From 03e33a15a6e6d3dafedd77b507955bd217a3ffaa Mon Sep 17 00:00:00 2001
From: Viktor Kopp <vifactor@gmail.com>
Date: Sat, 16 May 2020 16:19:44 +0200
Subject: [PATCH] Remove extension symbols count limit (#171)

- For files with no extension basename is returned as extension
- In file types table extensions sorted by files count in group
---
 .travis.yml               |  2 +-
 README.md                 |  3 ---
 analysis/gitrevision.py   |  6 ++----
 tools/__init__.py         | 18 ++++++++++++------
 tools/configuration.py    |  1 -
 tools/tests/__init__.py   |  0
 tools/tests/test_tools.py | 13 +++++++++++++
 7 files changed, 28 insertions(+), 15 deletions(-)
 create mode 100644 tools/tests/__init__.py
 create mode 100644 tools/tests/test_tools.py

diff --git a/.travis.yml b/.travis.yml
index ded23c5..1680a1b 100644
--- a/.travis.yml
+++ b/.travis.yml
@@ -8,4 +8,4 @@ install:
 virtualenv:
   system_site_packages: true
 script:
-  - python3 -m unittest
\ No newline at end of file
+  - python3 -m unittest -v
\ No newline at end of file
diff --git a/README.md b/README.md
index 004d9ec..3683755 100644
--- a/README.md
+++ b/README.md
@@ -94,7 +94,6 @@ Configuration file might contain following fields (all are optional):
 ```json
 {
     "max_domains": 10,
-    "max_ext_length": 10,
     "max_authors": 7,
     "max_plot_authors_count": 10,
     "max_authors_of_months": 6,
@@ -113,8 +112,6 @@ of contributors and activity levels, to avoid showing too much or too little
 information.
 
 * `max_domains`: number of e-mail domains to show in author stats
-* `max_ext_length`: max symbols count after `.` in a filename to 
-consider substring as a file extension
 * `max_authors`: number of authors in the "top authors" table 
 (other authors are listed without detailed stats)
 * `max_plot_authors_count`: number of authors to include in plots
diff --git a/analysis/gitrevision.py b/analysis/gitrevision.py
index 7f6a8b2..e4ebf74 100644
--- a/analysis/gitrevision.py
+++ b/analysis/gitrevision.py
@@ -1,5 +1,3 @@
-import functools
-
 from tools import get_file_extension
 from .gitdata import RevisionData, FilesData
 
@@ -26,8 +24,8 @@ def size(self):
     @property
     def files_extensions_summary(self):
         df = self.files_df[["size_bytes", "lines_count"]]
-        df["extension"] = self.files_df['file'].apply(functools.partial(get_file_extension, max_ext_length=6))
+        df["extension"] = self.files_df['file'].apply(get_file_extension)
         df = df.groupby(by="extension").agg({"size_bytes": ["sum"], "lines_count": ["sum", "count"]})
         df.columns = ["size_bytes", "lines_count", "files_count"]
         df.reset_index()
-        return df
+        return df.sort_values(by="files_count", ascending=False)
diff --git a/tools/__init__.py b/tools/__init__.py
index 99302eb..f0a8329 100644
--- a/tools/__init__.py
+++ b/tools/__init__.py
@@ -12,10 +12,16 @@ def split_email_address(email_address):
     return parts[0], parts[1]
 
 
-def get_file_extension(git_file_path, max_ext_length=5):
-    filename = os.path.basename(git_file_path)
+def get_file_extension(filepath: str):
+    assert filepath
+    filename = os.path.basename(filepath)
     basename_parts = filename.split('.')
-    ext = basename_parts[1] if len(basename_parts) == 2 and basename_parts[0] else ''
-    if len(ext) > max_ext_length:
-        ext = ''
-    return ext
+    if len(basename_parts) == 1:
+        # 'folder/filename'-case
+        return filename
+    elif len(basename_parts) == 2 and not basename_parts[0]:
+        # 'folder/.filename'-case
+        return filename
+    else:
+        # "normal" case
+        return basename_parts[-1]
diff --git a/tools/configuration.py b/tools/configuration.py
index ecb6d2f..36c5062 100644
--- a/tools/configuration.py
+++ b/tools/configuration.py
@@ -92,7 +92,6 @@ def __init__(self, args_orig, **kwargs):
     def _set_default_configuration(self):
         self.update({
             "max_domains": 10,
-            "max_ext_length": 10,
             "max_authors": 20,
             "max_plot_authors_count": 10,
             "max_authors_of_months": 6,
diff --git a/tools/tests/__init__.py b/tools/tests/__init__.py
new file mode 100644
index 0000000..e69de29
diff --git a/tools/tests/test_tools.py b/tools/tests/test_tools.py
new file mode 100644
index 0000000..3756f92
--- /dev/null
+++ b/tools/tests/test_tools.py
@@ -0,0 +1,13 @@
+import unittest
+
+import tools
+
+
+class TestTools(unittest.TestCase):
+
+    def test_get_file_extension(self):
+        self.assertEqual('extension', tools.get_file_extension("folder/filename.extension"))
+        self.assertEqual('.extension', tools.get_file_extension("folder/.extension"))
+        self.assertEqual('extension', tools.get_file_extension("folder/filename.suffix.extension"))
+        self.assertEqual('FILENAME', tools.get_file_extension("folder/FILENAME"))
+        self.assertEqual('extension', tools.get_file_extension("folder/.filename.suffix.extension"))