@ -1393,10 +1396,8 @@ def _numeric_op(self, right: Any, method_name: str) -> "Series": display_name = None elif np.issubdtype(np.dtype(type(right)), np.number): right_object = ArithmeticNumber(right, np.dtype(type(right))) - display_name = self.name elif isinstance(right, str): right_object = ArithmeticString(right) - display_name = self.name else: raise TypeError( f"unsupported operation type(s) [{method_name!r}] " @@ -1409,7 +1410,8 @@ def _numeric_op(self, right: Any, method_name: str) -> "Series": series = Series( _query_compiler=self._query_compiler.arithmetic_op_fields( - display_name, left_object + display_name=display_name, + arithmetic_object=left_object, ) ) diff --git a/eland/tasks.py b/eland/tasks.py index 61645777..fd3aeaf6 100644 --- a/eland/tasks.py +++ b/eland/tasks.py @@ -42,7 +42,7 @@ class Task(ABC): The task type (e.g. head, tail etc.) """ - def __init__(self, task_type: str): + def __init__(self, task_type: str) -> None: self._task_type = task_type @property @@ -345,7 +345,7 @@ def __repr__(self) -> str: class ArithmeticOpFieldsTask(Task): - def __init__(self, display_name: str, arithmetic_series: ArithmeticSeries): + def __init__(self, display_name: str, arithmetic_series: ArithmeticSeries) -> None: super().__init__("arithmetic_op_fields") self._display_name = display_name diff --git a/noxfile.py b/noxfile.py index 4c1cec06..27d78931 100644 --- a/noxfile.py +++ b/noxfile.py @@ -38,6 +38,7 @@ "eland/tasks.py", "eland/utils.py", "eland/groupby.py", + "eland/field_mappings.py", "eland/operations.py", "eland/ndframe.py", "eland/ml/__init__.py", diff --git a/tests/notebook/test_demo_notebook.ipynb b/tests/notebook/test_demo_notebook.ipynb index f387d296..e8e487d0 100644 --- a/tests/notebook/test_demo_notebook.ipynb +++ b/tests/notebook/test_demo_notebook.ipynb @@ -49,7 +49,7 @@ "metadata": {}, "outputs": [], "source": [ - "ed_flights = ed.DataFrame('http://localhost:9200', 'flights')" + "ed_flights = ed.DataFrame('localhost', 'flights')" ] }, { @@ -565,40 +565,18 @@ }, { "cell_type": "code", - "execution_count": 12, + "execution_count": null, "metadata": {}, - "outputs": [ - { - "data": { - "text/plain": [ - "False" - ] - }, - "execution_count": 12, - "metadata": {}, - "output_type": "execute_result" - } - ], + "outputs": [], "source": [ "pd_flights.empty" ] }, { "cell_type": "code", - "execution_count": 13, + "execution_count": null, "metadata": {}, - "outputs": [ - { - "data": { - "text/plain": [ - "False" - ] - }, - "execution_count": 13, - "metadata": {}, - "output_type": "execute_result" - } - ], + "outputs": [], "source": [ "ed_flights.empty" ] @@ -612,40 +590,18 @@ }, { "cell_type": "code", - "execution_count": 14, + "execution_count": null, "metadata": {}, - "outputs": [ - { - "data": { - "text/plain": [ - "(13059, 27)" - ] - }, - "execution_count": 14, - "metadata": {}, - "output_type": "execute_result" - } - ], + "outputs": [], "source": [ "pd_flights.shape" ] }, { "cell_type": "code", - "execution_count": 15, + "execution_count": null, "metadata": {}, - "outputs": [ - { - "data": { - "text/plain": [ - "(13059, 27)" - ] - }, - "execution_count": 15, - "metadata": {}, - "output_type": "execute_result" - } - ], + "outputs": [], "source": [ "ed_flights.shape" ] @@ -661,43 +617,18 @@ }, { "cell_type": "code", - "execution_count": 16, + "execution_count": null, "metadata": {}, - "outputs": [ - { - "data": { - "text/plain": [ - "Index(['0', '1', '2', '3', '4', '5', '6', '7', '8', '9',\n", - " ...