`recommend` command fails when `xml` format is specified

[AntonProkopyev]

Aibolit fails when xml report is requested.

$ python -m aibolit recommend --folder test/integration/check --format=xml --full
Version 1.3.0 is available, but you are using 0.0.0
ignore: []
Traceback (most recent call last):
  ...
  File "/home/me/aibolit/aibolit/__main__.py", line 536, in _process_file_result
    if pattern.get('pattern_code'):
       ^^^^^^^^^^^
AttributeError: 'list' object has no attribute 'get'

Test to reproduce the bug

diff --git a/test/recommend/test_recommend_pipeline.py b/test/recommend/test_recommend_pipeline.py
index c95986db..6e444249 100644
--- a/test/recommend/test_recommend_pipeline.py
+++ b/test/recommend/test_recommend_pipeline.py
@@ -156,6 +156,11 @@ class TestRecommendPipeline(TestCase):
         mock_cmd = self.__create_mock_cmd()
         create_xml_tree([], full_report=True, cmd=mock_cmd, exit_code=0)
 
+    def test_xml(self):
+        mock_input = self.__create_mock_input()
+        mock_cmd = self.__create_mock_cmd()
+        create_xml_tree(mock_input, cmd=mock_cmd, exit_code=2)
+
     def test_text_format(self):
         mock_input = self.__create_mock_input()
         new_mock = format_converter_for_pattern(mock_input)

[0crat]

@AntonProkopyev Thanks for reporting a new bug! You've earned +12 points. By reporting bugs, you help our project improve its quality. If you find anything else in the repository that doesn't look as good as you might expect, do not hesitate to report it. Your running balance is +84.

[agonychaser]

Hello Yegor, I'd like to contribute to this issue. As I understand, the problem is that _process_file_result method assumes that the result of components analysis looks like the following:

[
    {'pattern_code': 'P12', ...},
    {'pattern_code': 'P21', ...},
    {'pattern_code': 'P12', ...}
]

Whereas actually it is a nested list:

[[{'pattern_code': 'P23', 'pattern_name': 'Some patterns name', 'code_lines': [1, 2, 4], 'importance': 0.1}]]

So I think that we just should flatten the nested list beforehand:

flat_results = [item for sublist in result_for_file['results'] for item in sublist]
for i, pattern in enumerate(flat_results, start=1):
    if pattern.get('pattern_code'):
        pattern_score = _process_pattern(patterns_tag, pattern, i, patterns_number)
        importance_for_class.append(pattern_score) 

Would like to hear your opinion!

[agonychaser]

Well, I think that this is not the Yegor's duty :) @AntonProkopyev

[AntonProkopyev]

Hello Yegor, I'd like to contribute to this issue. As I understand, the problem is that _process_file_result method assumes that the result of components analysis looks like the following:

[
    {'pattern_code': 'P12', ...},
    {'pattern_code': 'P21', ...},
    {'pattern_code': 'P12', ...}
]

Whereas actually it is a nested list:

[[{'pattern_code': 'P23', 'pattern_name': 'Some patterns name', 'code_lines': [1, 2, 4], 'importance': 0.1}]]

So I think that we just should flatten the nested list beforehand:

flat_results = [item for sublist in result_for_file['results'] for item in sublist]
for i, pattern in enumerate(flat_results, start=1):
    if pattern.get('pattern_code'):
        pattern_score = _process_pattern(patterns_tag, pattern, i, patterns_number)
        importance_for_class.append(pattern_score) 

Would like to hear your opinion!

@onemoreslacker Hello! Totally agree, you could try this fix, just provide PR and e2e test to ensure the issue was fixed.