dpkits

Data processing Package

• Requirements

  • pandas
  • pyreadstat
  • numpy
  • zipfile
  • fastapi[UploadFile]

• Step 1: import classes

  # Convert data to pandas dataframe
  from dpkits.ap_data_converter import APDataConverter
  
  # Calculate LSM score
  from dpkits.calculate_lsm import LSMCalculation
  
  # Transpose data to stack and untack
  from dpkits.data_transpose import DataTranspose
  
  # Create the tables from converted dataframe 
  from dpkits.table_generator import DataTableGenerator
  
  # Format data tables 
  from dpkits.table_formater import TableFormatter
  

• Step 2: Convert data files to dataframe

  • class APDataConverter(files=None, file_name='', is_qme=True)
    • input 1 of files or file_name
    • files: list[UploadFile] default = None
    • file_name: str default = ''
    • is_qme: bool default = True
    • Returns:
      • df_data: pandas.Dataframe
      • df_info: pandas.Dataframe

    # Define input/output files name
    str_file_name = 'APDataTest'
    str_tbl_file_name = f'{str_file_name}_Topline.xlsx'
    
    converter = APDataConverter(file_name='APDataTesting.xlsx')
    
    df_data, df_info = converter.convert_df_mc() 
    
    # Use 'converter.convert_df_md()' if you need md data
    

• Step 3: Calculate LSM classificate (only for LSM projects)

  • class LSMCalculation.cal_lsm_6(df_data, df_info)
    • df_data: pandas.Dataframe
    • df_info: pandas.Dataframe
    • Returns:
      • df_data: pandas.Dataframe
      • df_info: pandas.Dataframe

    df_data, df_info = LSMCalculation.cal_lsm_6(df_data, df_info)
    
    # df_data, df_info will contains the columns CC1_Score to CC6_Score & LSM_Score
    

• Step 4: Data cleaning (if needed)

  # Use pandas's functions to clean/process data
  
  df_data['Gender_new'] = df_data['Gender']
  
  df_data.replace({
      'Q1_SP1': {1: 5, 2: 4, 3: 3, 4: 2, 5: 1},
      'Q1_SP2': {1: 5, 2: 4, 3: 3, 4: 2, 5: 1},
  }, inplace=True)
  
  df_data.loc[(df_data['Gender_new'] == 2) & (df_data['Age'] == 5),  ['Gender_new']] = [np.nan]
  df_info.loc[df_info['var_name'] == 'Q1_SP1', ['val_lbl']] = [{'1': 'a', '2': 'b', '3': 'c', '4': 'd', '5': 'e'}]
  
  df_info = pd.concat([df_info, pd.DataFrame(
      columns=['var_name', 'var_lbl', 'var_type', 'val_lbl'],
      data=[
          ['Gender_new', 'Please indicate your gender', 'SA', {'1': 'aaa', '2': 'bb', '3': 'cc'}]
      ]
  )], ignore_index=True)
  

• Step 5: Transpose data (if needed)

  • class DataTranspose.to_stack(df_data, df_info, dict_stack_structure)
    • df_data: pandas.Dataframe
    • df_info: pandas.Dataframe
    • dict_stack_structure: dict
    • Returns:
      • df_data_stack: pandas.Dataframe
      • df_info_stack: pandas.Dataframe

    dict_stack_structure = {
        'id_col': 'ResID',
        'sp_col': 'Ma_SP',
        'lst_scr': ['Gender', 'Age', 'City', 'HHI'],
        'dict_sp': {
            1: {
                'Ma_SP1': 'Ma_SP',
                'Q1_SP1': 'Q1',
                'Q2_SP1': 'Q2',
                'Q3_SP1': 'Q3',
            },
            2: {
                'Ma_SP2': 'Ma_SP',
                'Q1_SP2': 'Q1',
                'Q2_SP2': 'Q2',
                'Q3_SP2': 'Q3',
            },
        },
        'lst_fc': ['Awareness1', 'Frequency', 'Awareness2', 'Perception']
    }
    
    df_data_stack, df_info_stack = DataTranspose.to_stack(df_data, df_info, dict_stack_structure)
    

  • class DataTranspose.to_unstack(df_data_stack, df_info_stack, dict_unstack_structure)
    • df_data_stack: pandas.Dataframe which transpose from stack
    • df_info_stack: pandas.Dataframe which transpose from stack
    • dict_unstack_structure: dict
    • Returns:
      • df_data_unstack: pandas.Dataframe
      • df_info_unstack: pandas.Dataframe

    dict_unstack_structure = {
        'id_col': 'ResID',
        'sp_col': 'Ma_SP',
        'lst_col_part_head': ['Gender', 'Age', 'City', 'HHI'],
        'lst_col_part_body': ['Q1', 'Q2', 'Q3'],
        'lst_col_part_tail': ['Awareness1', 'Frequency', 'Awareness2', 'Perception']
    }
    
    df_data_unstack, df_info_unstack = DataTranspose.to_unstack(df_data_stack, df_info_stack, dict_unstack_structure)
    

