flowtask 5.8.4__cp310-cp310-manylinux2014_x86_64.manylinux_2_17_x86_64.whl

flowtask/components/TransformRows/TransformRows.py

@@ -0,0 +1,507 @@
+import asyncio
+import copy
+from typing import Any, Union
+from collections.abc import Callable
+import pandas
+import numpy as np
+from datamodel.typedefs.types import AttrDict
+from asyncdb.exceptions import NoDataFound
+from navconfig.logging import logging
+from querysource.types import strtobool
+from querysource.types.dt import transforms as qsdfunctions
+from . import functions as dffunctions
+from ...exceptions import (
+    ComponentError,
+    DataNotFound,
+    ConfigError
+)
+from ...utils.executor import getFunction
+from ...utils.functions import check_empty
+from ..flow import FlowComponent
+
+
+class TransformRows(FlowComponent):
+    """
+    TransformRows
+
+    Overview
+
+        The TransformRows class is a component for transforming, adding, or modifying rows in a Pandas DataFrame based on
+        specified criteria. It supports single and multiple DataFrame transformations, and various operations on columns.
+
+    .. table:: Properties
+    :widths: auto
+
+        +--------------------+----------+-------------------------------------------------------------------------------------------------------+
+        | Name               | Required | Description                                                                                           |
+        +--------------------+----------+-------------------------------------------------------------------------------------------------------+
+        | fields             |   No     | A dictionary defining the fields and corresponding transformations to be applied.                     |
+        +--------------------+----------+-------------------------------------------------------------------------------------------------------+
+        | filter_conditions  |   No     | A dictionary defining the filter conditions for transformations.                                      |
+        +--------------------+----------+-------------------------------------------------------------------------------------------------------+
+        | clean_notnull      |   No     | Boolean flag indicating if non-null values should be cleaned, defaults to True.                       |
+        +--------------------+----------+-------------------------------------------------------------------------------------------------------+
+        | replace_columns    |   No     | Boolean flag indicating if columns should be replaced, defaults to False.                             |
+        +--------------------+----------+-------------------------------------------------------------------------------------------------------+
+        | multi              |   No     | Boolean flag indicating if multiple DataFrame transformations should be supported, defaults to False. |
+        +--------------------+----------+-------------------------------------------------------------------------------------------------------+
+        | function           |   No     | View the list of function in the functions.py file on this directory                                  |
+        +--------------------+----------+-------------------------------------------------------------------------------------------------------+
+        | _applied           |   No     | List to store the applied transformations.                                                            |
+        +--------------------+----------+-------------------------------------------------------------------------------------------------------+
+
+    Return
+
+        The methods in this class manage the transformation of DataFrames, including initialization, execution, and result handling.
+
+
+
+        Example:
+
+        ```yaml
+        TransformRows:
+          fields:
+            display_name:
+              value:
+              - concat
+              - columns:
+                - first_name
+                - last_name
+            legal_address:
+              value:
+              - concat
+              - columns:
+                - legal_street_address_1
+                - legal_street_address_2
+            work_address:
+              value:
+              - concat
+              - columns:
+                - work_location_address_1
+                - work_location_address_2
+            first_name:
+              value:
+              - capitalize
+            last_name:
+              value:
+              - capitalize
+            warp_id:
+              value:
+              - nullif
+              - chars:
+                - '*'
+            old_warp_id:
+              value:
+              - nullif
+              - chars:
+                - '*'
+            worker_category_description:
+              value:
+              - case
+              - column: benefits_eligibility_class_code
+                condition: PART-TIME
+                match: Part Time
+                notmatch: Full Time
+            file_number:
+              value:
+              - ereplace
+              - columns:
+                - position_id
+                - payroll_group
+                newvalue: ''