\n", - " '13049', '13050', '13051', '13052', '13053', '13054', '13055', '13056', '13057', '13058'],\n", - " dtype='object', length=13059)" - ] - }, - "execution_count": 16, - "metadata": {}, - "output_type": "execute_result" - } - ], + "outputs": [], "source": [ "pd_flights.index" ] }, { "cell_type": "code", - "execution_count": 17, + "execution_count": null, "metadata": {}, - "outputs": [ - { - "data": { - "text/plain": [ - "" - ] - }, - "execution_count": 17, - "metadata": {}, - "output_type": "execute_result" - } - ], + "outputs": [], "source": [ "# NBVAL_IGNORE_OUTPUT\n", "ed_flights.index" @@ -705,20 +636,9 @@ }, { "cell_type": "code", - "execution_count": 18, + "execution_count": null, "metadata": {}, - "outputs": [ - { - "data": { - "text/plain": [ - "'_id'" - ] - }, - "execution_count": 18, - "metadata": {}, - "output_type": "execute_result" - } - ], + "outputs": [], "source": [ "ed_flights.index.es_index_field" ] @@ -734,49 +654,18 @@ }, { "cell_type": "code", - "execution_count": 19, + "execution_count": null, "metadata": {}, - "outputs": [ - { - "data": { - "text/plain": [ - "array([[841.2656419677076, False, 'Kibana Airlines', ..., 'Sunny', 0,\n", - " Timestamp('2018-01-01 00:00:00')],\n", - " [882.9826615595518, False, 'Logstash Airways', ..., 'Clear', 0,\n", - " Timestamp('2018-01-01 18:27:00')],\n", - " [190.6369038508356, False, 'Logstash Airways', ..., 'Rain', 0,\n", - " Timestamp('2018-01-01 17:11:14')],\n", - " ...,\n", - " [997.7518761454494, False, 'Logstash Airways', ..., 'Sunny', 6,\n", - " Timestamp('2018-02-11 04:09:27')],\n", - " [1102.8144645388556, False, 'JetBeats', ..., 'Hail', 6,\n", - " Timestamp('2018-02-11 08:28:21')],\n", - " [858.1443369038839, False, 'JetBeats', ..., 'Rain', 6,\n", - " Timestamp('2018-02-11 14:54:34')]], dtype=object)" - ] - }, - "execution_count": 19, - "metadata": {}, - "output_type": "execute_result" - } - ], + "outputs": [], "source": [ "pd_flights.values" ] }, { "cell_type": "code", - "execution_count": 20, + "execution_count": null, "metadata": {}, - "outputs": [ - { - "name": "stdout", - "output_type": "stream", - "text": [ - "This method would scan/scroll the entire Elasticsearch index(s) into memory. If this is explicitly required, and there is sufficient memory, call `ed.eland_to_pandas(ed_df).values`\n" - ] - } - ], + "outputs": [], "source": [ "try:\n", " ed_flights.values\n", @@ -800,198 +689,18 @@ }, { "cell_type": "code", - "execution_count": 21, + "execution_count": null, "metadata": {}, - "outputs": [ - { - "data": { - "text/html": [ - "
\n", - "\n", - "\n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - "
AvgTicketPriceCancelled...dayOfWeektimestamp
0841.265642False...02018-01-01 00:00:00
1882.982662False...02018-01-01 18:27:00
2190.636904False...02018-01-01 17:11:14
3181.694216True...02018-01-01 10:33:28
4730.041778False...02018-01-01 05:13:00
\n", - "

