Merge pull request #14 from dataiku/feature/sc-230874-customize-tab-names

feat: [sc-230874] [Multisheet Excel Export] Customize tab names

Author: alexbourret

CHANGELOG.md

@@ -1,5 +1,9 @@
 # Changelog
 
+## [Version 2.2.0](https://github.com/dataiku/dss-plugin-multisheet-excel-export/releases/tag/v2.2.0) - Major release - 2025-02
+
+- Add customization of tab names
+
 ## [Version 2.1.0](https://github.com/dataiku/dss-plugin-multisheet-excel-export/releases/tag/v2.1.0) - Major release - 2024-09
 - Bug fix: one temporary workbook is used per dataset to avoid out of memory issues while exporting large datasets. All these temporary workbooks are merged at the end to generate the final excel file
 - Optimizations: using of a cache for styles to avoid useless copies + openpyxl write only mode with lxml

custom-recipes/to-excel/recipe.json

@@ -45,6 +45,34 @@
             "type": "BOOLEAN",
             "defaultValue": false,
             "mandatory": true
+        },
+        {
+            "name": "renaming_sheets",
+            "label": "Renaming the sheets",
+            "description": "Name the sheets rather than use the datasets names",
+            "type": "BOOLEAN",
+            "defaultValue": false
+        },
+        {
+            "name": "dataset_to_sheet_mapping",
+            "label": "Sheet name",
+            "description": "",
+            "type": "OBJECT_LIST",
+            "visibilityCondition": "model.renaming_sheets == true",
+            "subParams": [
+                {
+                    "name": "dataset_name",
+                    "type": "DATASET",
+                    "label": "Dataset",
+                    "description": "To be valid, the dataset has to be selected in the recipe's input"
+                },
+                {
+                    "name": "sheet_name",
+                    "type": "STRING",
+                    "label": "Sheet name",
+                    "description": "Name of the targeted sheet for this dataset"
+                }
+            ]
         }
     ],
     "resourceKeys" : []

custom-recipes/to-excel/recipe.py

@@ -14,7 +14,7 @@
 from dataiku.customrecipe import get_output_names_for_role
 from dataiku.customrecipe import get_recipe_config
 from openpyxl import load_workbook, Workbook
-from xlsx_writer import datasets_to_xlsx
+from xlsx_writer import datasets_to_xlsx, assert_valid_sheet_name
 from typing import Union
 
 DEFAULT_DATAIKU_SHEET_NAME = "Sheet1"
@@ -50,6 +50,20 @@ def get_excel_worksheet(dataset: dataiku.Dataset, apply_conditional_formatting:
     return None
 
 
+def get_dataset_to_sheet_mapping(config):
+    renaming_sheets = config.get("renaming_sheets", False)
+    dataset_to_sheet_mapping = {}
+    if renaming_sheets:
+        dataset_to_sheet_mappings = config.get("dataset_to_sheet_mapping", {})
+        for mapping in dataset_to_sheet_mappings:
+            dataset_name = mapping.get("dataset_name")
+            sheet_name = mapping.get("sheet_name")
+            assert_valid_sheet_name(sheet_name)
+            dataset_to_sheet_mapping[dataset_name] = sheet_name
+            logger.info("Renaming dataset '{}' into sheet '{}'".format(dataset_name, sheet_name))
+    return dataset_to_sheet_mapping
+
+
 logger = logging.getLogger(__name__)
 logging.basicConfig(level=logging.INFO, format='Multi-Sheet Excel Exporter | %(levelname)s - %(message)s')
 
@@ -72,6 +86,7 @@ def get_excel_worksheet(dataset: dataiku.Dataset, apply_conditional_formatting:
 input_config = get_recipe_config()
 workbook_name = input_config.get('output_workbook_name', None)
 apply_conditional_formatting = input_config.get('export_conditional_formatting', False)
+dataset_to_sheet_mapping = get_dataset_to_sheet_mapping(input_config)
 
 if workbook_name is None:
     logger.warning("Received input received recipe config: {}".format(input_config))
@@ -89,7 +104,12 @@ def get_excel_worksheet(dataset: dataiku.Dataset, apply_conditional_formatting:
     tmp_file_path = tmp_file.name
     logger.info("Intend to write the output xls file to the following location: {}".format(tmp_file_path))
 