+            original_hire_date:
+              value:
+              - convert_to_datetime
+            hire_date:
+              value:
+              - convert_to_datetime
+            start_date:
+              value:
+              - convert_to_datetime
+            updated:
+              value:
+              - convert_to_datetime
+            gender_code:
+              value:
+              - convert_to_string
+            payroll_id:
+              value:
+              - convert_to_string
+            reports_to_payroll_id:
+              value:
+              - convert_to_string
+        ```
+
+    """  # noqa
+
+    def __init__(
+        self,
+        loop: asyncio.AbstractEventLoop = None,
+        job: Callable = None,
+        stat: Callable = None,
+        **kwargs,
+    ):
+        self.fields: dict = {}
+        self.filter_conditions: dict = {}
+        self.clean_notnull: bool = True
+        self.replace_columns: bool = False
+        self._applied: list = []
+        replace_cols = kwargs.pop("replace_columns", False)
+        if isinstance(replace_cols, str):
+            self.replace_columns = strtobool(replace_cols)
+        else:
+            self.replace_columns = bool(replace_cols)
+        # support for multiple dataframe transformations
+        try:
+            self.multi = bool(kwargs["multi"])
+            del kwargs["multi"]
+        except KeyError:
+            self.multi = False
+        if self.multi is False:
+            if "fields" in kwargs:
+                self.fields = kwargs["fields"]
+                del kwargs["fields"]
+        else:
+            self.fields = {}
+        super(TransformRows, self).__init__(loop=loop, job=job, stat=stat, **kwargs)
+        self._logger.debug(f"Using Replace Columns: {self.replace_columns}")
+
+    async def start(self, **kwargs):
+        """Obtain Pandas Dataframe."""
+        if self.previous:
+            self.data = self.input
+        else:
+            raise ComponentError("a Previous Component was not found.")
+        if check_empty(self.data):
+            raise DataNotFound("No data was found")
+
+    async def run(self):
+        if self.data is None:
+            return False
+        if isinstance(self.data, pandas.DataFrame):
+            self.add_metric("started_rows", self.data.shape[0])
+            self.add_metric("started_columns", self.data.shape[1])
+            df = await self.transform(self, self.data)
+            self._result = df
+            # avoid threat the Dataframe as a Copy
+            self._result.is_copy = None
+            return self._result
+        elif self.multi is True:
+            # iteration over every Pandas DT:
+            try:
+                result = self.data.items()
+            except Exception as err:
+                raise ComponentError(
+                    f"TransformRows: Invalid Result type for Multiple: {err}"
+                ) from err
+            self._result = {}
+            for name, rs in result:
+                try:
+                    el = getattr(self, name)
+                except AttributeError:
+                    self._result[name] = rs
+                    continue
+                df = await self.transform(AttrDict(el), rs)
+                self._result[name] = df
+            return self._result
+        else:
+            raise NoDataFound(
+                "TransformRows: Pandas Dataframe Empty or is not a Dataframe"
+            )
+
+    async def _scalar_column(
+        self,
+        it: pandas.DataFrame,
+        field: str,
+        val: Any
+    ) -> None:
+        """_scalar_column.
+
+        Operating the Dataframe Function Column as an Scalar Value.
+        """
+        try:
+            if val.startswith('{'):
+                it[field] = self.mask_replacement(val)
+            else:
+                it[field] = it[val]
+                if self.replace_columns is True:
+                    it.drop(val, axis="columns", inplace=True)
+        except KeyError:
+            self._logger.error(
+                f"Column doesn't exists: {val}"
+            )
+        except Exception as e:
+            self._logger.error(
+                f"Error dropping Column: {val}, {e}"
+            )
+
+    def _get_df_function(self, fn: str) -> Union[Callable, None]:
+        func = getattr(dffunctions, fn, None)
+        if func is None:
+            func = getattr(qsdfunctions, fn)
+        return func
+
+    async def _check_condition(self, df, field: str, fname: str, conditions: dict):
+        """Check if a field meets the specified condition."""
+        # Check column existence
+        if conditions.get("exists") and field not in df.columns:
+            return False
+
+        # Check for specific criteria (e.g., column must have non-null values)
+        if "not_null" in conditions and conditions["not_null"]:
+            try:
+                if df[field].notnull().all():
+                    return False
+            except KeyError:
+                return False
+
+        # TODO: add more conditions to check
+        return True
+
+    async def _apply_one(
+        self,
+        it: pandas.DataFrame,
+        field: str,
+        val: Any
+    ) -> None:
+        """_apply_one.