5 rows × 27 columns

\n", - "
" - ], - "text/plain": [ - " AvgTicketPrice Cancelled ... dayOfWeek timestamp\n", - "0 841.265642 False ... 0 2018-01-01 00:00:00\n", - "1 882.982662 False ... 0 2018-01-01 18:27:00\n", - "2 190.636904 False ... 0 2018-01-01 17:11:14\n", - "3 181.694216 True ... 0 2018-01-01 10:33:28\n", - "4 730.041778 False ... 0 2018-01-01 05:13:00\n", - "\n", - "[5 rows x 27 columns]" - ] - }, - "execution_count": 21, - "metadata": {}, - "output_type": "execute_result" - } - ], + "outputs": [], "source": [ "pd_flights.head()" ] }, { "cell_type": "code", - "execution_count": 22, + "execution_count": null, "metadata": {}, - "outputs": [ - { - "data": { - "text/html": [ - "
\n", - "\n", - "\n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - "
AvgTicketPriceCancelled...dayOfWeektimestamp
0841.265642False...02018-01-01 00:00:00
1882.982662False...02018-01-01 18:27:00
2190.636904False...02018-01-01 17:11:14
3181.694216True...02018-01-01 10:33:28
4730.041778False...02018-01-01 05:13:00
\n", - "
\n", - "

