Merge pull request #120 from arnav-mandal1234/stats

Adds Stat class and calculates additional statistics

Author: steven-esser

src/deltacode/__init__.py

@@ -52,12 +52,14 @@ def __init__(self, new_path, old_path, options):
         self.options = options
         self.deltas = []
         self.errors = []
+        self.stats = Stat(self.new.files_count, self.old.files_count)
 
         if self.new.path != '' and self.old.path != '':
             self.determine_delta()
             self.determine_moved()
             self.license_diff()
             self.copyright_diff()
+            self.stats.calculate_stats()
             # Sort deltas by score, descending, i.e., high > low, and then by
             # factors, alphabetically.  Run the least significant sort first.
             self.deltas.sort(key=lambda Delta: Delta.factors, reverse=False)
@@ -106,6 +108,7 @@ def determine_delta(self):
                 except KeyError:
                     delta = Delta(100, new_file, None)
                     delta.status = 'added'
+                    self.stats.num_added += 1
                     self.deltas.append(delta)
                     continue
 
@@ -117,11 +120,13 @@ def determine_delta(self):
                     if new_file.sha1 == f.sha1:
                         delta = Delta(0, new_file, f)
                         delta.status = 'unmodified'
+                        self.stats.num_unmodified += 1
                         self.deltas.append(delta)
                         continue
                     else:
                         delta = Delta(20, new_file, f)
                         delta.status = 'modified'
+                        self.stats.num_modified += 1
                         self.deltas.append(delta)
 
         # now time to find the added.
@@ -138,6 +143,7 @@ def determine_delta(self):
                 except KeyError:
                     delta = Delta(0, None, old_file)
                     delta.status = 'removed'
+                    self.stats.num_removed += 1
                     self.deltas.append(delta)
                     continue
 
@@ -181,6 +187,9 @@ def update_deltas(self, added, removed):
         """
         delta = Delta(0, added.new_file, removed.old_file)
         delta.status = 'moved'
+        self.stats.num_moved += 1
+        self.stats.num_added -= 1
+        self.stats.num_removed -= 1
         self.deltas.append(delta)
         self.deltas.remove(added)
         self.deltas.remove(removed)
@@ -312,3 +321,48 @@ def to_dict(self):
             ('new', new_file),
             ('old', old_file),
         ])
+
+class Stat(object):
+    """
+    Contains all the stats for the file changes in the new directory
+    with respect to the old directory.
+    """
+    def __init__(self, new_files_count, old_files_count):
+        self.new_files_count = new_files_count
+        self.old_files_count = old_files_count
+        self.num_added = 0
+        self.num_removed = 0
+        self.num_moved = 0
+        self.num_modified = 0
+        self.num_unmodified = 0
+        self.percent_added = 0
+        self.percent_removed = 0
+        self.percent_moved = 0
+        self.percent_modified = 0
+        self.percent_unmodified = 0
+
+    def calculate_stats(self):
+        """
+        Calculates the percentage change of new directory with respect to
+        the old directory.
+        """
+        self.percent_added = utils.calculate_percent(self.num_added, self.old_files_count)
+        self.percent_removed = utils.calculate_percent(self.num_removed, self.old_files_count)
+        self.percent_moved = utils.calculate_percent(self.num_moved, self.old_files_count)
+        self.percent_modified = utils.calculate_percent(self.num_modified, self.old_files_count)
+        self.percent_unmodified = utils.calculate_percent(self.num_unmodified, self.old_files_count)
+
+    def to_dict(self):
+        """
+        Return an OrderedDict comprising all the percent attributes of the object.
+        """
+        return OrderedDict([
+            ('old_files_count', self.old_files_count),
+            ('new_files_count', self.new_files_count),
+            ('percent_added', self.percent_added),
+            ('percent_removed', self.percent_removed),
+            ('percent_moved', self.percent_moved),
+            ('percent_modified', self.percent_modified),
+            ('percent_unmodified', self.percent_unmodified),
+        ])
+

src/deltacode/cli.py

@@ -48,6 +48,7 @@ def write_json(deltacode, outfile, all_delta_types=False):
         ('deltacode_version', __version__),
         ('deltacode_errors', collect_errors(deltacode)),
         ('deltas_count', len([d for d in deltas(deltacode, all_delta_types)])),
+        ('delta_stats', deltacode.stats.to_dict()),
         ('deltas', deltas(deltacode, all_delta_types))
     ])
 

src/deltacode/utils.py

@@ -23,7 +23,7 @@
 #  Visit https://github.com/nexB/deltacode/ for support and download.
 #
 
-from __future__ import absolute_import
+from __future__ import absolute_import, division
 
 from collections import defaultdict
 
@@ -182,14 +182,19 @@ def deltas(deltacode, all_delta_types=False):
         elif not delta.is_unmodified():
             yield delta.to_dict()
 
+def calculate_percent(value, total):
+    """
+    Return the rounded value percentage of total.
+    """
+    ratio = (value / total) * 100
+    return round(ratio, 2)
 
 class AlignmentException(Exception):
     """
     Named exception for alignment errors.
     """
     pass
 