+
+        Processing "fields" attributes as a Dictionary.
+        Apply to one column only.
+        """
+        conditions = {}
+        if "conditions" in val:
+            conditions = val.get('conditions')
+            if not isinstance(conditions, dict):
+                self._logger.warning(
+                    f"TranformRows conditions need to be a dictionary, got {type(conditions)}"
+                )
+        if "value" in val:
+            operation = val.get('value')
+        else:
+            operation = val
+        args = {}
+        try:
+            fname = operation.pop(0)
+        except AttributeError:
+            fname = operation
+            operation = []
+        self._applied.append(f"Function: {fname!s} args: {operation}")
+        if not await self._check_condition(it, field, fname, conditions):
+            self._logger.warning(
+                f"Column {field} not met conditions {conditions}"
+            )
+            return it  # Skip transformation if condition fails
+        # fname is a field and len > 1
+        if fname in it.columns and len(operation) == 0:
+            # simple field replacement
+            it[field] = it[fname]
+            it = it.copy()
+            return it
+
+        # only calls functions on TransformRows scope:
+        # first, check if is a Pandas-based Function
+        try:
+            args = operation[0]
+        except IndexError:
+            args = {}
+        try:
+            func = self._get_df_function(fname)
+            self._logger.notice(
+                f"Calling Function: {fname!s} with args: {operation}"
+            )
+            if fname == "fill_column":
+                args["variables"] = self._variables
+            if args:
+                it = func(df=it, field=field, **args)
+            else:
+                it = func(df=it, field=field)
+            it = it.copy()
+            return it
+        except AttributeError:
+            pass
+        # try to apply an scalar function:
+        func = getFunction(fname)
+        if callable(func):
+            # SCALAR FUNCTION
+            try:
+                tmp = operation[0]
+                for a, b in tmp.items():
+                    if isinstance(b, list):
+                        for idx, el in enumerate(b):
+                            if el in self._variables:
+                                b[idx] = el.replace(
+                                    str(el),
+                                    str(self._variables[str(el)]),
+                                )
+                    if b in self._variables:
+                        args[a] = b.replace(
+                            str(b), str(self._variables[str(b)])
+                        )
+                result = func(**args)
+            except (KeyError, IndexError, ValueError):
+                result = func()
+            except Exception as err:
+                self._logger.warning(
+                    f"Error or missing DF Function: {err!s}"
+                )
+                # using scalar functions to set value in columns
+                func = getFunction(fname)
+                if not callable(func):
+                    return it
+                self._logger.debug(
+                    f"Calling Scalar: {fname!s}: {func}"
+                )
+                try:
+                    args = operation[0]
+                    tmp = operation[0]
+                except IndexError:
+                    args = {}
+                try:
+                    for a, b in tmp.items():
+                        if b in self._variables:
+                            args[a] = b.replace(
+                                str(b), str(self._variables[str(b)])
+                            )
+                    result = func(**args)
+                except (TypeError, KeyError, IndexError, ValueError):
+                    # result doesn't need keyword arguments
+                    result = func()
+                except Exception as e:
+                    print(func, fname, field, val)
+                    self._logger.exception(
+                        f"Error Running an Scalar Function {fname!s} \
+                        to set Dataframe: {e}"
+                    )
+                    return it
+            r = {field: result}
+            it = it.assign(**r)
+            it = it.copy()
+            return it
+        else:
+            self._logger.warning(
+                f"Function {func} is not callable."
+            )
+        return it
+
+    async def _apply_multiple(
+        self,
+        it: pandas.DataFrame,
+        field: str,
+        val: Any
+    ) -> None:
+        """_apply_multiple.
+
+        Apply Multiples functions to a single column.
