From 799fb5e9584de13850cb704a04cd3397fde35e1e Mon Sep 17 00:00:00 2001
From: Jay DeLuca <jaydeluca4@gmail.com>
Date: Tue, 24 Dec 2024 14:25:54 -0500
Subject: [PATCH] list files in count by script

---
 .gitignore                  |  6 +++++-
 count_by_instrumentation.py | 19 +++++++++++++------
 data_filter.py              |  1 +
 main.py                     |  4 ++++
 utilities.py                | 19 ++++++++++++++++---
 5 files changed, 39 insertions(+), 10 deletions(-)

diff --git a/.gitignore b/.gitignore
index 5d289db..48f8b11 100644
--- a/.gitignore
+++ b/.gitignore
@@ -9,4 +9,8 @@ __pycache__/*
 tests/cache
 tests/cache/*
 
-.coverage
\ No newline at end of file
+.coverage
+
+latest-groovy.png
+latest-with-java.png
+latest-groovy-detailed.png
\ No newline at end of file
diff --git a/count_by_instrumentation.py b/count_by_instrumentation.py
index 991f5fa..58cb3dc 100644
--- a/count_by_instrumentation.py
+++ b/count_by_instrumentation.py
@@ -13,8 +13,9 @@
 
 
 class App:
-    def __init__(self, languages: List[str], path_prefix: str, keyword: str):
-        self.client = GithubClient()
+    def __init__(self, languages: List[str], path_prefix: str, keyword: str,
+                 client: GithubClient = GithubClient()):
+        self.client = client
         self.data_filter = DataFilter(languages=languages,
                                       path_prefix=path_prefix, keyword=keyword)
 
@@ -35,19 +36,21 @@ def main(args):
         keyword="test"
     )
 
-    today = datetime.now().date().strftime("%Y-%m-%dT%H:%M:%SZ")
+    today = (datetime.now().date() + pd.Timedelta(days=1)).strftime(
+        "%Y-%m-%dT%H:%M:%SZ")
 
     commit = app.get_commit_by_date(date=today, repository=args.repo)
     repo_files = app.get_repository_by_commit(
         repository=args.repo,
         commit=commit
     )
-    file_counts, file_sizes = count_by_language_and_file_extension(
+    file_info = count_by_language_and_file_extension(
         files=repo_files["files"],
         languages=[args.language])
 
     # Print the table showing file counts and sizes
-    data = [(key, file_counts[key], file_sizes[key]) for key in file_counts.keys()]
+    data = [(key, file_info.file_counts[key], file_info.file_sizes[key]) for key in
+            file_info.file_counts.keys()]
     df2 = pd.DataFrame(data, columns=['Key', 'File Count', 'Total File Size'])
     df2 = df2.sort_values(by='Total File Size', key=lambda col: col.astype(int),
                           ascending=False)
@@ -56,7 +59,7 @@ def main(args):
     print(f"| Total | {df2['File Count'].sum()} | {df2['Total File Size'].sum()} |")
 
     # Create a pie chart for file counts only
-    df = pd.DataFrame(list(file_counts.items()), columns=['Key', 'Value'])
+    df = pd.DataFrame(list(file_info.file_counts.items()), columns=['Key', 'Value'])
     df = df.sort_values(by='Value', key=lambda col: col.astype(int), ascending=False)
 
     sns.set_theme()
@@ -69,6 +72,10 @@ def main(args):
     plt.title(f'Remaining {args.language} files by Instrumentation')
     plt.ylabel('')
 
+    print("\n")
+    for item in file_info.matched_files:
+        print(item)
+
     if args.output is not None:
         plt.savefig(args.output)
     else:
diff --git a/data_filter.py b/data_filter.py
index 20637ba..77079d4 100644
--- a/data_filter.py
+++ b/data_filter.py
@@ -42,6 +42,7 @@ def get_file_counts_and_lengths(self, payload):
         for i in tree:
             if self.matches_meta(i) \
                     and self.matches_file_extensions(i["path"]) \
+                    and "grails" not in i["path"] \
                     and self.matches_directory(i["path"]):
                 data_result.append(CodeFile(path=i["path"], size=i["size"]))
 
diff --git a/main.py b/main.py
index 65204e2..bfeea10 100644
--- a/main.py
+++ b/main.py
@@ -89,6 +89,10 @@ def main(args):
 
     for lang, counts in language_counts.items():
         df = pd.DataFrame({'Date': dates, 'Count': counts})
+
+        # remove any junk data
+        df = df[df['Count'] != 0]
+
         df['Date'] = pd.to_datetime(df['Date'])
         sns.lineplot(x='Date', y='Count', label=lang.capitalize(), data=df, marker='o')
 
diff --git a/utilities.py b/utilities.py
index c98e37e..ee4f0ce 100644
--- a/utilities.py
+++ b/utilities.py
@@ -1,3 +1,4 @@
+from dataclasses import dataclass
 from datetime import datetime, timedelta
 from collections import defaultdict
 from typing import List, Dict
@@ -5,6 +6,13 @@
 from CodeFile import CodeFile
 
 
+@dataclass
+class FileCountInfo:
+    file_counts: Dict[str, int]
+    file_sizes: Dict[str, int]
+    matched_files: List[str]
+
+
 def get_dates_between(start_date_str, end_date, interval):
     date_format = "%Y-%m-%d"
     output_format = "%Y-%m-%dT%H:%M:%SZ"
@@ -38,12 +46,14 @@ def count_by_file_extension(files: List[str], languages: List[str]) -> dict:
     for file in files:
         for ext in languages:
             extension = f".{ext.lower()}"
-            if file.endswith(extension):
+            if file.endswith(extension) and "grails" not in file:
                 file_counts[ext] += 1
     return file_counts
 
 
-def count_by_language_and_file_extension(files: List[CodeFile], languages: List[str]) -> Dict[str, Dict[str, int]]:
+def count_by_language_and_file_extension(files: List[CodeFile],
+                                         languages: List[str]) -> FileCountInfo:
+    matched_files = []
     file_counts = defaultdict(int)
     file_sizes = defaultdict(int)
     for file in files:
@@ -55,7 +65,10 @@ def count_by_language_and_file_extension(files: List[CodeFile], languages: List[
         if extension in languages:
             file_counts[instrumentation] += 1
             file_sizes[instrumentation] += file.size
-    return file_counts, file_sizes
+            matched_files.append(file.path)
+
+    return FileCountInfo(file_counts=file_counts, file_sizes=file_sizes,
+                         matched_files=matched_files)
 
 
 def convert_to_plot(input_dict: dict, items):