-    datasets_to_xlsx(input_datasets_names, tmp_file_path, lambda name: get_excel_worksheet(dataiku.Dataset(name), apply_conditional_formatting))
+    datasets_to_xlsx(
+        input_datasets_names,
+        tmp_file_path,
+        lambda name: get_excel_worksheet(dataiku.Dataset(name), apply_conditional_formatting),
+        dataset_to_sheet_mapping=dataset_to_sheet_mapping
+    )
 
     with open(tmp_file_path, 'rb', encoding=None) as f:
         output_folder.upload_stream(output_file_name, f)

plugin.json

@@ -1,6 +1,6 @@
 {
     "id" : "multisheet-excel-export",
-    "version" : "2.1.0",
+    "version" : "2.2.0",
 
 
     "meta" : {

python-lib/xlsx_writer.py

@@ -171,7 +171,7 @@ def copy_sheet_to_workbook(source_sheet: Worksheet, target_workbook: Workbook) -
     return target_sheet
 
 
-def rename_too_long_dataset_names(input_dataset_names: List[str]) -> Dict[str, str]:
+def rename_too_long_dataset_names(input_dataset_names: List[str], dataset_to_sheet_mapping={}) -> Dict[str, str]:
     """
     Excel allows for only maximum 30 chars in the sheet names, so if some DS have more than 30 chars :
         - truncate the name to 28 chars
@@ -184,7 +184,9 @@ def rename_too_long_dataset_names(input_dataset_names: List[str]) -> Dict[str, s
     index_rename = -1
     renaming_length = EXCEL_MAX_LEN_SHEET_NAME - 2
     for name in input_dataset_names:
-        if len(name) > EXCEL_MAX_LEN_SHEET_NAME:
+        if name in dataset_to_sheet_mapping:
+            return_map[name] = dataset_to_sheet_mapping.get(name)
+        elif len(name) > EXCEL_MAX_LEN_SHEET_NAME:
             index_rename += 1
             rename = f"{name[0:renaming_length]}{index_rename:02d}"
             # Almost impossible case : a DS already has this name
@@ -200,7 +202,7 @@ def rename_too_long_dataset_names(input_dataset_names: List[str]) -> Dict[str, s
     return return_map
 
 
-def datasets_to_xlsx(input_dataset_names, xlsx_abs_path, worksheet_provider):
+def datasets_to_xlsx(input_dataset_names, xlsx_abs_path, worksheet_provider, dataset_to_sheet_mapping={}):
     """
     Write each input dataset into one temporary excel file and merge all these excel files into the final excel file
     :param input_dataset_names: the list of dataset, using one temporary workbook per dataset
@@ -210,7 +212,7 @@ def datasets_to_xlsx(input_dataset_names, xlsx_abs_path, worksheet_provider):
 
     logger.info(f"Building output excel file '{xlsx_abs_path}'...")
 
-    template_workbook, workbook_tmp_files = get_temporary_workbooks(input_dataset_names, worksheet_provider)
+    template_workbook, workbook_tmp_files = get_temporary_workbooks(input_dataset_names, worksheet_provider, dataset_to_sheet_mapping=dataset_to_sheet_mapping)
 
     # Save template workbook with styles and unzip it
     template_workbook_extract_dir = get_template_workbook_directory(template_workbook)
@@ -227,7 +229,7 @@ def datasets_to_xlsx(input_dataset_names, xlsx_abs_path, worksheet_provider):
     logger.info("Done writing output xlsx file.")
 
 
-def get_temporary_workbooks(input_dataset_names, worksheet_provider):
+def get_temporary_workbooks(input_dataset_names, worksheet_provider, dataset_to_sheet_mapping={}):
     """
     Create a template workbook and one temporary workbook per dataset stored on disk
     :param input_dataset_names: the list of dataset, using one temporary workbook per dataset
@@ -243,7 +245,7 @@ def get_temporary_workbooks(input_dataset_names, worksheet_provider):
     # List containing all temporary workbooks generated from dataset
     workbook_tmp_files = []
 
-    renaming_map = rename_too_long_dataset_names(input_dataset_names)
+    renaming_map = rename_too_long_dataset_names(input_dataset_names, dataset_to_sheet_mapping=dataset_to_sheet_mapping)
 
     for name in input_dataset_names:
         dataset_worksheet = worksheet_provider(name)
@@ -368,3 +370,8 @@ def add_style_if_not_exist(style, list):
                                                                                                               len(fills),
                                                                                                               len(alignments),
                                                                                                               len(number_formats)))
+
+
+def assert_valid_sheet_name(sheet_name):
+    if sheet_name is not None and len(sheet_name) > EXCEL_MAX_LEN_SHEET_NAME:
+        raise Exception("The sheet name '{}' is too long. Maximum is {} characters".format(sheet_name, EXCEL_MAX_LEN_SHEET_NAME))