Some CRUS stats

[firejuggler]

http://mersenneforum.org/showpost.ph...51&postcount=1
first post of the Software/instructions/questions thread.

[rogue]

It isn't clear to me if those are the current scripts.

[rogue]

I presume that the links in this post have the current scripts.

[mdettweiler]

Originally Posted by rogue

I presume that the links in this post have the current scripts.

Indeed, they do--those are the latest scripts that had been sent to me and are currently in use on the website.

[rogue]

I'm trying to run them, but the version of awk on lion doesn't support strtonum(). I will try to install gawk from macports and see if that will do the trick.

[rogue]

You can add this to the 21_main.txt script. You can call the tables whatever you want. I have Table05 and Table07 (I had renumbered them in my copy of the script). You then have to modify 42_final_html.txt to add these two tables to the html generation. Marking conjectures are reserved is a bit more work.

## top20 unproven by test limit ########
END {cct1=0
for (q=1; q<=NR; q++) {
dlim[q] = "0000000000"
if (searchlim[q]!="proven" && searchlim[q]!="part tested") {
if (length(senotemp[q])==3) {dlim[q] = "0000000" senotemp[q]}
if (length(senotemp[q])==4) {dlim[q] = "000000" senotemp[q]}
if (length(senotemp[q])==5) {dlim[q] = "00000" senotemp[q]}
if (length(senotemp[q])==6) {dlim[q] = "0000" senotemp[q]}
if (length(senotemp[q])==7) {dlim[q] = "000" senotemp[q]}
if (length(senotemp[q])==8) {dlim[q] = "00" senotemp[q]}
if (length(senotemp[q])==9) {dlim[q] = "0" senotemp[q]}
dalim[q] = senotemp[q]
ckpro[q]=ck[q];basepro[q]=base[q];highnpro[q]=highn[q]
}
}
n1=asort(dlim,dtemp1)
#
print "\n<a name=\"Table05\"></a><h4>Top-20 unproven conjectures by test limit</h4></tr>" > "temp_"projshort"_html.htm"
print projbeg"<table border=\"1\">\n" > "temp_"projshort"_html.htm"
print projbeg"<tr class=\"toprow\">\n<th colspan=\"4\">"projlong"</th>\n"projend"</tr>\n" > "temp_"projshort"_html.htm"
print projbeg"<tr class=\"secrow\">\n<th>Base</th>\n<th>CK</th>\n<th>k remaining</th>\n<th>Test Limit</th>\n"projend"</tr>\n" > "temp_"projshort"_html.htm"
#
print "Top-20 unproven conjectures by test limit\n" > "temp_"projshort"_tables.txt"
printf (" %s \n",projlong) > "temp_"projshort"_tables.txt"
printf (" %5s %10s %15s %12s \n", " Base", "CK", "k remaining", "Test Limit") > "temp_"projshort"_tables.txt"

for (z=n1; z>=n1-19; z--) {
if (cct1==20) {break}
for (j=q; j>=1; j--) { if (dlim[j]==dtemp1[z]) {
if (dtemp[z]=="0") {continue}
if ((dtemp[z]==dtemp1[z-1]) && (cct1<19)) {continue}
printf ("| %5s %10s %10s %12s |\n",base[j],ck[j],kremain[j],dalim[j]) > "temp_"projshort"_tables.txt"
printf (projbeg"<tr align=\"right\">\n<td>%s</td>\n<td>%s</td>\n<td>%s</td><td>%s</td>\n"projend"</tr>\n", base[j], ck[j], kremain[j], dalim[j]) > "temp_"projshort"_html.htm"
cct1++
delete dlim[j]
}
if (cct1==20) {break}
}
}
print projend"</table>\n" > "temp_"projshort"_html.htm"
}


