Why gitpython results doesn't match to the github website?

Question

I use the following python code to extract the diff (the hunks) between two commits.

from git import Repo
!git clone https://github.com/apache/commons-math.git
repo = Repo("/content/commons-math")
file_path = 'commons-math-legacy/src/test/java/org/apache/commons/math4/legacy/distribution/EmpiricalDistributionTest.java'
parent = 'd080f0d8251d58728024955764a5c0c75acf8277'
commit = '9d1741bfe4a7808cfa0c313891a717adf98a3087'
hunks = repo.git.diff(parent, commit, file_path, ignore_blank_lines=True, ignore_space_at_eol=True)

The hunks show that the specified file is a new file that is created by adding 689 lines:

diff --git a/commons-math-legacy/src/test/java/org/apache/commons/math4/legacy/distribution/EmpiricalDistributionTest.java b/commons-math-legacy/src/test/java/org/apache/commons/math4/legacy/distribution/EmpiricalDistributionTest.java
new file mode 100644
index 000000000..dfdfdd946
--- /dev/null
+++ b/commons-math-legacy/src/test/java/org/apache/commons/math4/legacy/distribution/EmpiricalDistributionTest.java
@@ -0,0 +1,689 @@
+/*
+ * Licensed to the Apache Software Foundation (ASF) under one or more
.
.
.

But, when I open the corresponding GitHub commit page, and check the details for EmpiricalDistributionTest.java, it shows that this file is renamed (the containing folder is changed) and a few lines are updated. My first question is why the results from GitPython doesn't match to the GitHub interface? And the second question is how could I configure GitPython to get the same results as GitHub website interface?

I found that this problem happens when a file is moved to another folder and the content of the file changes in that commit. In java projects, when the containing package of a class changes, the folder names and the content of the file changes. But, I have no idea why GitPython cannot detect this situation as an update on an existing file. Thanks in advance for your help.

Answer 1

The reason why the results from GitPython do not match the GitHub interface is that GitPython is only looking at the diff between the two commits that I specified, while the GitHub interface is showing the full history of the file. When a file is moved to a different folder and its contents are updated, Git stores this as two separate changes: first, the file is deleted from its old location, and second, a new file is created at the new location with the updated content.

To configure GitPython to get the same results as the GitHub website interface, use the --follow option when calling the git.diff command. This option instructs Git to follow the history of the file even if it has been moved or renamed. Here's an updated version of the code that includes the --follow option:

hunks = repo.git.diff(parent, commit, file_path, ignore_blank_lines=True, ignore_space_at_eol=True, follow=True)

With the --follow option, the git.diff command will detect that the file has been moved and renamed, and it will show the full history of the file.