5 rows × 27 columns

" - ], - "text/plain": [ - " AvgTicketPrice Cancelled ... dayOfWeek timestamp\n", - "0 841.265642 False ... 0 2018-01-01 00:00:00\n", - "1 882.982662 False ... 0 2018-01-01 18:27:00\n", - "2 190.636904 False ... 0 2018-01-01 17:11:14\n", - "3 181.694216 True ... 0 2018-01-01 10:33:28\n", - "4 730.041778 False ... 0 2018-01-01 05:13:00\n", - "\n", - "[5 rows x 27 columns]" - ] - }, - "execution_count": 22, - "metadata": {}, - "output_type": "execute_result" - } - ], + "outputs": [], "source": [ "ed_flights.head()" ] @@ -1005,198 +714,18 @@ }, { "cell_type": "code", - "execution_count": 23, + "execution_count": null, "metadata": {}, - "outputs": [ - { - "data": { - "text/html": [ - "
\n", - "\n", - "\n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - "
AvgTicketPriceCancelled...dayOfWeektimestamp
130541080.446279False...62018-02-11 20:42:25
13055646.612941False...62018-02-11 01:41:57
13056997.751876False...62018-02-11 04:09:27
130571102.814465False...62018-02-11 08:28:21
13058858.144337False...62018-02-11 14:54:34
\n", - "

5 rows × 27 columns

\n", - "
" - ], - "text/plain": [ - " AvgTicketPrice Cancelled ... dayOfWeek timestamp\n", - "13054 1080.446279 False ... 6 2018-02-11 20:42:25\n", - "13055 646.612941 False ... 6 2018-02-11 01:41:57\n", - "13056 997.751876 False ... 6 2018-02-11 04:09:27\n", - "13057 1102.814465 False ... 6 2018-02-11 08:28:21\n", - "13058 858.144337 False ... 6 2018-02-11 14:54:34\n", - "\n", - "[5 rows x 27 columns]" - ] - }, - "execution_count": 23, - "metadata": {}, - "output_type": "execute_result" - } - ], + "outputs": [], "source": [ "pd_flights.tail()" ] }, { "cell_type": "code", - "execution_count": 24, + "execution_count": null, "metadata": {}, - "outputs": [ - { - "data": { - "text/html": [ - "
\n", - "\n", - "\n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - "
AvgTicketPriceCancelled...dayOfWeektimestamp
130541080.446279False...62018-02-11 20:42:25
13055646.612941False...62018-02-11 01:41:57
13056997.751876False...62018-02-11 04:09:27
130571102.814465False...62018-02-11 08:28:21
13058858.144337False...62018-02-11 14:54:34
\n", - "
\n", - "

