Add MWTabFile.download_results_file() for programmatic results file download

[sagarbasnet74]

Summary

The CLI supports --results-files but MWTabFile has no programmatic way to download results files. This PR adds MWTabFile.download_results_file() to fill that gap.

Changes

• Added download_results_file() to MWTabFile class in src/mwtab/mwtab.py
• Uses existing study_id and analysis_id managed properties
• Supports both MS_RESULTS_FILE and NMR_RESULTS_FILE sections
• Allows optional requests.Session for connection pooling
• Falls back to default filename if results dict lacks filename key

Tests

• 8 new test cases in tests/test_download_results_file.py
• All 330 tests pass (322 original + 8 new)

Usage

from mwtab import mwtab as mt
from mwtab import fileio, mwrest

url = mwrest.GenericMWURL({
    "context": "study",
    "input_item": "study_id", 
    "input_value": "ST000001",
    "output_item": "mwtab"
}).url

mwfile = next(fileio.read_with_class(url, mt.MWTabFile, {}))
path = mwfile.download_results_file(output_dir=".")

[hunter-moseley]

Thank you for this pull request.
Due to your prodding, we are looking to add the results file download functionality into the API.
However, we will probably implement it through the mwrest submodule.

[sagarbasnet74]

Thank you for the feedback, Dr. Moseley! That makes sense — keeping REST operations in mwrest.py maintains cleaner separation of concerns.

[sagarbasnet74]

Hi Dr. Moseley, Thank you for the quick feedback on PR #15. I completely agree that implementing results file download through mwrest.py is architecturally cleaner — my MWTabFile approach was pragmatic but your modular design is better long-term. I'd love to learn more about your planned mwrest implementation and discuss where I could contribute most effectively — whether that's building out the mwrest version, writing tests for it, or other projects in the lab's pipeline like kegg_pull or rcha_metab. To be direct: I am applying to the MS in Bioinformatics program at the University of Kentucky and would very much like to join your lab. I have hands-on computational research experience, and your work on metabolomics pipelines, multi-omics integration, and open-source scientific tooling sits exactly at the intersection of what I want to spend the next several years on. Beyond the results file downloader, I also explored extending mwtab with a Polars backend for the pandas bridge (for faster large-scale metabolite data processing) and a KEGG cross-annotation module via kegg_pull — both motivated by gaps I noticed while working through the codebase. I'd welcome your thoughts on whether these directions align with the lab's roadmap. I am attaching my CV and would be grateful for 15 minutes of your time to discuss both the technical work and potential fit in your lab. Thank you for building mwtab and making it open — it is genuinely one of the most well-engineered research codebases I have worked with. Best regards, Sagar Basnet GitHub: https://github.com/sagarbasnet74

…

On Tue, 9 Jun 2026 at 06:39, Hunter Moseley ***@***.***> wrote: *hunter-moseley* left a comment (MoseleyBioinformaticsLab/mwtab#15) <#15 (comment)> Thank you for this pull request. Due to your prodding, we are looking to add the results file download functionality into the API. However, we will probably implement it through the mwrest submodule. — Reply to this email directly, view it on GitHub <#15?email_source=notifications&email_token=A4ERHUBICZN5MB4C5B55ETT465N3PA5CNFSNUABFM5UWIORPF5TWS5BNNB2WEL2JONZXKZKDN5WW2ZLOOQXTINRVGQ4TKMZZGM3KM4TFMFZW63VGMF2XI2DPOKSWK5TFNZ2KYZTPN52GK4S7MNWGSY3L#issuecomment-4654953936>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/A4ERHUHKVMNOAXHLLMKIC6T465N3PAVCNFSM6AAAAACZ5AK2YSVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHM2DMNJUHE2TGOJTGY> . Triage notifications, keep track of coding agent tasks and review pull requests on the go with GitHub Mobile for iOS <https://github.com/notifications/mobile/ios/A4ERHUCGWEXSYSQFWDHH5AD465N3PA5CNFSNUABFM5UWIORPF5TWS5BNNB2WEL2JONZXKZKDN5WW2ZLOOQXTINRVGQ4TKMZZGM3KM4TFMFZW63VGMF2XI2DPOKSWK5TFNZ2KUZTPN52GK4S7NFXXG> and Android <https://github.com/notifications/mobile/android/A4ERHUGTX6EYBQ6U3A5AZ3T465N3PA5CNFSNUABFM5UWIORPF5TWS5BNNB2WEL2JONZXKZKDN5WW2ZLOOQXTINRVGQ4TKMZZGM3KM4TFMFZW63VGMF2XI2DPOKSWK5TFNZ2K4ZTPN52GK4S7MFXGI4TPNFSA>. Download it today! You are receiving this because you authored the thread.Message ID: ***@***.***>

[ptth222]

I ended up moving the results file downloading into the MWRESTFile class, so it is now available there and through the read_mwrest function in the fileio module. I only had it in the CLI because I couldn't really imagine a use case where it would be needed in the API. I would expect that anyone wanting to work with these files would just use the CLI to grab them and then manipulate them programmatically. Any code written to download anything programmatically is just duplicating what the CLI already does, so it just seems unlikely to me that someone would reinvent the wheel for that.

Even though we won't be using it, I do have some feedback on the code you wrote if you are interested. None of this is meant to criticize or be negative.

1. You can't assume Study and/or Analysis IDs exist. You need error checking on them. The "download_and_save_mwrest_file" function in cli.py has it as an example.
2. You can ignore finding the filename. The filename should always be of the form "STXXXXXX_ANXXXXXX_Results.txt". Filenames that don't look like that are a mistake and that name doesn't appear when you get the file from the Workbench site.
3. You can ignore finding the RESULT_FILE key. You don't need the RESULT_FILE key because it is notorious for being malformed and might not get read in correctly. It's probably better to check the response status when trying to download or looking for "RESULTS_FILE" in the raw lines like "download_and_save_mwrest_file" in cli.py does. I can pretty much guarantee that as it is right now it would not get the same results as using the --results-files option.
4. "base_url" should instead use the "RESULTS_FILE_BASE_URL" from cli.py. This would likely need to be moved to mwtab.py and then imported into cli.py.
5. Instead of using the requests package you can reuse code from fileio.py. You can look at "download_and_save_mwrest_file" in cli.py for how it does it.

Note that after my next commit "download_and_save_mwrest_file" changes a lot since I moved the results file stuff to MWRESTFile.

[hunter-moseley]

Sagar,

Let's move the rest of your discussion off this issue and into email.