Extract Python Tracebacks from text
A module & command-line tool for extracting Python tracebacks from text.
$ tbgrep file1 file2 file3
$ tail -f logfile | tbgrep
$ tbgrep --stats logfile [...] == 99 occurences ================================================== Traceback (most recent call last): File "/usr/lib/python2.4/site-packages/bodhi/admin.py", line 209, in _masher_request req_params=kwargs) File "/usr/lib/python2.4/site-packages/fedora/client/proxyclient.py", line 285, in send_request raise AuthError(_('Unable to log into server. Invalid' AuthError: Unable to log into server. Invalid authentication tokens. Send new username and password ================================================================== 733 unique tracebacks extracted
Once instantiated, you pass each line to the process method, which will return either None, or a string of a traceback.
from tbgrep import TracebackGrep extractor = TracebackGrep() for line in file('logfile'): tb = extractor.process(line) if tb: print(tb)
Instead of displaying each traceback found in the file, tbgrep also supports generating statistics about all tracebacks in the file.
extractor = TracebackGrep(stats=True) for line in file('logfile'): extractor.process(line) extractor.print_stats()
There are also some functions that allow you to search for tracebacks in a more convenient way.
from tbgrep import ( tracebacks_from_lines, tracebacks_from_file, last_traceback_from_file) for tb in tracebacks_from_file(file('logfile')): print(tb) for tb in tracebacks_from_file(file('logfile'), reverse=True): print(tb) print(last_traceback_from_file(file('logfile')))
tbgrep can extract tracebacks from logs of various formats. For example, CherryPy starts the traceback on a line with other details (like module, timestamp, etc), but the rest of the trace starts at the beginning of the line. Apache logs, on the other hand, will prefix each line of the traceback with this information. tbgrep is designed to handle these kinds of situations