viewerstats.py

#!runpy.sh

"""\

This module contains code for analyzing ViewerStats data as uploaded by the viewer.

$LicenseInfo:firstyear=2021&license=viewerlgpl$
Second Life Viewer Source Code
Copyright (C) 2021, Linden Research, Inc.

This library is free software; you can redistribute it and/or
modify it under the terms of the GNU Lesser General Public
License as published by the Free Software Foundation;
version 2.1 of the License only.

This library is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public
License along with this library; if not, write to the Free Software
Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA  02110-1301  USA

Linden Research, Inc., 945 Battery Street, San Francisco, CA  94111  USA
$/LicenseInfo$
"""

import argparse
import numpy as np
import pandas as pd
import json
from collections import Counter, defaultdict
from llbase import llsd
import io
import re
import os
import sys

def show_stats_by_key(recs,indices,settings_sd = None):
    result = ()
    cnt = Counter()
    per_key_cnt = defaultdict(Counter)
    for r in recs:
        try:
            d = r
            for idx in indices:
                d = d[idx]
            for k,v in d.items():
                if isinstance(v,dict):
                    continue
                cnt[k] += 1
                if isinstance(v,list):
                    v = tuple(v)
                per_key_cnt[k][v] += 1
        except Exception as e:
            print "err", e
            print "d", d, "k", k, "v", v
            raise
    mc = cnt.most_common()
    print "========================="
    keyprefix = ""
    if len(indices)>0:
        keyprefix = ".".join(indices) + "."
    for i,m in enumerate(mc):
        k = m[0]
        bigc = m[1]
        unset_cnt = len(recs) - bigc
        kmc = per_key_cnt[k].most_common(5)
        print i, keyprefix+str(k), bigc
        if settings_sd is not None and k in settings_sd and "Value" in settings_sd[k]:
            print "    ", "default",settings_sd[k]["Value"],"count",unset_cnt
        for v in kmc:
            print "    ", "value",v[0],"count",v[1]
    if settings_sd is not None:
        print "Total keys in settings", len(settings_sd.keys())
        unused_keys = list(set(settings_sd.keys()) - set(cnt.keys()))
        unused_keys_non_str = [k for k in unused_keys if settings_sd[k]["Type"] != "String"]
        unused_keys_str = [k for k in unused_keys if settings_sd[k]["Type"] == "String"]

        # Things that no one in the sample has set to a non-default value. Possible candidates for removal.
        print "\nUnused_keys_non_str", len(unused_keys_non_str)
        print   "======================"
        print "\n".join(sorted(unused_keys_non_str))

        # Strings are not currently logged, so we have no info on usage.
        print "\nString keys (usage unknown)", len(unused_keys_str)
        print   "======================"
        print "\n".join(sorted(unused_keys_str))

        # Things that someone has set but that aren't recognized settings.
        unrec_keys = list(set(cnt.keys()) - set(settings_sd.keys()))
        print "\nUnrecognized keys", len(unrec_keys)
        print   "======================"
        print "\n".join(sorted(unrec_keys))

        result = (settings_sd.keys(), unused_keys_str, unused_keys_non_str, unrec_keys)
    return result

def parse_settings_xml(fname):
    # assume we're in scripts/metrics
    fname = "../../indra/newview/app_settings/" + fname
    with open(fname,"r") as f:
        contents = f.read()
        return llsd.parse_xml(contents)

def read_raw_settings_xml(fname):
    # assume we're in scripts/metrics
    fname = "../../indra/newview/app_settings/" + fname
    contents = None
    with open(fname,"r") as f:
        contents = f.read()
    return contents

def write_settings_xml(fname, contents):
    # assume we're in scripts/metrics
    fname = "../../indra/newview/app_settings/" + fname
    with open(fname,"w") as f:
        f.write(llsd.format_pretty_xml(contents))
        f.close()