• Step 6: OE Running

• Step 7: Export *.sav & *.xlsx

  • class converter.generate_multiple_data_files(dict_dfs=dict_dfs, is_md=False, is_export_sav=True, is_export_xlsx=True, is_zip=True)
    • df_data: pandas.Dataframe
      • dict_dfs: dict
      • is_md: bool default False
      • is_export_sav: bool default True
      • is_export_xlsx: bool default True
      • is_zip: bool default True
      • Returns: NONE

      dict_dfs = {
          1: {
              'data': df_data,
              'info': df_info,
              'tail_name': 'ByCode',
              'sheet_name': 'ByCode',
              'is_recode_to_lbl': False,
          },
          2: {
              'data': df_data,
              'info': df_info,
              'tail_name': 'ByLabel',
              'sheet_name': 'ByLabel',
              'is_recode_to_lbl': True,
          },
          3: {
              'data': df_data_stack,
              'info': df_info_stack,
              'tail_name': 'Stack',
              'sheet_name': 'Stack',
              'is_recode_to_lbl': False,
          },
          4: {
              'data': df_data_unstack,
              'info': df_info_unstack,
              'tail_name': 'Unstack',
              'sheet_name': 'Unstack',
              'is_recode_to_lbl': False,
          },
      }
      
      converter.generate_multiple_data_files(dict_dfs=dict_dfs, is_md=False, is_export_sav=True, is_export_xlsx=True, is_zip=True)
      

• Step 8: Export data tables

  • init DataTableGenerator(df_data=df_data, df_info=df_info, xlsx_name=str_tbl_file_name)
    • df_data: pandas.Dataframe
    • df_info: pandas.Dataframe
    • xlsx_name: str
    • Returns: NONE
  • class DataTableGenerator.run_tables_by_js_files(lst_func_to_run)
    • lst_func_to_run: list
    • Returns: NONE
  • init TableFormatter(xlsx_name=str_tbl_file_name)
    • xlsx_name: str
    • Returns: NONE
  • class TableFormatter.format_sig_table()
    • Returns: NONE

  lst_side_qres = [
      {"qre_name": "CC1", "sort": "des"},
      {"qre_name": "$CC3", "sort": "asc"},
      {"qre_name": "$CC4", "sort": "des"},
      {"qre_name": "$CC6"},
      {"qre_name": "$CC10"},
      {"qre_name": "LSM"},
      {"qre_name": "Gender"},
      {"qre_name": "Age"},
      {"qre_name": "City"},
      {"qre_name": "HHI"},
      
      # MA Question with net/combine (can apply to SA questions)
      {"qre_name": "$Q15", "cats": {
          'net_code': {
              '900001|combine|Group 1 + 2': {
                  '1': 'Yellow/dull teeth',
                  '3': 'Dental plaque',
                  '5': 'Bad breath',
                  '7': 'Aphthousulcer',
                  '2': 'Sensitive teeth',
                  '4': 'Caries',
                  '6': 'Gingivitis (bleeding, swollen gums)',
              },
              '900002|net|Group 1': {
                  '1': 'Yellow/dull teeth',
                  '3': 'Dental plaque',
                  '5': 'Bad breath',
                  '7': 'Aphthousulcer',
              },
              '900003|net|Group 2': {
                  '2': 'Sensitive teeth',
                  '4': 'Caries',
                  '6': 'Gingivitis (bleeding, swollen gums)',
              },
          },
          '8': 'Other (specify)',
          '9': 'No problem',
      }},
  
      # Scale question with full properties
      {
          "qre_name": "Perception",
          "cats": {
              '1': 'Totally disagree', '2': 'Disagree', '3': 'Neutral', '4': 'Agree', '5': 'Totally agree',
              'net_code': {
                  '900001|combine|B2B': {'1': 'Totally disagree', '2': 'Disagree'},
                  '900002|combine|Medium': {'3': 'Neutral'},
                  '900003|combine|T2B': {'4': 'Agree', '5': 'Totally agree'},
              }
          },
          "mean": {1: 1, 2: 2, 3: 3, 4: 4, 5: 5}
      },
  ]
  
  lst_header_qres = [
      [
          {
              "qre_name": "Age",
              "qre_lbl": "Age",
              "cats": {
                  'TOTAL': 'TOTAL',
                  '2': '18 - 24', '3': '25 - 30', '4': '31 - 39', '5': '40 - 50', '6': 'Trên 50'
              }
          },
          {
              "qre_name": "@City2",
              "qre_lbl": "Location",
              "cats": {
                  'City.isin([1, 5, 10, 11, 12])': 'All South',
                  'City.isin([2, 4, 16, 17, 18])': 'All North',
              }
          },
      ],
  ]
  
  lst_func_to_run = [
      {
          'func_name': 'run_standard_table_sig',
          'tables_to_run': [
              'Tbl_1_Pct',  # this table use df_data & df_info to run
              'Tbl_1_Count',  # this table use df_data & df_info to run
          ],
          'tables_format': {
  
              "Tbl_1_Pct": {
                  "tbl_name": "Table 1 - Pct",
                  "tbl_filter": "City > 0",
                  "is_count": 0,
                  "is_pct_sign": 1,
                  "is_hide_oe_zero_cats": 1,
                  "sig_test_info": {
                      "sig_type": "",  # ind / rel
                      "sig_cols": [],
                      "lst_sig_lvl": []
                  },
                  "lst_side_qres": lst_side_qres,
                  "lst_header_qres": lst_header_qres
              },
  
              "Tbl_1_Count": {
                  "tbl_name": "Table 1 - Count",
                  "tbl_filter": "City > 0",
                  "is_count": 1,
                  "is_pct_sign": 0,
                  "is_hide_oe_zero_cats": 1,
                  "sig_test_info": {
                      "sig_type": "",
                      "sig_cols": [],
                      "lst_sig_lvl": []
                  },
                  "lst_side_qres": lst_side_qres,
                  "lst_header_qres": lst_header_qres
              },
          },
  
      },
  ]
  
  dtg = DataTableGenerator(df_data=df_data, df_info=df_info, xlsx_name=str_tbl_file_name)
  dtg.run_tables_by_js_files(lst_func_to_run)
  
  dtf = TableFormatter(xlsx_name=str_tbl_file_name)
  dtf.format_sig_table()
  

This is a simple example package. You can use Github-flavored Markdown to write your content.