Add Scanner info cover sheet into xlsx (#156)

Signed-off-by: jiyeong.seok <[email protected]>

Author: dd-jy

src/fosslight_util/cover.py

@@ -0,0 +1,44 @@
+#!/usr/bin/env python
+# -*- coding: utf-8 -*-
+# Copyright (c) 2024 LG Electronics Inc.
+# SPDX-License-Identifier: Apache-2.0
+
+import os
+import sys
+from fosslight_util.help import print_package_version
+
+
+class CoverItem:
+    tool_name_key = "Tool name"
+    tool_version_key = "Tool version"
+    start_time_key = "Start time"
+    python_ver_key = "Python version"
+    analyzed_path_key = "Analyzed path"
+    comment_key = "Comment"
+
+    def __init__(self, tool_name="", start_time="", input_path="", comment=""):
+        self.tool_name = tool_name
+        if start_time:
+            date, time = start_time.split('_')
+            self.start_time = f'{date}, {time[0:2]}:{time[2:4]}'
+        else:
+            self.start_time = ""
+        self.input_path = os.path.abspath(input_path)
+        self.comment = comment
+
+        self.tool_version = print_package_version(self.tool_name, "", False)
+        self.python_version = f'{sys.version_info.major}.{sys.version_info.minor}'
+
+    def __del__(self):
+        pass
+
+    def get_print_json(self):
+        json_item = {}
+        json_item[self.tool_name_key] = self.tool_name
+        json_item[self.tool_version_key] = self.tool_version
+        json_item[self.start_time_key] = self.start_time
+        json_item[self.python_ver_key] = self.python_version
+        json_item[self.analyzed_path_key] = self.input_path
+        json_item[self.comment_key] = self.comment
+
+        return json_item

src/fosslight_util/help.py

@@ -52,10 +52,12 @@ def print_package_version(pkg_name, msg="", exitopt=True):
     if msg == "":
         msg = f"{pkg_name} Version:"
     cur_version = pkg_resources.get_distribution(pkg_name).version
-    print(f'{msg} {cur_version}')
 
     if exitopt:
+        print(f'{msg} {cur_version}')
         sys.exit(0)
+    else:
+        return cur_version
 
 
 def print_help_msg_download(exitOpt=True):

src/fosslight_util/output_format.py

@@ -3,7 +3,7 @@
 # Copyright (c) 2021 LG Electronics Inc.
 # SPDX-License-Identifier: Apache-2.0
 import os
-from fosslight_util.write_excel import write_result_to_excel, write_result_to_csv, remove_empty_sheet
+from fosslight_util.write_excel import write_result_to_excel, write_result_to_csv
 from fosslight_util.write_opossum import write_opossum
 from fosslight_util.write_yaml import write_yaml
 
@@ -56,29 +56,24 @@ def check_output_format(output='', format='', customized_format={}):
     return success, msg, output_path, output_file, output_extension
 
 
-def write_output_file(output_file_without_ext, file_extension, sheet_list, extended_header={}, hide_header={}):
+def write_output_file(output_file_without_ext, file_extension, sheet_list, extended_header={}, hide_header={}, cover=""):
     success = True
     msg = ''
 
-    is_not_null, sheet_list = remove_empty_sheet(sheet_list)
-    if is_not_null:
-        if file_extension == '':
-            file_extension = '.xlsx'
-        result_file = output_file_without_ext + file_extension
+    if file_extension == '':
+        file_extension = '.xlsx'
+    result_file = output_file_without_ext + file_extension
 
