Solve minor issue for h5 files #268
Conversation
To solve error issue when using h5 files (old python digitization) AttributeError: 'Values' object has no attribute 'tmpname'
|
I found an easier way to fix this. Just by not flipping the image when reading h5 files. I'll update the PR soon. Hopefully h5 fiels won't be used anymore. |
| @@ -892,7 +892,7 @@ def reconstruct(self,evrange=(-1,-1,-1)): | |||
| postfix = 'mid.gz' | |||
| options.tmpname = "%s/%s%05d.%s" % (tmpdir,prefix,int(options.run),postfix) | |||
| else: | |||
| if options.rawdata_tier == 'root': | |||
| if options.rawdata_tier == 'root' or options.rawdata_tier == 'h5': | |||
| file_url = sw.swift_root_file(options.tag, int(options.run)) | |||
There was a problem hiding this comment.
Why was this added? swift_root_file returns a .root address which cannot work for h5, right?. If you are not downloading the h5 files, I would add a if statement and a print with a sys.exit() in case you reach this part of the code and it seems you need to download a h5 file. Now h5 is not downloadable and we will add this feature when/if it will be necessary. Does this make sense?
utilities.py
Outdated
| if rawdata_tier == 'h5': | ||
| img_fr = rootfile[key][:].T #necessary because uproot inverts column and rows with x and y | ||
| else: | ||
| img_fr = rootfile[key].values().T #necessary because uproot inverts column and rows with x and y |
There was a problem hiding this comment.
the if tag=='MAN' needs to be inside the root tier case
This way there is no need to change the utilities.py
|
Now everything should be better. The actual changes are only in the recostruction.py at: line 499 The rest (spaces and blank lines) must be a bug of github |
No description provided.