def write_raw_settings_xml(fname, string):
    # assume we're in scripts/metrics
    fname = "../../indra/newview/app_settings/" + fname
    with io.open(fname,"w", newline='\n') as f:
        f.write(string.decode('latin1'))
        f.close()

def remove_settings(string, to_remove):
    for r in to_remove:
        subs_str = r"<key>" + r + r"<.*?</map>\n"
        string = re.sub(subs_str,"",string,flags=re.S|re.DOTALL)
    return string

def get_used_strings(root_dir):
    used_str = set()
    skipped_ext = set()
    for dir_name, sub_dir_list, file_list in os.walk(root_dir):
        for fname in file_list:
            if fname in ["settings.xml", "settings.xml.edit", "settings_per_account.xml"]:
                print "skip", fname
                continue
            (base,ext) = os.path.splitext(fname)
            #if ext not in [".cpp", ".hpp", ".h", ".xml"]:
            #    skipped_ext.add(ext)
            #    continue
            
            full_name = os.path.join(dir_name,fname)

            with open(full_name,"r") as f:
                #print full_name
                lines = f.readlines()
                for l in lines:
                    ms = re.findall(r'[>\"]([A-Za-z0-9_]+)[\"<]',l)
                    for m in ms:
                        #print "used_str",m
                        used_str.add(m)
    print "skipped extensions", skipped_ext
    print "got used_str", len(used_str)
    return used_str
                
    
if __name__ == "__main__":

    parser = argparse.ArgumentParser(description="process tab-separated table containing viewerstats logs")
    parser.add_argument("--verbose", action="store_true",help="verbose flag")
    parser.add_argument("--preferences", action="store_true", help="analyze preference info")
    parser.add_argument("--remove_unused", action="store_true", help="remove unused preferences")
    parser.add_argument("--column", help="name of column containing viewerstats info")
    parser.add_argument("infiles", nargs="+", help="name of .tsv files to process")
    args = parser.parse_args()

    for fname in args.infiles:
        print "process", fname
        df = pd.read_csv(fname,sep='\t')
        #print "DF", df.describe()
        jstrs = df['RAW_LOG:BODY']
        #print "JSTRS", jstrs.describe()
        recs = []
        for i,jstr in enumerate(jstrs):
            recs.append(json.loads(jstr))
        show_stats_by_key(recs,[])
        show_stats_by_key(recs,["agent"])
        if args.preferences:
            print "\nSETTINGS.XML"
            settings_sd = parse_settings_xml("settings.xml")
            #for skey,svals in settings_sd.items(): 
            #    print skey, "=>", svals
            (all_str,_,_,_) = show_stats_by_key(recs,["preferences","settings"],settings_sd)
            print

            #print "\nSETTINGS_PER_ACCOUNT.XML"
            #settings_pa_sd = parse_settings_xml("settings_per_account.xml")
            #show_stats_by_key(recs,["preferences","settings_per_account"],settings_pa_sd)

            if args.remove_unused:
                # walk codebase looking for strings
                all_str_set = set(all_str)
                used_strings = get_used_strings("../../indra")
                used_strings_set = set(used_strings)
                unref_strings = all_str_set-used_strings_set
                # Some settings names are generated by appending to a prefix. Need to look for this case.
                prefix_used = set()
                print "checking unref_strings", len(unref_strings)
                for u in unref_strings:
                    for k in range(6,len(u)):
                        prefix = u[0:k]
                        if prefix in all_str_set and prefix in used_strings_set:
                            prefix_used.add(u)
                            #print "PREFIX_USED",u,prefix
                print "PREFIX_USED", len(prefix_used), ",".join(list(prefix_used))
                print
                unref_strings = unref_strings - prefix_used
                        
                print "\nUNREF_IN_CODE " + str(len(unref_strings)) + "\n"
                print "\n".join(list(unref_strings))
                settings_str = read_raw_settings_xml("settings.xml")
                # Do this via direct string munging to generate minimal changeset
                settings_edited = remove_settings(settings_str,unref_strings)
                write_raw_settings_xml("settings.xml.edit",settings_edited)