skewness

Author: kddubey

github_fetches/sklearn/count.py

@@ -1,6 +1,11 @@
+from collections import defaultdict
 import json
 from pathlib import Path
 
+import seaborn as sns
+
+sns.set_theme(style="darkgrid")
+
 
 def open_json(path: str) -> dict:
     with open(path, "r") as f:
@@ -19,8 +24,30 @@ def open_json(path: str) -> dict:
         ]
     issue_num_to_data.update(issues_batch)
 
-num_issues_with_links = sum(
-    bool(issue["body_links"] + issue["comment_links"])
+issue_num_to_links: dict[str, list[str]] = {
+    issue_num: issue["body_links"] + issue["comment_links"]
     for issue_num, issue in issue_num_to_data.items()
+    if issue["body_links"] + issue["comment_links"]
+}
+issue_num_to_pages = {}
+for issue_num, links in issue_num_to_links.items():
+    processed = {link.split("/")[-1].split("#")[0] for link in links}
+    pages = {link for link in processed if link.endswith(".html")}
+    if pages:
+        issue_num_to_pages[issue_num] = pages
+
+page_to_issue_nums = defaultdict(set)
+for issue_num, pages in issue_num_to_pages.items():
+    for page in pages:
+        page_to_issue_nums[page].add(issue_num)
+page_to_issue_nums = sorted(
+    page_to_issue_nums.items(), key=lambda x: len(x[1]), reverse=True
 )
-print(num_issues_with_links / len(issue_num_to_data))
+num_issues_per_page = [len(issue_nums) for _, issue_nums in page_to_issue_nums]
+
+num_issues_per_page = num_issues_per_page[2:]
+# The first two pages are really basic and shouldn't be counted I think:
+# contributing.html
+# develop.html
+
+sns.lineplot(x=range(len(num_issues_per_page)), y=num_issues_per_page)