5 rows × 27 columns

" - ], - "text/plain": [ - " AvgTicketPrice Cancelled ... dayOfWeek timestamp\n", - "13054 1080.446279 False ... 6 2018-02-11 20:42:25\n", - "13055 646.612941 False ... 6 2018-02-11 01:41:57\n", - "13056 997.751876 False ... 6 2018-02-11 04:09:27\n", - "13057 1102.814465 False ... 6 2018-02-11 08:28:21\n", - "13058 858.144337 False ... 6 2018-02-11 14:54:34\n", - "\n", - "[5 rows x 27 columns]" - ] - }, - "execution_count": 24, - "metadata": {}, - "output_type": "execute_result" - } - ], + "outputs": [], "source": [ "ed_flights.tail()" ] @@ -1210,52 +739,18 @@ }, { "cell_type": "code", - "execution_count": 25, + "execution_count": null, "metadata": {}, - "outputs": [ - { - "data": { - "text/plain": [ - "Index(['AvgTicketPrice', 'Cancelled', 'Carrier', 'Dest', 'DestAirportID', 'DestCityName',\n", - " 'DestCountry', 'DestLocation', 'DestRegion', 'DestWeather', 'DistanceKilometers',\n", - " 'DistanceMiles', 'FlightDelay', 'FlightDelayMin', 'FlightDelayType', 'FlightNum',\n", - " 'FlightTimeHour', 'FlightTimeMin', 'Origin', 'OriginAirportID', 'OriginCityName',\n", - " 'OriginCountry', 'OriginLocation', 'OriginRegion', 'OriginWeather', 'dayOfWeek',\n", - " 'timestamp'],\n", - " dtype='object')" - ] - }, - "execution_count": 25, - "metadata": {}, - "output_type": "execute_result" - } - ], + "outputs": [], "source": [ "pd_flights.keys()" ] }, { "cell_type": "code", - "execution_count": 26, + "execution_count": null, "metadata": {}, - "outputs": [ - { - "data": { - "text/plain": [ - "Index(['AvgTicketPrice', 'Cancelled', 'Carrier', 'Dest', 'DestAirportID', 'DestCityName',\n", - " 'DestCountry', 'DestLocation', 'DestRegion', 'DestWeather', 'DistanceKilometers',\n", - " 'DistanceMiles', 'FlightDelay', 'FlightDelayMin', 'FlightDelayType', 'FlightNum',\n", - " 'FlightTimeHour', 'FlightTimeMin', 'Origin', 'OriginAirportID', 'OriginCityName',\n", - " 'OriginCountry', 'OriginLocation', 'OriginRegion', 'OriginWeather', 'dayOfWeek',\n", - " 'timestamp'],\n", - " dtype='object')" - ] - }, - "execution_count": 26, - "metadata": {}, - "output_type": "execute_result" - } - ], + "outputs": [], "source": [ "ed_flights.keys()" ] @@ -1269,179 +764,27 @@ }, { "cell_type": "code", - "execution_count": 27, + "execution_count": null, "metadata": {}, - "outputs": [ - { - "data": { - "text/plain": [ - "0 Kibana Airlines\n", - "1 Logstash Airways\n", - "2 Logstash Airways\n", - "3 Kibana Airlines\n", - "4 Kibana Airlines\n", - " ... \n", - "13054 Logstash Airways\n", - "13055 Logstash Airways\n", - "13056 Logstash Airways\n", - "13057 JetBeats\n", - "13058 JetBeats\n", - "Name: Carrier, Length: 13059, dtype: object" - ] - }, - "execution_count": 27, - "metadata": {}, - "output_type": "execute_result" - } - ], + "outputs": [], "source": [ - "pd_flights.get('Carrier')" - ] - }, - { - "cell_type": "code", - "execution_count": 28, - "metadata": {}, - "outputs": [ - { - "data": { - "text/plain": [ - "0 Kibana Airlines\n", - "1 Logstash Airways\n", - "2 Logstash Airways\n", - "3 Kibana Airlines\n", - "4 Kibana Airlines\n", - " ... \n", - "13054 Logstash Airways\n", - "13055 Logstash Airways\n", - "13056 Logstash Airways\n", - "13057 JetBeats\n", - "13058 JetBeats\n", - "Name: Carrier, Length: 13059, dtype: object" - ] - }, - "execution_count": 28, - "metadata": {}, - "output_type": "execute_result" - } - ], + "pd_flights.get('Carrier')" + ] + }, + { + "cell_type": "code", + "execution_count": null, + "metadata": {}, + "outputs": [], "source": [ "ed_flights.get('Carrier')" ] }, { "cell_type": "code", - "execution_count": 29, + "execution_count": null, "metadata": {}, - "outputs": [ - { - "data": { - "text/html": [ - "
\n", - "\n", - "\n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - "
CarrierOrigin
0Kibana AirlinesFrankfurt am Main Airport
1Logstash AirwaysCape Town International Airport
2Logstash AirwaysVenice Marco Polo Airport
3Kibana AirlinesNaples International Airport
4Kibana AirlinesLicenciado Benito Juarez International Airport
.........
13054Logstash AirwaysPisa International Airport
13055Logstash AirwaysWinnipeg / James Armstrong Richardson Internat...
13056Logstash AirwaysLicenciado Benito Juarez International Airport
13057JetBeatsItami Airport
13058JetBeatsAdelaide International Airport
\n", - "

