I have download trevin.gandhi's code and i found that my code is same with him, but i failed twice. why?
You are using pip version 9.0.1, however version 18.0 is available.
You should consider upgrading via the 'pip install --upgrade pip' command.
Traceback (most recent call last):
File "/tmp/codalab/tmpB9zIc5/run/ingestion_program/ingestion.py", line 215, in <module>
D = DataManager(basename, input_dir, replace_missing=True, max_samples=max_samples, verbose=verbose,testdata=0)
File "/tmp/codalab/tmpB9zIc5/run/ingestion_program/data_manager.py", line 82, in __init__
self.getInfo (info_file)
File "/tmp/codalab/tmpB9zIc5/run/ingestion_program/data_manager.py", line 356, in getInfo
self.getFormatData(os.path.join(input_dir, basename + '_train1.data'))
File "/tmp/codalab/tmpB9zIc5/run/ingestion_program/data_manager.py", line 404, in getFormatData
data = data_converter.file_to_array (filename)
File "/tmp/codalab/tmpB9zIc5/run/ingestion_program/data_converter.py", line 24, in file_to_array
with open(filename, "r") as data_file:
IOError: [Errno 2] No such file or directory: '/tmp/codalab/tmpB9zIc5/run/input/B_train1.data'
Hi,
Please look at this post:
https://competitions.codalab.org/forums/16559/2584/
Posted by: hugo.jair @ Aug. 10, 2018, 7:02 a.m.I have the same issue.
Will you post in the forum as soon as you have fixed this bug?
Posted by: OhYeah @ Aug. 10, 2018, 8:26 a.m.Of course, that is the idea.
Thank you for your comprehension!
Posted by: hugo.jair @ Aug. 10, 2018, 8:32 a.m.Thanks for your reply, I am not fix this problem and i finish my own code with the templete and won the forst prize?(maybe), although i am the first, but i am not sure that why auc is lower and better=.=
Posted by: Lean-Y @ Aug. 10, 2018, 12:23 p.m.Hi,
I am sorry, I did not understand what you were asking. Could you please rephrase? many thanks in advance
Best
I just had a similar error:
You are using pip version 9.0.1, however version 18.0 is available.
You should consider upgrading via the 'pip install --upgrade pip' command.
Traceback (most recent call last):
File "/tmp/codalab/tmpOoRovC/run/ingestion_program/ingestion.py", line 215, in <module>
D = DataManager(basename, input_dir, replace_missing=True, max_samples=max_samples, verbose=verbose,testdata=0)
File "/tmp/codalab/tmpOoRovC/run/ingestion_program/data_manager.py", line 82, in __init__
self.getInfo (info_file)
File "/tmp/codalab/tmpOoRovC/run/ingestion_program/data_manager.py", line 356, in getInfo
self.getFormatData(os.path.join(input_dir, basename + '_train1.data'))
File "/tmp/codalab/tmpOoRovC/run/ingestion_program/data_manager.py", line 404, in getFormatData
data = data_converter.file_to_array (filename)
File "/tmp/codalab/tmpOoRovC/run/ingestion_program/data_converter.py", line 24, in file_to_array
with open(filename, "r") as data_file:
IOError: [Errno 2] No such file or directory: '/tmp/codalab/tmpOoRovC/run/input/C_train1.data'
By looking at ingestion_program_stdout_file.txt
I see that for the datasets "AA" and "B", the loading worked fine:
Info file found : /tmp/codalab/tmpOoRovC/run/input/AA/AA_public.info
Info file found : /tmp/codalab/tmpOoRovC/run/input/B/B_public.info
But somehow for the dataset "C" it didn't work:
Info file NOT found : /tmp/codalab/tmpOoRovC/run/input/C_public.info
The issue is that the file "C_public.info" is not directly in the folder "input" but in the folder "input/C".
Why is it looking directly in "input" instead of "input/C", but for other datasets like "AA" and "B" it worked for me? Whereas for the author of this thread it failed on "B".
I think I found out why: the issue come from the file "data_manager.py" at lines 68-69.
if basename in input_dir or os.path.isfile(os.path.join(input_dir, basename + '_train.data')) :
self.input_dir = input_dir
else:
self.input_dir = os.path.join (input_dir , basename )
I suspect that the bug come from "basename in input_dir" because:
For me it failed on the dataset "C" because for my worker on Codelab , my "input_dir" was "'/tmp/codalab/tmpOoRovC/run/input" which contain the Upper case letter "C".
For the author of the thread it failed on the dataset "B" because for his worker on Codelab his "input_dir" was "/tmp/codalab/tmpB9zIc5/run/input" which contain the Upper case letter "B".
A possible solution is to remove lines 68, 69 and 70 to keep only line 70:
self.input_dir = os.path.join (input_dir , basename )
Hi,
Thank you for the suggestion, but this does not fix the problem either. In either case, the public_info files are located in the same directory. Thanks anyways!
best
Posted by: hugo.jair @ Aug. 13, 2018, 4:41 p.m.