Feat spectra sim

[jiaqingxie]

Summary

Closes: #(issue)

References

• Contribution guide: CONTRIBUTING.md
• Skills catalog and categories: skills/README.md

If this PR adds a new skill

• The new skill folder is under skills/<skill-name>/
• SKILL.md frontmatter includes name and description
• name matches folder name and uses kebab-case
• I checked skills/README.md and assigned this skill to the correct category
• I considered whether this should be merged into an existing skill instead of creating a new one
• I updated skills/README.md (name, description, status, and link if applicable)

Chosen category in skills/README.md:

Merge/Separate decision note:

Validation

# Example:
# npx skills add . --list
# npx skills add . --skill <skill-name>

Impact

• Affected skills:
• Backward compatibility:
• Risks / limitations:

Checklist

• Change type: New skill / Update skill / Bug fix / Docs / Other
• I read CONTRIBUTING.md
• I searched for duplicate PRs/issues
• I tested changes locally
• I updated related docs (README / skills catalog / examples) if needed

[gemini-code-assist]

Code Review

This pull request introduces two new skills, f-nmr-simulation and p-nmr-simulation, for predicting solid-state 19F and 31P NMR chemical shifts using the NMRNet model. Additionally, the ms-spectra-simulation skill has been refactored to utilize the fioRa online application via Playwright browser automation, replacing the previous local installation. Review feedback suggests optimizing the download of large model weights to avoid potential memory issues and using the csv module for safer payload construction in the mass spectrometry simulation script.

I am having trouble creating individual review comments. Click here to see my feedback.

skills/f-nmr-simulation/f_nmr_simulation.py (57)

Using zf.read(zip_path) loads the entire file into memory. Since the model weights are approximately 560 MB, this can lead to Out-Of-Memory (OOM) errors in environments with limited RAM. It is more efficient to stream the file content directly to disk using shutil.copyfileobj.

            import shutil
            with zf.open(zip_path) as source, open(local_path, "wb") as target:
                shutil.copyfileobj(source, target)

skills/p-nmr-simulation/p_nmr_simulation.py (57)

Using zf.read(zip_path) loads the entire file into memory. Since the model weights are approximately 560 MB, this can lead to Out-Of-Memory (OOM) errors in environments with limited RAM. It is more efficient to stream the file content directly to disk using shutil.copyfileobj.

            import shutil
            with zf.open(zip_path) as source, open(local_path, "wb") as target:
                shutil.copyfileobj(source, target)

skills/ms-spectra-simulation/ms_spectra_simulation.py (70-74)

The CSV payload for the fioRa app is constructed using f-strings without escaping. If args.name or args.smiles contains commas or newlines, the resulting CSV will be malformed. Using the csv module ensures that fields are correctly quoted and escaped.

def build_textbox_payload(args: argparse.Namespace) -> str:
    import io
    import csv
    output = io.StringIO()
    writer = csv.writer(output, lineterminator="\n")
    writer.writerow(["Name", "SMILES", "Precursor_type", "CE", "Instrument_type"])
    writer.writerow([args.name, args.smiles, args.precursor, f"{args.ce:g}", args.instrument])
    return output.getvalue()

[little1d]

@jiaqingxie 可能需要对应修改一下 skills/README.md，把这两个新的 skill 描述加上，其他没什么问题