## top20 unproven by test limit decimal length ########
END {cct1=0
for (q=1; q<=NR; q++) {
dlim[q] = "0000000000"
dlen[q] = "0"
if (searchlim[q]!="proven" && searchlim[q]!="part tested") {
dalim[q] = senotemp[q]
dalen[q] = sprintf("%.0f", strtonum(senotemp[q]) * log(strtonum(base[q])) / log(10))
if (length(dalen[q])==3) {dlen[q] = "0000000" dalen[q]}
if (length(dalen[q])==4) {dlen[q] = "000000" dalen[q]}
if (length(dalen[q])==5) {dlen[q] = "00000" dalen[q]}
if (length(dalen[q])==6) {dlen[q] = "0000" dalen[q]}
if (length(dalen[q])==7) {dlen[q] = "000" dalen[q]}
if (length(dalen[q])==8) {dlen[q] = "00" dalen[q]}
if (length(dalen[q])==9) {dlen[q] = "0" dalen[q]}
}
}
n1=asort(dlen,dtemp1)
#
print "\n<a name=\"Table07\"></a><h4>Top-20 unproven conjectures by test limit decimal length</h4></tr>" > "temp_"projshort"_html.htm"
print projbeg"<table border=\"1\">\n" > "temp_"projshort"_html.htm"
print projbeg"<tr class=\"toprow\">\n<th colspan=\"5\">"projlong"</th>\n"projend"</tr>\n" > "temp_"projshort"_html.htm"
print projbeg"<tr class=\"secrow\">\n<th>Base</th>\n<th>CK</th>\n<th>k remaining</th>\n<th>Test Limit</th><th>Decimal Length</th>\n"projend"</tr>\n" > "temp_"projshort"_html.htm"
#
print "Top-20 unproven conjectures by test limit decimal length\n" > "temp_"projshort"_tables.txt"
printf (" %s \n",projlong) > "temp_"projshort"_tables.txt"
printf (" %5s %10s %15s %12s %8s \n", " Base", "CK", "k remaining", "Test Limit", "Decimal Length") > "temp_"projshort"_tables.txt"

for (z=n1; z>=n1-19; z--) {
if (cct1==20) {break}
for (j=q; j>=1; j--) { if (dlen[j]==dtemp1[z]) {
if (dtemp[z]=="0") {continue}
if ((dtemp[z]==dtemp1[z-1]) && (cct1<19)) {continue}
printf ("| %5s %10s %10s %12s %8s |\n",base[j],ck[j],kremain[j],dalim[j],dalen[j]) > "temp_"projshort"_tables.txt"
printf (projbeg"<tr align=\"right\">\n<td>%s</td>\n<td>%s</td>\n<td>%s</td><td>%s</td><td>%s</td>\n"projend"</tr>\n", base[j], ck[j], kremain[j], dalim[j], dalen[j]) > "temp_"projshort"_html.htm"
cct1++
delete dlen[j]
}
if (cct1==20) {break}
}
}
print projend"</table>\n" > "temp_"projshort"_html.htm"
}

[mdettweiler]

Hmm...okay. I'm not particularly familiar with the inner workings of the scripts myself; basically I've just plopped them into the web site as-is and set up cron jobs to run them on a schedule. Could you possibly send me your latest version of the complete files for the modified scripts (which I'm guessing you have anyway from testing things on your end)? Thanks.

[rogue]

I've attached the two scripts I modified.

[rogue]

I'm working on the changes to show reserved bases on the crus-stats page. In the process I'm going to clean up the awk scripts. They could use a little reformatting to enhance readability.

[rogue]

I've attached the updated scripts. I did the following:

1) wget the reservation pages
2) Use the reservation pages to show reserved conjectures in yellow
3) Add two tables: "Top-20 unproven conjectures by highest tested n" and "Top-20 unproven conjectures by highest tested decimal length"
4) Cleaned up the scripts so enhance readability (and maintainability)
5) Add a "partially tested" column. This reflects bases where some k, but not all k have been tested.
6) Replace text like this "There are 7 conjectures with CK>=1T (0.7% of total). 85.7% remain untested, 0.0% are proven" with a table.
7) Fixed the count of calculation of "largest group of consecutive proven" values. Under some conditions, this could produce erroneous results.

For some reason, my build of gawk had a problem with the built-in getline() function. It works fine in BEGIN sections, but not in END sections. This causes the weight to not be correct. Someone will need to run this script independently to ensure that it populates that table correctly.

You can see the new layout here. Of course Gary will need to decide if he wants to run with these as I can't update the server.

In theory, it should be easy to modify the scripts behind http://www.noprimeleftbehind.net/crus/tab/CRUS_tab.htm to do something similar, but I will leave that to the interested reader.

In theory, it should be easy to use the scripts to build an html page that has all reservations (and just reservations). I view that as a quick reference page to see what is reserved, when and by whom. IMO, it would only show conjectures with reservations, thus be a much smaller version of the current reservation pages. Again, that can be left to the interested reader.

[Batalov]

"Top-20 Proven Conjectures by Highest Conjectured K" section (vstats/crus-stats.htm#Table10) appears to be sorted reverse-literal (not reverse-numeric): CK of 98 is top in both sections.