# \`/code\_execution/data\` directory

**URL:** <https://community.drivendata.org/t/code-execution-data-directory/10668>\
**Category:** NASA Pose Bowl\
**Created:** [April 26, 2024, 1:09pm UTC](https://community.drivendata.org/t/code-execution-data-directory/10668 "2024-04-26T13:09:03Z")\
**Posts on this page:** 4\
**Page:** 1

<div class="post-metadata">

**Author:** ![miturbe](https://avatars.discourse-cdn.com/v4/letter/m/cab0a1/32.png) [@miturbe](https://community.drivendata.org/u/miturbe)\
**Post date:** [April 26, 2024, 1:09pm UTC](https://community.drivendata.org/t/code-execution-data-directory/10668/1 "2024-04-26T13:09:03Z")

</div>

I have a question regarding the following paragraph: " **Remember** that for local testing purposes, the `/code_execution/data` directory is just a mounted version of what you have saved locally in this project’s `data` directory. So you will just be using the publicly available training files for local testing. In the official code execution environment, `/code_execution/data` will contain the _actual test data_ , which no participants have access to, and this is what will be used to compute your score for the leaderboard."

Q: is the `/code_execution/data` directory, for the private test data, organised in a set of foleders where each folder contains a chain of images?

---

<div class="post-metadata">

**Author:** ![isms](https://yyz2.discourse-cdn.com/flex028/user_avatar/community.drivendata.org/isms/32/9_2.png) [@isms](https://community.drivendata.org/u/isms)\
**Post date:** [April 26, 2024, 2:30pm UTC](https://community.drivendata.org/t/code-execution-data-directory/10668/2 "2024-04-26T14:30:57Z")

</div>

Yes, it’s the same directory structure as the training data, just with different data.

---

<div class="post-metadata">

**Author:** ![miturbe](https://avatars.discourse-cdn.com/v4/letter/m/cab0a1/32.png) [@miturbe](https://community.drivendata.org/u/miturbe)\
**Post date:** [April 27, 2024, 9:50am UTC](https://community.drivendata.org/t/code-execution-data-directory/10668/3 "2024-04-27T09:50:45Z")

</div>

Taking advantage of the current topic I would like to clarify another doubt.

1st) As descrived in the runtime repository “/code\_execution” is the working directory from where the “main.sh” file will be executed?

2nd) Assuming the 1st statment is as described above, I would like to know if the following example should result in obtaining the full path?

source\_path = Path(r"data/images/b038c264fe").resolve()

Result: /mnt/c/code\_execution/data/images/b038c264fe where _/mnt/c_ is an assumption of the directory where all the files of my working directory are located.

3rd) And in the case of this code line: model\_path = Path(r"submission\_src/best.pt").resolve() will the path part “submission\_src/” pose any challenge when executing it? Should it be changed to just “submission/” without \_src? even if the folder where it is contained is called “submission\_src/” to allow a make pack-submission command to zip the files?

I get confused with the directories of submission:src

---

<div class="post-metadata">

**Author:** ![isms](https://yyz2.discourse-cdn.com/flex028/user_avatar/community.drivendata.org/isms/32/9_2.png) [@isms](https://community.drivendata.org/u/isms)\
**Post date:** [April 30, 2024, 11:58pm UTC](https://community.drivendata.org/t/code-execution-data-directory/10668/4 "2024-04-30T23:58:04Z")

</div>

Hi @miturbe - I’m happy to answer this in general, but the larger issue is that you should definitely set yourself up to test locally with docker. If you do that, when a question like this comes up you could just put a `print` statement in your script to get a quick answer.

Let me know if you have trouble getting that set up.