13059 rows × 2 columns

\n", - "
" - ], - "text/plain": [ - " Carrier Origin\n", - "0 Kibana Airlines Frankfurt am Main Airport\n", - "1 Logstash Airways Cape Town International Airport\n", - "2 Logstash Airways Venice Marco Polo Airport\n", - "3 Kibana Airlines Naples International Airport\n", - "4 Kibana Airlines Licenciado Benito Juarez International Airport\n", - "... ... ...\n", - "13054 Logstash Airways Pisa International Airport\n", - "13055 Logstash Airways Winnipeg / James Armstrong Richardson Internat...\n", - "13056 Logstash Airways Licenciado Benito Juarez International Airport\n", - "13057 JetBeats Itami Airport\n", - "13058 JetBeats Adelaide International Airport\n", - "\n", - "[13059 rows x 2 columns]" - ] - }, - "execution_count": 29, - "metadata": {}, - "output_type": "execute_result" - } - ], + "outputs": [], "source": [ "pd_flights.get(['Carrier', 'Origin'])" ] @@ -1455,17 +798,9 @@ }, { "cell_type": "code", - "execution_count": 30, + "execution_count": null, "metadata": {}, - "outputs": [ - { - "name": "stdout", - "output_type": "stream", - "text": [ - "unhashable type: 'list'\n" - ] - } - ], + "outputs": [], "source": [ "try:\n", " ed_flights.get(['Carrier', 'Origin'])\n", @@ -1482,153 +817,9 @@ }, { "cell_type": "code", - "execution_count": 31, + "execution_count": null, "metadata": {}, - "outputs": [ - { - "data": { - "text/html": [ - "
\n", - "\n", - "\n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - "
AvgTicketPriceCancelled...dayOfWeektimestamp
8960.869736True...02018-01-01 12:09:35
26975.812632True...02018-01-01 15:38:32
311946.358410True...02018-01-01 11:51:12
651975.383864True...22018-01-03 21:13:17
950907.836523True...22018-01-03 05:14:51
..................
12820909.973606True...52018-02-10 05:11:35
12906983.429244True...62018-02-11 06:19:58
129181136.678150True...62018-02-11 16:03:10
129191105.211803True...62018-02-11 05:36:05
130131055.350213True...62018-02-11 13:20:16
\n", - "

68 rows × 27 columns

\n", - "
" - ], - "text/plain": [ - " AvgTicketPrice Cancelled ... dayOfWeek timestamp\n", - "8 960.869736 True ... 0 2018-01-01 12:09:35\n", - "26 975.812632 True ... 0 2018-01-01 15:38:32\n", - "311 946.358410 True ... 0 2018-01-01 11:51:12\n", - "651 975.383864 True ... 2 2018-01-03 21:13:17\n", - "950 907.836523 True ... 2 2018-01-03 05:14:51\n", - "... ... ... ... ... ...\n", - "12820 909.973606 True ... 5 2018-02-10 05:11:35\n", - "12906 983.429244 True ... 6 2018-02-11 06:19:58\n", - "12918 1136.678150 True ... 6 2018-02-11 16:03:10\n", - "12919 1105.211803 True ... 6 2018-02-11 05:36:05\n", - "13013 1055.350213 True ... 6 2018-02-11 13:20:16\n", - "\n", - "[68 rows x 27 columns]" - ] - }, - "execution_count": 31, - "metadata": {}, - "output_type": "execute_result" - } - ], + "outputs": [], "source": [ "pd_flights.query('Carrier == \"Kibana Airlines\" & AvgTicketPrice > 900.0 & Cancelled == True')" ] @@ -2508,7 +1699,7 @@ " \n", " \n", " AvgTicketPrice\n", - " DistanceKilometers\n", + " Cancelled\n", " ...\n", " FlightTimeMin\n", " dayOfWeek\n", @@ -2526,18 +1717,18 @@ " \n", " mean\n", " 628.253689\n", - " 7092.142457\n", + " 0.128494\n", " ...\n", " 511.127842\n", " 2.835975\n", " \n", " \n", " std\n", - " 266.386661\n", - " 4578.263193\n", + " 266.407061\n", + " 0.334664\n", " ...\n", - " 334.741135\n", - " 1.939365\n", + " 334.766770\n", + " 1.939513\n", " \n", " \n", " min\n", @@ -2550,7 +1741,7 @@ " \n", " 25%\n", " 410.008918\n", - " 2470.545974\n", + " 0.000000\n", " ...\n", " 251.938710\n", " 1.000000\n", @@ -2558,7 +1749,7 @@ " \n", " 50%\n", " 640.387285\n", - " 7612.072403\n", + " 0.000000\n", " ...\n", " 503.148975\n", " 3.000000\n", @@ -2566,7 +1757,7 @@ " \n", " 75%\n", " 842.213490\n", - " 9735.660463\n", + " 0.000000\n", " ...\n", " 720.505705\n", " 4.000000\n", @@ -2574,28 +1765,28 @@ " \n", " max\n", " 1199.729004\n", - " 19881.482422\n", + " 1.000000\n", " ...\n", " 1902.901978\n", " 6.000000\n", " \n", " \n", "\n", - "

