MassIVE has changed the FTP server URLs

[madina1203]

There is an issue when I want to download the file. I have watched that there is no delay between retries, but code is running perfectly 8 secs, which seems weird. It is more like connection interruption somehow. Is there any other way to solve this issue?

proj=ppx.find_project("MSV000090001", fetch=True, timeout=30)
proj._parser.max_reconnects = 50
proj.download("raw/RP_IDX_RAW/yeast_H2O_OTMS2_IDX_2_MminusHOnly.raw")

[madina1203]

I increased retry number to 50, but it is not helpful

[wfondrie]

Hi @madina1203 - thanks for sharing your issue!

Can you post the error that you receive as well? Unfortunately, sometimes the repositories change their API suddenly and I suspect that may be the case here.

[wfondrie]

I was able to confirm that MassIVE now adds v01 to their FTP site URLs, which is causing the issue.

What used to be:

ftp://massive.ucsd.edu/MSV000090001

Is now:

ftp://massive.ucsd.edu/v01/MSV000090001

A patch is inbound!

[madina1203]

This was the error: Traceback (most recent call last): File "/Users/madinabekbergenova/Desktop/PhD/phd_ms_spectrum/main.py", line 5, in <module> downloaded_raw = proj.download("raw/RP_IDX_RAW/yeast_H2O_OTMS2_IDX_2_MminusHOnly.raw") File "/Users/madinabekbergenova/Desktop/PhD/phd_ms_spectrum/venv/lib/python3.10/site-packages/ppx/project.py", line 238, in download return self._parser.download( File "/Users/madinabekbergenova/Desktop/PhD/phd_ms_spectrum/venv/lib/python3.10/site-packages/ppx/ftp.py", line 262, in download self._download_file( File "/Users/madinabekbergenova/Desktop/PhD/phd_ms_spectrum/venv/lib/python3.10/site-packages/ppx/ftp.py", line 135, in _download_file self.connect() File "/Users/madinabekbergenova/Desktop/PhD/phd_ms_spectrum/venv/lib/python3.10/site-packages/ppx/ftp.py", line 84, in connect self._with_reconnects(self._connect) File "/Users/madinabekbergenova/Desktop/PhD/phd_ms_spectrum/venv/lib/python3.10/site-packages/ppx/ftp.py", line 113, in _with_reconnects raise error_temp( ftplib.error_temp: Failed after 50 reconnect(s), the last error was: 550 Failed to change directory. Thank you for your reply.

[wfondrie]

Alas, it looks like this will more problematic than I originally thought. There are currently now 7 subdirectories in MassIVE (v0[1-7]), which will take a bit of work to account for.

[RobAlbn]

Hi @wfondrie, While you are working on fixing this bug, is there any way to tell ppx the correct URL and download files from MassIVE anyway? I tried to change the value of the property url in different ways, but it didn't work. Thank you very much!

[RobAlbn]

Hi again @wfondrie, I was able to circumvent the bug by following S.B's second piece of advice here: https://stackoverflow.com/questions/76881788/how-do-i-replace-a-python-property-without-a-setter-for-an-instance. Basically, I changed the value of the property url so that it is the correct URL of the dataset I would like to download. Apparently, it's working since it has been downloading for more than 10 minutes now.

[RobAlbn]

Thank you for linking the pull request! Is it already possible to use conda to install the ppx version that correctly retrieves the MassIVE FTP URLs?

[wfondrie]

The updated version is now available on PyPI, so you can install it with pip. However, it should be updated on Bioconda within a few hours!

[RobAlbn]

Thank you very much! And kudos on ppx since it greatly facilitates proteomic data download