[Reprozip-users] Fwd: Reprozip Project
remi.rampin at nyu.edu
Tue Oct 31 14:30:49 EDT 2017
2017-10-31 10:21 EDT, Ali Salari <ali.salar007 at gmail.com>:
> I appreciate your complete response, I've updated the graph.py file and it
> works correctly.
> Actually, my problem remains unsolved. I've included just a few files but
> those are still too large to display. I have a database with 8, 10, 5
> thousands of processes, opened files and executed files respectively.
> have you pleased an idea to display such amount of data?
> on the other hand, I've reduced the database size, how can I put the
> modified DB address instantly as the input DB of graph.py?
The graph command expects both the trace.sqlite3 database and the
config.yml file (it reads the name of runs, and input/outputs from there).
If you unpack those to a directory, you can use reprounzip graph -d
thedirectory graphoutput.dot as you curate the database.
I know that graphs are usually much too big to be displayed properly. The
only facilities that the graph commands are to make those easier are:
- Hiding files that are from packages/not from packages/not input files
- Hiding files that match a regex
- Regrouping files that match a regex or a prefix to limit the number of
nodes and edges
This is actually very versatile, although it is not very easy to use. But
you should also be able to poke at the trace database directly.
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the Reprozip-users