-        if file_extension == '.xlsx':
-            success, msg = write_result_to_excel(result_file, sheet_list, extended_header, hide_header)
-        elif file_extension == '.csv':
-            success, msg, result_file = write_result_to_csv(result_file, sheet_list)
-        elif file_extension == '.json':
-            success, msg = write_opossum(result_file, sheet_list)
-        elif file_extension == '.yaml':
-            success, msg, result_file = write_yaml(result_file, sheet_list, False)
-        else:
-            success = False
-            msg = f'Not supported file extension({file_extension})'
+    if file_extension == '.xlsx':
+        success, msg = write_result_to_excel(result_file, sheet_list, extended_header, hide_header, cover)
+    elif file_extension == '.csv':
+        success, msg, result_file = write_result_to_csv(result_file, sheet_list)
+    elif file_extension == '.json':
+        success, msg = write_opossum(result_file, sheet_list)
+    elif file_extension == '.yaml':
+        success, msg, result_file = write_yaml(result_file, sheet_list, False)
     else:
-        result_file = ""
-        msg = "Nothing is detected from the scanner so output file is not generated."
+        success = False
+        msg = f'Not supported file extension({file_extension})'
 
     return success, msg, result_file

src/fosslight_util/write_excel.py

@@ -13,6 +13,7 @@
 from pathlib import Path
 import fosslight_util.constant as constant
 from jsonmerge import merge
+from fosslight_util.cover import CoverItem
 
 _HEADER = {'BIN (': ['ID', 'Binary Name', 'Source Code Path',
                      'NOTICE.html', 'OSS Name', 'OSS Version',
@@ -31,6 +32,7 @@
 IDX_FILE = 0
 IDX_EXCLUDE = 7
 logger = logging.getLogger(constant.LOGGER_NAME)
+COVER_SHEET_NAME = 'Scanner Info'
 
 
 def write_excel_and_csv(filename_without_extension, sheet_list, ignore_os=False, extended_header={}, hide_header={}):
@@ -105,9 +107,9 @@ def get_header_row(sheet_name, sheet_content, extended_header={}):
             if sheet_name.startswith(header_key):
                 selected_header = merged_headers[header_key]
                 break
-
-    if not selected_header:
-        selected_header = sheet_content.pop(0)
+    if len(sheet_content) > 0:
+        if not selected_header:
+            selected_header = sheet_content.pop(0)
     return selected_header, sheet_content
 
 
@@ -166,16 +168,18 @@ def write_result_to_csv(output_file, sheet_list_origin, separate_sheet=False, ex
     return success, error_msg, output
 
 
-def write_result_to_excel(out_file_name, sheet_list, extended_header={}, hide_header={}):
+def write_result_to_excel(out_file_name, sheet_list, extended_header={}, hide_header={}, cover=""):
     success = True
     error_msg = ""
 
     try:
-        if sheet_list:
-            output_dir = os.path.dirname(out_file_name)
-            Path(output_dir).mkdir(parents=True, exist_ok=True)
+        output_dir = os.path.dirname(out_file_name)
+        Path(output_dir).mkdir(parents=True, exist_ok=True)
 
-            workbook = xlsxwriter.Workbook(out_file_name)
+        workbook = xlsxwriter.Workbook(out_file_name)
+        if cover:
+            write_cover_sheet(workbook, cover)
+        if sheet_list:
             for sheet_name, sheet_contents in sheet_list.items():
                 selected_header, sheet_content_without_header = get_header_row(sheet_name, sheet_contents[:], extended_header)
                 try:
@@ -188,13 +192,33 @@ def write_result_to_excel(out_file_name, sheet_list, extended_header={}, hide_he
 
                 if hide_header:
                     hide_column(worksheet, selected_header, hide_header)
-            workbook.close()
+        workbook.close()
     except Exception as ex:
         error_msg = str(ex)
         success = False
     return success, error_msg
 
