Re: [Challenge2022] datanetAPI fails when dataset name is over 100 characters

Hi, After running simulation i got an error reading the resulting tar file with the DatanetAPI iterator.   File "RouteNet_Fermi/datanetAPI.py", line 805, in __iter__     results_file = tar.extractfile(dir_info.name <http://dir_info.name>+"/simulationResults.txt")   File ".conda/envs/gnnch/lib/python3.9/tarfile.py", line 2113, in extractfile     tarinfo = self.getmember(member)   File ".conda/envs/gnnch/lib/python3.9/tarfile.py", line 1799, in getmember     raise KeyError("filename %r not found" % name) KeyError: "filename 'results_2022-10-03__23-42-06__hard3_lbw_from_traffic-above_bsz_small_2_avgbw_large_1_netsz_10_100x1_p_0.5_29_0_0//simulationResults.txt' not found" But the file simulationResults.txt is present in the tarfile. The reason seems to be that tarfile library adds a slash / character at the end of dir_info.name <http://dir_info.name> when the name is over 100 characters. And since DatanetAPI code adds a '/' when calling extractfile, this results in double slash which is apparently invalid. I fixed the error locally in datanetAPI.py by checking if the directory name has / at the end and if yes, to remove it. But i would appreciate it if you update your API so that the samples we submit that have long dir names could still be opened on your end. Thanks _______________________________________________ Challenge2022 mailing list Challenge2022(a)bnn.upc.edu https://mail.bnn.upc.edu/cgi-bin/mailman/listinfo/challenge2022