Modify apl storage

[comcon1]

• changes schema of apl.json
  From:

{
 "0.0": 62.4,
 "20.0": 63.1
}

to:

[
   [0, 62.4],
   [20.0, 63.1]
]

• remove code duplication in data-retrieving API
• add get_density functionality

NOTE! Can be merged only together with corresponding schema changes in BilayerData.

---

📚 Documentation preview 📚: https://databank--467.org.readthedocs.build/

[comcon1]

Here is a script for conversion of Simulations folder in the BilayerData repository:

#!/usr/bin/env python3
"""
Convert apl.json dicts to sorted Nx2 matrix arrays.
Searches subfolders, saves as apl_matrix.json.
"""

import json
from pathlib import Path

from fairmd.lipids.auxiliary import CompactJSONEncoder

def convert_apl_file(input_path):
    """Convert single apl.json to matrix."""
    with open(input_path, "r") as f:
        data = json.load(f)

    # Sort by price key (float), convert to [price, value] pairs
    matrix = [[float(k), v] for k, v in sorted(data.items(), key=lambda x: float(x[0]))]

    with open(input_path, "w") as f:
        json.dump(matrix, f, indent=2, cls=CompactJSONEncoder)

    print(f"✓ Converted {input_path} → {input_path} ({len(matrix)} rows)")


def main():
    root = Path(".")
    apl_files = list(root.rglob("apl.json"))

    if not apl_files:
        print("No apl.json files found in subfolders.")
        return

    print(f"Found {len(apl_files)} apl.json files:")
    for f in apl_files:
        print(f"  {f}")

    for apl_file in apl_files:
        try:
            convert_apl_file(apl_file)
        except Exception as e:
            print(f"✗ Error in {apl_file}: {e}")


if __name__ == "__main__":
    main()

Should be applied after the script is accepted.

[batukav]

Can we not apply a schema like

{"time": [0.0, 20.0], "apl": [62.4, 63.1]}

This would make accessing the time series values much easier. If @comcon1 schema is required for the other parts of the project like the website etc, I am fine with the original request. I just wanted to know.

[comcon1]

We can apply this to everything, yes. But will it be comfortable to compare in git diffs? Best regards, Alex Sent from [Proton Mail](https://proton.me/mail/home) for Android.

…

-------- Original Message --------

On Sunday, 02/22/26 at 15:13 Batuhan Kav ***@***.***> wrote: @batukav commented on this pull request. Can we not apply a schema like {"time": [0.0, 20.0], "apl": [62.4, 63.1]} This would make accessing the time series values much easier. If ***@***.***(https://github.com/comcon1) schema is required for the other parts of the project like the website etc, I am fine with the original request. I just wanted to know. — Reply to this email directly, [view it on GitHub](#467 (review)), or [unsubscribe](https://github.com/notifications/unsubscribe-auth/AGBXJOH6FKUH47KLJ4YRZNT4NG2QRAVCNFSM6AAAAACV2SG7E2VHI2DSMVQWIX3LMV43YUDVNRWFEZLROVSXG5CSMV3GSZLXHMZTQMZXG43TIOBUGY). You are receiving this because you were mentioned.Message ID: ***@***.***>

[comcon1]

Can we not apply a schema like

If to apply this schema, then probably to all data of this kind? To densitites and FormFactors? Will it be comfortable to see the values like this. I'm not sure. I want to see by eyes that density at 0 is 200 etc.

[batukav]

The schema would be the same for all time/spatial series data, including density profile, form factors etc.

After some thoughts, I agree that we should go with the initial schema suggestion with the condition that each data point will be a single line as shown below. It's good for the eyes, easy on git, easy to audit and easy to load into numpy.

[
  [0.0, 62.4],
  [20.0, 63.1],
  [40.0, 63.5]
]

[batukav]

Please consider making each entry a single line, for example

[
  [
    100000.0078125, 61.4804
  ],
  [
...

It will make using CLI commands like grep also easier.

[comcon1]

You think it

[
  [
    100000.0078125, 61.4804
  ],
  [
   ...

is better than

[
  [100000.0078125, 61.4804],
  [...
```?
It is three times more lines! It still has one list entity on the line.

[comcon1]

A.. I see!
I have one number per line in my edits. It's quite stupid. Should fix it definitely, thx!

[comcon1]

I have used now the compact encoder to get a nice format. The formatter-snippet is updated so it can be applied directly to the database.