 
+def write_cover_sheet(workbook, cover):
+    worksheet = workbook.add_worksheet(COVER_SHEET_NAME)
+
+    format_bold = workbook.add_format({'bold': True})
+    worksheet.merge_range('A1:B1', 'About the scanner', format_bold)
+
+    key_format = workbook.add_format({'bold': True, 'font_color': 'white', 'bg_color': 'navy'})
+    item_format = workbook.add_format()
+    item_format.set_text_wrap()
+
+    cover_json = cover.get_print_json()
+    row = 1
+    for item in cover_json:
+        worksheet.write(row, 0, item, key_format)
+        worksheet.write(row, 1, cover_json[item], item_format)
+        row += 1
+    worksheet.set_column(0, 0, 30)
+    worksheet.set_column(1, 1, 100)
+
+
 def write_result_to_sheet(worksheet, sheet_contents):
     row = 1
     for row_item in sheet_contents:
@@ -221,7 +245,33 @@ def create_worksheet(workbook, sheet_name, header_row):
     return worksheet
 
 
-def merge_excels(find_excel_dir, final_out, merge_files=''):
+def merge_cover_comment(find_excel_dir, merge_files=''):
+    FIND_EXTENSION = '.xlsx'
+    merge_comment = []
+    cover_comment = ''
+    try:
+        files = os.listdir(find_excel_dir)
+
+        if len([name for name in files if name.endswith(FIND_EXTENSION)]) > 0:
+            for file in files:
+                if merge_files:
+                    if file not in merge_files:
+                        continue
+                if file.endswith(FIND_EXTENSION):
+                    file = os.path.join(find_excel_dir, file)
+                    df_excel = pd.read_excel(file, sheet_name=COVER_SHEET_NAME, index_col=0, engine='openpyxl')
+                    if not df_excel.empty:
+                        tool_name = df_excel.loc[CoverItem.tool_name_key].values[0]
+                        comment = df_excel.loc[CoverItem.comment_key].values[0]
+                        merge_comment.append(str(f"[{tool_name}] {comment}"))
+            cover_comment = '\n'.join(merge_comment)
+    except Exception as ex:
+        logger.warning(f'Fail to merge comment of Scanner info: {str(ex)}')
+
+    return cover_comment
+
+
+def merge_excels(find_excel_dir, final_out, merge_files='', cover=''):
     success = True
     msg = ""
     FIND_EXTENSION = '.xlsx'
@@ -231,7 +281,7 @@ def merge_excels(find_excel_dir, final_out, merge_files=''):
 
         if len([name for name in files if name.endswith(FIND_EXTENSION)]) > 0:
             writer = pd.ExcelWriter(final_out)
-
+            write_cover_sheet(writer.book, cover)
             for file in files:
                 if merge_files:
                     if file not in merge_files:
@@ -243,13 +293,14 @@ def merge_excels(find_excel_dir, final_out, merge_files=''):
                     excel_file = pd.ExcelFile(file, engine='openpyxl')
 
                     for sheet_name in excel_file.sheet_names:
-                        sheet_name_to_copy = f"{f_short_name}_{sheet_name}"
+                        if sheet_name == COVER_SHEET_NAME:
+                            continue
                         df_excel = pd.read_excel(
                             file, sheet_name=sheet_name, engine='openpyxl')
-                        if sheet_name not in added_sheet_names:
-                            sheet_name_to_copy = sheet_name
-                        df_excel.to_excel(writer, sheet_name_to_copy,
-                                          index=False)
+                        if sheet_name in added_sheet_names:
+                            sheet_name = f"{f_short_name}_{sheet_name}"
+                        df_excel.to_excel(writer, sheet_name, index=False)
+                        added_sheet_names.append(sheet_name)
             writer.close()
     except Exception as ex:
         msg = str(ex)

tests/test_excel_and_csv.py

@@ -50,7 +50,6 @@ def main():
     sheet_contents['SRC'] = src_sheet_items
     sheet_contents['BIN_TEST'] = bin_sheet_items
     sheet_contents['SRC_NULL'] = []
-    sheet_contents['NULL_SHEET'] = []
     sheet_contents['CUSTOM_HEADER_SHEET'] = sheet_items
 
     logger.warning("TESTING - Writing an excel and csv")