8 rows × 7 columns

\n", + "

8 rows × 9 columns

\n", "" ], "text/plain": [ - " AvgTicketPrice DistanceKilometers ... FlightTimeMin dayOfWeek\n", - "count 13059.000000 13059.000000 ... 13059.000000 13059.000000\n", - "mean 628.253689 7092.142457 ... 511.127842 2.835975\n", - "std 266.386661 4578.263193 ... 334.741135 1.939365\n", - "min 100.020531 0.000000 ... 0.000000 0.000000\n", - "25% 410.008918 2470.545974 ... 251.938710 1.000000\n", - "50% 640.387285 7612.072403 ... 503.148975 3.000000\n", - "75% 842.213490 9735.660463 ... 720.505705 4.000000\n", - "max 1199.729004 19881.482422 ... 1902.901978 6.000000\n", + " AvgTicketPrice Cancelled ... FlightTimeMin dayOfWeek\n", + "count 13059.000000 13059.000000 ... 13059.000000 13059.000000\n", + "mean 628.253689 0.128494 ... 511.127842 2.835975\n", + "std 266.407061 0.334664 ... 334.766770 1.939513\n", + "min 100.020531 0.000000 ... 0.000000 0.000000\n", + "25% 410.008918 0.000000 ... 251.938710 1.000000\n", + "50% 640.387285 0.000000 ... 503.148975 3.000000\n", + "75% 842.213490 0.000000 ... 720.505705 4.000000\n", + "max 1199.729004 1.000000 ... 1902.901978 6.000000\n", "\n", - "[8 rows x 7 columns]" + "[8 rows x 9 columns]" ] }, "execution_count": 40, @@ -2708,7 +1899,7 @@ " 26 timestamp 13059 non-null datetime64[ns]\n", "dtypes: bool(2), datetime64[ns](1), float64(5), int64(2), object(17)\n", "memory usage: 64.000 bytes\n", - "Elasticsearch storage usage: 5.046 MB\n" + "Elasticsearch storage usage: 4.923 MB\n" ] } ], @@ -3383,7 +2574,7 @@ "outputs": [ { "data": { - "image/png": "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", + "image/png": "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\n", "text/plain": [ "
" ] @@ -3406,7 +2597,7 @@ "outputs": [ { "data": { - "image/png": "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", + "image/png": "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\n", "text/plain": [ "
" ] @@ -3455,11 +2646,11 @@ " is_source_field: False\n", "Mappings:\n", " capabilities:\n", - " es_field_name is_source es_dtype es_date_format pd_dtype is_searchable is_aggregatable is_scripted aggregatable_es_field_name\n", - "timestamp timestamp True date strict_date_hour_minute_second datetime64[ns] True True False timestamp\n", - "OriginAirportID OriginAirportID True keyword None object True True False OriginAirportID\n", - "DestAirportID DestAirportID True keyword None object True True False DestAirportID\n", - "FlightDelayMin FlightDelayMin True integer None int64 True True False FlightDelayMin\n", + " display_name es_field_name is_source es_dtype es_date_format pd_dtype is_searchable is_aggregatable is_scripted aggregatable_es_field_name\n", + "0 timestamp timestamp True date strict_date_hour_minute_second datetime64[ns] True True False timestamp\n", + "1 OriginAirportID OriginAirportID True keyword None object True True False OriginAirportID\n", + "2 DestAirportID DestAirportID True keyword None object True True False DestAirportID\n", + "3 FlightDelayMin FlightDelayMin True integer None int64 True True False FlightDelayMin\n", "Operations:\n", " tasks: [('boolean_filter': ('boolean_filter': {'bool': {'must': [{'term': {'OriginAirportID': 'AMS'}}, {'range': {'FlightDelayMin': {'gt': 60}}}]}})), ('tail': ('sort_field': '_doc', 'count': 5))]\n", " size: 5\n", @@ -3477,8 +2668,11 @@ } ], "metadata": { + "interpreter": { + "hash": "d22e07e3c0eda54bcbbcc3abaf38008b628f195b142529d1ab17deb0e4ffdc82" + }, "kernelspec": { - "display_name": "Python 3", + "display_name": "Python 3 (ipykernel)", "language": "python", "name": "python3" }, @@ -3492,7 +2686,7 @@ "name": "python", "nbconvert_exporter": "python", "pygments_lexer": "ipython3", - "version": "3.8.5" + "version": "3.9.5" } }, "nbformat": 4, diff --git a/tests/series/test_arithmetics_pytest.py b/tests/series/test_arithmetics_pytest.py index 4a6251af..2a94a0f8 100644 --- a/tests/series/test_arithmetics_pytest.py +++ b/tests/series/test_arithmetics_pytest.py @@ -15,67 +15,31 @@ # specific language governing permissions and limitations # under the License. -from datetime import datetime - # File called _pytest for PyCharm compatability import numpy as np import pytest -from eland import Series -from tests.common import TestData, assert_pandas_eland_series_equal +from tests.common import TestData, assert_almost_equal, assert_pandas_eland_series_equal class TestSeriesArithmetics(TestData): - def test_ecommerce_datetime_comparisons(self): - pd_df = self.pd_ecommerce() - ed_df = self.ed_ecommerce() - - ops = ["__le__", "__lt__", "__gt__", "__ge__", "__eq__", "__ne__"] - - # this datetime object is timezone naive - datetime_obj = datetime(2016, 12, 18) - - # FIXME: the following timezone conversions are just a temporary fix - # to run the datetime comparison tests - # - # The problem: - # - the datetime objects of the pandas DataFrame are timezone aware and - # can't be compared with timezone naive datetime objects - # - the datetime objects of the eland