-
 def align_trees(a_files, b_files):
     """
     Given two sequences of File objects 'a' and 'b', return a tuple of

tests/data/deltacode/all_stat_order_check_new.json

@@ -0,0 +1,69 @@
+{
+  "scancode_notice": "Generated with ScanCode and provided on an \"AS IS\" BASIS, WITHOUT WARRANTIES\nOR CONDITIONS OF ANY KIND, either express or implied. No content created from\nScanCode should be considered or used as legal advice. Consult an Attorney\nfor any legal advice.\nScanCode is a free software code scanning tool from nexB Inc. and others.\nVisit https://github.com/nexB/scancode-toolkit/ for support and download.",
+  "scancode_version": "2.9.0b1.post8.2a658c314",
+  "scancode_options": {
+    "input": "C:\\code\\nexb\\dev\\codebase01\\sorting\\sorted01_new",
+    "--copyright": true,
+    "--info": true,
+    "--json-pp": "C:\\code\\nexb\\dev\\deltacode\\tests\\data\\deltacode\\scan_sorted01_new.json",
+    "--license": true,
+    "--package": true
+  },
+  "files_count": 6,
+  "files": [
+    {
+      "path": "new/a.txt",
+      "type": "file",
+      "name": "a.txt",
+      "size": 100,
+      "sha1": "abtrbt",
+      "licenses": [
+      ]
+    },
+    {
+      "path": "new/b.txt",
+      "type": "file",
+      "name": "b.txt",
+      "size": 100,
+      "sha1": "dasdas",
+      "licenses": [
+      ]
+    },
+    {
+      "path": "new/c.txt",
+      "type": "file",
+      "name": "c.txt",
+      "size": 110,
+      "sha1": "avfvf",
+      "licenses": [
+      ]
+    },
+    {
+      "path": "new/d.txt",
+      "type": "file",
+      "name": "d.txt",
+      "size": 110,
+      "sha1": "avfvf",
+      "licenses": [
+      ]
+    },
+    {
+      "path": "new/e2.txt",
+      "type": "file",
+      "name": "e2.txt",
+      "size": 210,
+      "sha1": "avfvvrvf",
+      "licenses": [
+      ]
+    },
+    {
+      "path": "new/f2.txt",
+      "type": "file",
+      "name": "f2.txt",
+      "size": 110,
+      "sha1": "ivfvvrvf",
+      "licenses": [
+      ]
+    }
+  ]
+}

tests/data/deltacode/all_stat_order_check_old.json

@@ -0,0 +1,78 @@
+{
+  "scancode_notice": "Generated with ScanCode and provided on an \"AS IS\" BASIS, WITHOUT WARRANTIES\nOR CONDITIONS OF ANY KIND, either express or implied. No content created from\nScanCode should be considered or used as legal advice. Consult an Attorney\nfor any legal advice.\nScanCode is a free software code scanning tool from nexB Inc. and others.\nVisit https://github.com/nexB/scancode-toolkit/ for support and download.",
+  "scancode_version": "2.9.0b1.post8.2a658c314",
+  "scancode_options": {
+    "input": "C:\\code\\nexb\\dev\\codebase01\\sorting\\sorted01_old",
+    "--copyright": true,
+    "--info": true,
+    "--json-pp": "C:\\code\\nexb\\dev\\deltacode\\tests\\data\\deltacode\\scan_sorted01_old.json",
+    "--license": true,
+    "--package": true
+  },
+  "files_count": 7,
+  "files": [
+    {
+      "path": "old/a.txt",
+      "type": "file",
+      "name": "a.txt",
+      "size": 100,
+      "sha1": "abtrbt",
+      "licenses": [
+      ]
+    },
+    {
+      "path": "old/b.txt",
+      "type": "file",
+      "name": "b.txt",
+      "size": 100,
+      "sha1": "dasdas",
+      "licenses": [
+      ]
+    },
+    {
+      "path": "old/c.txt",
+      "type": "file",
+      "name": "c.txt",
+      "size": 100,
+      "sha1": "awd",
+      "licenses": [
+      ]
+    },
+    {
+      "path": "old/a/d.txt",
+      "type": "file",
+      "name": "d.txt",
+      "size": 110,
+      "sha1": "avfvf",
+      "licenses": [
+      ]
+    },
+    {
+      "path": "old/e1.txt",
+      "type": "file",
+      "name": "e1.txt",
+      "size": 110,
+      "sha1": "cegr",
+      "licenses": [
+      ]
+    },
+    {
+      "path": "old/f1.txt",
+      "type": "file",
+      "name": "f1.txt",
+      "size": 310,
+      "sha1": "gkbme",
+      "licenses": [
+      ]
+    },
+    {
+      "path": "old/g.txt",
+      "type": "file",
+      "name": "g.txt",
+      "size": 410,
+      "sha1": "btbtg",
+      "licenses": [
+      ]
+    }
+  ]
+}

tests/data/deltacode/scan_sorted01_new.json

@@ -9,7 +9,7 @@
     "--license": true,
     "--package": true
   },
-  "files_count": 5,
+  "files_count": 4,
   "files": [
     {
       "path": "sorted01_new",

tests/test_deltacode.py

@@ -1592,3 +1592,27 @@ def test_DeltaCode_permissive_add_public_domain(self):
         assert [d.status for d in deltas_object if d.new_file.path == 'a1.py'] == ['modified']
         assert [d.factors for d in deltas_object if d.new_file.path == 'a2.py'].pop() == []
         assert [d.status for d in deltas_object if d.new_file.path == 'a2.py'] == ['unmodified']
+
+    def test_Stat_calculation_and_ordering(self):
+        new_scan = self.get_test_loc('deltacode/all_stat_order_check_new.json')
+        old_scan = self.get_test_loc('deltacode/all_stat_order_check_old.json')
+
+        options = OrderedDict([
+            ('--all-delta-types', True)
+        ])
+
+        expected = OrderedDict([
+            ('old_files_count', 7),
+            ('new_files_count', 6),
+            ('percent_added', 28.57),
+            ('percent_removed', 42.86),
+            ('percent_moved', 14.29),
+            ('percent_modified', 14.29),
+            ('percent_unmodified', 28.57)
+        ])
+
+        deltacode_object = DeltaCode(new_scan, old_scan, options)
+
+        stats_object = deltacode_object.stats.to_dict()
+
+        assert stats_object == expected