Announce: Log Detective - AI tool to analyze build logs failures

[Jiri Kyjovsky]

Hey, Fedora devs!

Ever find yourself lost in the labyrinth of build logs, desperately seeking
the elusive culprit behind a failed build? Enter Log Detective – your
future personal AI guru dedicated to unravelling the mysteries of build log
failures within the RPM community.

*What is Log Detective?*

Log Detective will be an AI tool for analyzing failed build logs in the RPM
community. We are currently in the early stage of development and plan to
train the AI using real-world data from maintainers, creating a specialized
model for error identification. Thus we created a data collection portal
[1], where you can contribute your failed build logs.

*Why Log Detective?*

Diving into build logs feels like navigating a jungle of thousands of
lines. When a build fails, spotting the glitch easily turns into a
head-scratcher even for pros. "ERROR" in logs doesn't guarantee always a
quick fix, and these errors don't reveal themselves at the log's end.
Untangling this requires a keen understanding of packaging difficulties.

*Help us shape the future of Log Detective!*

We're reaching out to developers like you to contribute by uploading your
recent failed build logs on our website [1] and explaining why the build
failed. Your input is crucial in building the dataset that will train Log
Detective's AI model. By participating, you're playing a key role in
creating a tool to streamline error resolution for the entire RPM
community. Join us in making Fedora even more powerful!

*How to contribute data to Log Detective data collection website?*

Visit log-detective.com [1] and share your recent failed build. We can
fetch the logs from build systems like Copr [2], Koji [3], services like
Packit [4] or arbitrary URL. Highlight the lines in the log associated with
the failure and describe how it can be fixed. The more detailed, the
merrier for our final tool, and the more hints for you and other developers
when your build fails in future.

*Join the Log Detective!*

Drop by our GitHub repository [5] to share your ideas. Let's collaborate to
build a tool that changes the game for handling build log failures.

Log Detective isn't just a tool, it's your AI sidekick for defeating build
log challenges. Be part of the revolution!

Cheers,
The Log Detective Crew

[1] - https://log-detective.com/
[2] - https://copr.fedorainfracloud.org/
[3] - https://koji.fedoraproject.org/
[4] - https://packit.dev/
[5] - https://github.com/fedora-copr/log-detective-website

[Tristan Cacqueray]

On Wed, Jan 17, 2024 at 12:34 Jiri Kyjovsky wrote:
>  Hey, Fedora devs!
>
>  Ever find yourself lost in the labyrinth of build logs, desperately seeking
>  the elusive culprit behind a failed build? Enter Log Detective – your
>  future personal AI guru dedicated to unravelling the mysteries of build log
>  failures within the RPM community.
>
>  *What is Log Detective?*
>
>  Log Detective will be an AI tool for analyzing failed build logs in the RPM
>  community. We are currently in the early stage of development and plan to
>  train the AI using real-world data from maintainers, creating a specialized
>  model for error identification. Thus we created a data collection portal
>  [1], where you can contribute your failed build logs.
>
>  *Why Log Detective?*
>
>  Diving into build logs feels like navigating a jungle of thousands of
>  lines. When a build fails, spotting the glitch easily turns into a
>  head-scratcher even for pros. "ERROR" in logs doesn't guarantee always a
>  quick fix, and these errors don't reveal themselves at the log's end.
>  Untangling this requires a keen understanding of packaging difficulties.
>
>  *Help us shape the future of Log Detective!*
>
>  We're reaching out to developers like you to contribute by uploading your
>  recent failed build logs on our website [1] and explaining why the build
>  failed. Your input is crucial in building the dataset that will train Log
>  Detective's AI model. By participating, you're playing a key role in
>  creating a tool to streamline error resolution for the entire RPM
>  community. Join us in making Fedora even more powerful!
>
>  *How to contribute data to Log Detective data collection website?*
>
>  Visit log-detective.com [1] and share your recent failed build. We can
>  fetch the logs from build systems like Copr [2], Koji [3], services like
>  Packit [4] or arbitrary URL. Highlight the lines in the log associated with
>  the failure and describe how it can be fixed. The more detailed, the
>  merrier for our final tool, and the more hints for you and other developers
>  when your build fails in future.
>
>  *Join the Log Detective!*
>
>  Drop by our GitHub repository [5] to share your ideas. Let's collaborate to
>  build a tool that changes the game for handling build log failures.
>
>  Log Detective isn't just a tool, it's your AI sidekick for defeating build
>  log challenges. Be part of the revolution!
>
>  Cheers,
>  The Log Detective Crew
>
>  [1] - https://log-detective.com/
>  [2] - https://copr.fedorainfracloud.org/
>  [3] - https://koji.fedoraproject.org/
>  [4] - https://packit.dev/
>  [5] - https://github.com/fedora-copr/log-detective-website 
Hello,

I'm building a similar project named logjuicer using the
following dataset for evaluation:
  https://github.com/logjuicer/logjuicer-tests

How do you store the collected data and is there a schema/library to
access it?

Cheers,
-Tristan

[Jiri Kyjovsky]

Hello Tristan,

We store the data in json format. You can access and download the data directly at our
main page at the bottom - "Download collected data". Or via
https://log-detective.com/download

Cheers
Jiri

[Jakub Kadlcik]

Hello all,
I recorded a very short video showing how you can contribute to the Log
Detective project.

https://www.youtube.com/watch?v=_-O7ryKCnlQ

Any help will be greatly appreciated :-)
Jakub



On Wed, Jan 17, 2024 at 8:26 PM Jiri Kyjovsky <jkyjovsk(a)redhat.com&gt; wrote:

>  Hello Tristan,
>
>  We store the data in json format. You can access and download the data
>  directly at our main page at the bottom - "Download collected data". Or via
>  https://log-detective.com/download
>
>  Cheers
>  Jiri
>  --
>  _______________________________________________
>  devel mailing list -- devel(a)lists.fedoraproject.org
>  To unsubscribe send an email to devel-leave(a)lists.fedoraproject.org
>  Fedora Code of Conduct:
>  https://docs.fedoraproject.org/en-US/project/code-of-conduct/
>  List Guidelines: https://fedoraproject.org/wiki/Mailing_list_guidelines
>  List Archives:
>  https://lists.fedoraproject.org/archives/list/[e-mail]
>  Do not reply to spam, report it:
>  https://pagure.io/fedora-infrastructure/new_issue
>