+        """
+        if isinstance(val, dict):
+            for fn, args in val.items():
+                try:
+                    func = self._get_df_function(fn)
+                    self._logger.notice(
+                        f"Calling Function: {fn!s} with args: {args}"
+                    )
+                    if not args:
+                        args = {}
+                    it = func(df=it, field=field, **args)
+                    it = it.copy()
+                except AttributeError:
+                    pass
+            return it
+        for value in val:
+            # iterate over a list of possible functions:
+            if isinstance(value, str):
+                # create the column based on value
+                await self._scalar_column(it, field, value)
+                continue
+            if isinstance(value, dict):
+                # function dictionary: {'split': {'separator': ','}}
+                # TODO: for now, only work with dataframe functions:
+                for fn, args in value.items():
+                    try:
+                        func = self._get_df_function(fn)
+                        self._logger.notice(
+                            f"Calling Function: {fn!s} with args: {args}"
+                        )
+                        if not args:
+                            args = {}
+                        it = func(df=it, field=field, **args)
+                        it = it.copy()
+                    except AttributeError:
+                        pass
+            elif isinstance(value, list):
+                # list of functions to be applied:
+                # split the list element in fn and arguments
+                fn, args = value if len(value) == 2 else (value[0], {})
+                func = self._get_df_function(fn)
+                if func is not None:
+                    it = func(df=it, field=field, **args)
+                    it = it.copy()
+                else:
+                    self._logger.warning(
+                        f"Function {fn} is not callable."
+                    )
+            else:
+                raise ConfigError(
+                    f"Unable to work with this Function {value}: {type(value)}"
+                )
+        return it
+
+    async def transform(self, elem, df):
+        try:
+            fields = copy.deepcopy(elem.fields)
+        except KeyError:
+            fields = {}
+        if not isinstance(df, pandas.DataFrame):
+            raise NoDataFound("Pandas Dataframe Empty or is not a Dataframe")
+        if hasattr(elem, "clean_dates"):
+            u = df.select_dtypes(include=["datetime64[ns]"])
+            df[u.columns] = df[u.columns].replace({np.nan: None})
+        it = df.copy()
+        for field, val in fields.items():
+            if isinstance(val, str):
+                await self._scalar_column(it, field, val)
+                continue
+            elif isinstance(val, dict) and 'value' in val:
+                # Function to be applied:
+                it = await self._apply_one(it, field, val)
+                continue
+            elif isinstance(val, (dict, list)):
+                # Applying multiple functions to same column:
+                it = await self._apply_multiple(it, field, val)
+                continue
+        # at the end
+        df = it
+        # starting changes:
+        if hasattr(elem, "clean_str"):
+            df.is_copy = None
+            u = df.select_dtypes(include=["object", "string"])
+            u = u.applymap(lambda x: x.strip() if isinstance(x, str) else x)
+            df[u.columns] = df[u.columns].fillna("")
+        if hasattr(elem, "drop_empty"):
+            # First filter out those rows which
+            # does not contain any data
+            df.dropna(how="all")
+            # removing empty cols
+            df.is_copy = None
+            df.dropna(axis=1, how="all")
+            df.dropna(axis=0, how="all")
+        if self._debug is True:
+            print("TRANSFORM ===")
+            print(df)
+            print("::: Printing Column Information === ")
+            for column, t in df.dtypes.items():
+                print(column, "->", t, "->", df[column].iloc[0])
+        # avoid threat the Dataframe as a Copy
+        df.is_copy = None
+        self._result = df
+        try:
+            self.add_metric("ended_rows", df.shape[0])
+            self.add_metric("ended_columns", df.shape[1])
+            self.add_metric("Transformations", self._applied)
+        except Exception as err:
+            logging.error(f"TransformRows: Error setting Metrics: {err}")
+        return self._result
+
+    def close(self):
+        pass

flowtask/components/TransformRows/__init__.py

@@ -0,0 +1,9 @@
+"""
+TransformRows.
+
+TransformRows allow making column transformation over a Pandas Dataframe.
+"""
+
+from .TransformRows import TransformRows
+
+__all__ = ["TransformRows"]