You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
docs: incorporate some AI guidance for contributors
This what @thisisnic suggested via apache/arrow#48952 and has discussed
on the Apache Arrow developers mailing list. I think this is good
guidance to start with for AI contributions
Signed-off-by: R. Tyler Croy <[email protected]>
Copy file name to clipboardExpand all lines: CONTRIBUTING.md
+38Lines changed: 38 additions & 0 deletions
Display the source diff
Display the rich diff
Original file line number
Diff line number
Diff line change
@@ -8,6 +8,44 @@ If you want to start contributing, first look at our good first issues: https://
8
8
9
9
If you want to contribute something more substantial, see our "Projects seeking contributors" section on our roadmap: https://github.com/delta-io/delta-rs/issues/1128
10
10
11
+
## AI-generated code
12
+
13
+
We recognise that AI coding assistants are now a regular part of many
14
+
developers' workflows and can improve productivity. Thoughtful use of these
15
+
tools can be beneficial, but AI-generated PRs can sometimes lead to
16
+
undesirable additional maintainer burden. PRs that appear to be fully
17
+
generated by AI with little to no engagement from the author may be closed
18
+
without further review.
19
+
20
+
Human-generated mistakes tend to be easier to spot and reason about, and
21
+
code review is intended to be a collaborative learning experience that
22
+
benefits both submitter and reviewer. When a PR appears to have been
23
+
generated without much engagement from the submitter, reviewers with access
24
+
to AI tools could more efficiently generate the code directly, and since
25
+
the submitter is not likely to learn from the review process, their time is
26
+
more productively spent researching and reporting on the issue.
27
+
28
+
We are not opposed to the use of AI tools in generating PRs, but recommend
29
+
the following:
30
+
31
+
* Only submit a PR if you are able to debug and own the changes yourself -
32
+
review all generated code to understand every detail. [Apache Datafusion has a useful explanation of **why fully AI-generated PRs without understanding are not helpful**](https://datafusion.apache.org/contributor-guide/index.html#why-fully-ai-generated-prs-without-understanding-are-not-helpful).
33
+
* Match the style and conventions used in the rest of the codebase, including
34
+
PR titles and descriptions
35
+
* Be upfront about AI usage and summarise what was AI-generated
36
+
* If there are parts you don't fully understand, leave comments on your own PR
37
+
explaining what steps you took to verify correctness
38
+
* Watch for AI's tendency to generate overly verbose comments, unnecessary
39
+
test cases, and incorrect fixes
40
+
* Break down large PRs into smaller ones to make review easier
41
+
42
+
PR authors are also responsible for disclosing any copyrighted materials in
43
+
submitted contributions. See the `[Apache Software
44
+
Foundation's](https://apache.org)[guidance on AI-generated
45
+
code](https://www.apache.org/legal/generative-tooling.html) for further
46
+
information on licensing considerations.
47
+
48
+
11
49
## Claiming an issue
12
50
13
51
If you want to claim an issue to work on, you can write the word `take` as a comment in it and you will be automatically assigned.
0 commit comments