Adding the download button for QeAppWorkChains.

[mikibonacci]

This fixes #823

We provide a DownloadDataWidget to download, separately:

• AiiDA archive: export_{node.pk}.aiida
• AiiDA raw data export_{node.pk}_raw.zip

both are zip files.

[codecov]

Codecov Report

Attention: Patch coverage is 39.13043% with 42 lines in your changes missing coverage. Please review.

Project coverage is 67.10%. Comparing base (23c3fe1) to head (150d805).
Report is 122 commits behind head on main.

Files with missing lines	Patch %	Lines
src/aiidalab_qe/app/result/utils/download_data.py	37.31%	42 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #850      +/-   ##
==========================================
- Coverage   67.53%   67.10%   -0.43%     
==========================================
  Files          50       51       +1     
  Lines        4620     4688      +68     
==========================================
+ Hits         3120     3146      +26     
- Misses       1500     1542      +42     

Flag	Coverage Δ
python-3.11	67.10% <39.13%> (-0.43%)	⬇️
python-3.9	67.13% <39.13%> (-0.43%)	⬇️

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[mikibonacci]

Hi @superstar54 and @AndresOrtegaGuerrero , this PR is ready for review.

The main new thing is the DownloadDataWidget, in the new utils subfolder of the result one. I left mostly untouched the implementation of a download output button found in the WorkChainOutputs, as I have not fully understood how it works. I preferred to implement a widget which contains both archive and raw_data download buttons.

Files are downloaded as zip archives.

A part the implementation, which probably needs to be improved, only one issue: with google chrome the downloads are seen as insecure (I think because they are zip files, and we have http instead of https).
Could you please confirm that also you have this issue?

Thanks!

[AndresOrtegaGuerrero]

@mikibonacci thank you , do you think we can include the import aiida node , is too much for this PR ?

[mikibonacci]

Hi @AndresOrtegaGuerrero , thanks for looking at this PR. for the import button, I think I can do a second PR so we don't add too much to this one. But also, I am a bit concerned on what the user may try to upload... for example, maybe they expect that if a PwBaseWorkChain is imported, this will be visible in the app... but this is not always true (true only if we import the whole QeAppWorkChain). Actually, we should be safe enough, as the user typically will just export and import using the app buttons.

To summarize, I think we can do it in another PR, maybe not a critical one. We have already the upload widget from euphonic vibro-app, so should be easy to use similar code and just play with the import API of AiiDA

[mikibonacci]

Hi @AndresOrtegaGuerrero , I addressed the docstring and the download methods merging. I also put a reloading of the workchain node in the produce_bitestream method, as I encountered a problem with SQLAlchemy (apparently if I don't download again the node, the session in which I produce the files is different from the one of the notebook, and then it does not work - still <100% clear to me)

[AndresOrtegaGuerrero]

Hi @superstar54 and @AndresOrtegaGuerrero , this PR is ready for review.

The main new thing is the DownloadDataWidget, in the new utils subfolder of the result one. I left mostly untouched the implementation of a download output button found in the WorkChainOutputs, as I have not fully understood how it works. I preferred to implement a widget which contains both archive and raw_data download buttons.

Files are downloaded as zip archives.

A part the implementation, which probably needs to be improved, only one issue: with google chrome the downloads are seen as insecure (I think because they are zip files, and we have http instead of https). Could you please confirm that also you have this issue?

Thanks!

One question why the aiida file is saved as a zip?

[mikibonacci]

Hi @AndresOrtegaGuerrero , it is just the standard archive format when you create an archive (see this in aiida-core create.py file at line 173.
Maybe I can just put the button description to be "Download AiiDA archive data" and not "Download AiiDA archive zip data", as it may be a bit misleading (the user may try to unzip the archive file).

What do you think?

[superstar54]

Hi @mikibonacci , thanks for the work!

One concern regarding large data. What if the download takes long time? Will it block the whole notebook?

I clicked the download button, and then went the step 1, I got this error:

DetachedInstanceError: Instance <DbNode at 0x7f05fd706280> is not bound to a Session; attribute refresh operation cannot proceed (Background on this error at: https://sqlalche.me/e/20/bhk3)

[superstar54]

Hi @mikibonacci , thanks for the update. It's a good way to download it in another thread!

Maybe one bug: when download the raw data, it always in this status: Downloading data...

[mikibonacci]

Hi @superstar54 , thanks for reviewing the PR. I addressed your valuable comments and suggestions.

However, I don't understand the bug you mention: are the data downloaded in your local machine but the "Downloading data..." still there?

Because if the data are a lot, it will take a while, and I think we should leave the string there up to completion

[superstar54]

In my test, tt's a small calculation, the aidia file is downloaded within 5 s. But for the raw data, after 10 mins, it still shows Downloading data..., and no data is not downloaded.

[mikibonacci]

Hi @superstar54 , I think I solved the issue. I just disable the raw data download if the ProcessDumper cannot be imported. I have not found a way to catch the error from the second thread (for now).

[superstar54]

LGTM!