DataFrame are timezone naive (which - # should be fixed) - # - however if the eland DataFrame is converted to a pandas DataFrame - # (using the `to_pandas` function) the datetime objects become timezone aware - # - # This tests converts the datetime objects of the pandas Series to - # timezone naive ones and utilizes a class to make the datetime objects of the - # eland Series timezone naive before the result of `to_pandas` is returned. - # The `to_pandas` function is executed by the `assert_pandas_eland_series_equal` - # function, which compares the eland and pandas Series - - # convert to timezone naive datetime object - pd_df["order_date"] = pd_df["order_date"].dt.tz_localize(None) - - class ModifiedElandSeries(Series): - def to_pandas(self): - """remove timezone awareness before returning the pandas dataframe""" - series = super().to_pandas() - series = series.dt.tz_localize(None) - return series - - for op in ops: - pd_series = pd_df[getattr(pd_df["order_date"], op)(datetime_obj)][ - "order_date" - ] - ed_series = ed_df[getattr(ed_df["order_date"], op)(datetime_obj)][ - "order_date" - ] - - # "type cast" to modified class (inherits from ed.Series) that overrides the `to_pandas` function - ed_series.__class__ = ModifiedElandSeries - - assert_pandas_eland_series_equal( - pd_series, ed_series, check_less_precise=True - ) + ops = [ + "__add__", + "__truediv__", + "__floordiv__", + "__pow__", + "__mod__", + "__mul__", + "__sub__", + "add", + "truediv", + "floordiv", + "pow", + "mod", + "mul", + "sub", + ] + funcs = ["max", "min", "mean", "sum", "median", "var", "std"] def test_ecommerce_series_invalid_div(self): pd_df = self.pd_ecommerce() @@ -124,47 +88,26 @@ def test_ecommerce_series_simple_series_addition(self): assert_pandas_eland_series_equal(pd_series, ed_series, rtol=True) - def test_ecommerce_series_basic_arithmetics(self): + @pytest.mark.parametrize("op", ops) + def test_ecommerce_series_basic_arithmetics(self, op): pd_df = self.pd_ecommerce().head(100) ed_df = self.ed_ecommerce().head(100) - ops = [ - "__add__", - "__truediv__", - "__floordiv__", - "__pow__", - "__mod__", - "__mul__", - "__sub__", - "add", - "truediv", - "floordiv", - "pow", - "mod", - "mul", - "sub", - ] - - for op in ops: - pd_series = getattr(pd_df["taxful_total_price"], op)( - pd_df["total_quantity"] - ) - ed_series = getattr(ed_df["taxful_total_price"], op)( - ed_df["total_quantity"] - ) - assert_pandas_eland_series_equal(pd_series, ed_series, rtol=True) + pd_series = getattr(pd_df["taxful_total_price"], op)(pd_df["total_quantity"]) + ed_series = getattr(ed_df["taxful_total_price"], op)(ed_df["total_quantity"]) + assert_pandas_eland_series_equal(pd_series, ed_series, rtol=True) - pd_series = getattr(pd_df["taxful_total_price"], op)(10.56) - ed_series = getattr(ed_df["taxful_total_price"], op)(10.56) - assert_pandas_eland_series_equal(pd_series, ed_series, rtol=True) + pd_series = getattr(pd_df["taxful_total_price"], op)(10.56) + ed_series = getattr(ed_df["taxful_total_price"], op)(10.56) + assert_pandas_eland_series_equal(pd_series, ed_series, rtol=True) - pd_series = getattr(pd_df["taxful_total_price"], op)(np.float32(1.879)) - ed_series = getattr(ed_df["taxful_total_price"], op)(np.float32(1.879)) - assert_pandas_eland_series_equal(pd_series, ed_series, rtol=True) + pd_series = getattr(pd_df["taxful_total_price"], op)(np.float32(1.879)) + ed_series = getattr(ed_df["taxful_total_price"], op)(np.float32(1.879)) + assert_pandas_eland_series_equal(pd_series, ed_series, rtol=True) - pd_series = getattr(pd_df["taxful_total_price"], op)(int(8)) - ed_series = getattr(ed_df["taxful_total_price"], op)(int(8)) - assert_pandas_eland_series_equal(pd_series, ed_series, rtol=True) + pd_series = getattr(pd_df["taxful_total_price"], op)(int(8)) + ed_series = getattr(ed_df["taxful_total_price"], op)(int(8)) + assert_pandas_eland_series_equal(pd_series, ed_series, rtol=True) def test_supported_series_dtypes_ops(self): pd_df = self.pd_ecommerce().head(100) @@ -353,3 +296,26 @@ def test_supported_series_dtypes_rops(self): pd_series = getattr(pd_df["total_quantity"], op)(pd_df["currency"]) with pytest.raises(TypeError): ed_series = getattr(ed_df["total_quantity"], op)(ed_df["currency"]) + + def test_scripted_series_nunique(self): + pd_df = self.pd_flights() + ed_df = self.ed_flights() + + ed_nunique = ed_df["DestCountry"] + ed_df["OriginCountry"] + pd_nunique = pd_df["DestCountry"] + pd_df["OriginCountry"] + + assert ed_nunique.nunique() == pd_nunique.nunique() + + @pytest.mark.parametrize("func", funcs) + @pytest.mark.parametrize("op", ops) + def test_scripted_series_ops(self, func, op): + pd_df = self.pd_ecommerce() + ed_df = self.ed_ecommerce() + + pd_series = getattr(pd_df["taxful_total_price"], op)(pd_df["total_quantity"]) + ed_series = getattr(ed_df["taxful_total_price"], op)(ed_df["total_quantity"]) + + ed_agg = getattr(pd_series, func)() + pd_agg = getattr(ed_series, func)() + + assert_almost_equal(pd_agg, ed_agg)