Hi!
I have vdr 1.7.27 and epgsearch 1.0.1.beta2, from yavdr..
I'm recording "Game of Thrones" from Canal Digital Nordic. The third season is starting in a few days..
The episodes have no episode name, and the description field simply says: "(1:10/s3) 3. kausi maailman suosituimmasta HBO-sarjasta."
The text part is always the same, just the episode number varies. So just a single character difference. But epgsearch thinks these are repeats! Even if I set required match to 100%, it still thinks they are repeats.. shouldn't 100% mean exact match?
Hi,
probably the following code is the reason for this behaviour:
epgsearchtools.c:
// last try with Levenshtein Distance, only compare the first 1000 chars double fMatch = FuzzyMatch(eDescr, rDescr, 1000); double tmp_matchlimit = matchLimit/100.0; if(maxLength - minLength < 5) { tmp_matchlimit = 0.95; LogFile.Log(2,"difference between both descriptions is < 5 setting matchlimit to: %.2f %%", tmp_matchlimit*100); }
Please change the line
if(maxLength - minLength < 5)
to
if(maxLength - minLength < 5 && matchLimit < 95)
and let me know if it works for you.
cheers, Christian
Am 04.04.2013 20:04, schrieb Teemu Suikki:
Hi!
I have vdr 1.7.27 and epgsearch 1.0.1.beta2, from yavdr..
I'm recording "Game of Thrones" from Canal Digital Nordic. The third season is starting in a few days..
The episodes have no episode name, and the description field simply says: "(1:10/s3) 3. kausi maailman suosituimmasta HBO-sarjasta."
The text part is always the same, just the episode number varies. So just a single character difference. But epgsearch thinks these are repeats! Even if I set required match to 100%, it still thinks they are repeats.. shouldn't 100% mean exact match?
vdr mailing list vdr@linuxtv.org http://www.linuxtv.org/cgi-bin/mailman/listinfo/vdr
I tried this, no change..
Then I added some debug. For some reason, matchLimit is 90 even if I have set it to 100! There must be some other bug that resets matchLimit..
If I set matchLimit=100 inside the function, it works as expected if your patch is installed. 5.4.2013 0.32 "Christian Wieninger" cwieninger@gmx.de kirjoitti:
Hi,
probably the following code is the reason for this behaviour:
epgsearchtools.c:
// last try with Levenshtein Distance, only compare the first 1000 chars double fMatch = FuzzyMatch(eDescr, rDescr, 1000); double tmp_matchlimit = matchLimit/100.0; if(maxLength - minLength < 5) { tmp_matchlimit = 0.95; LogFile.Log(2,"difference between both descriptions is < 5 setting matchlimit to: %.2f %%", tmp_matchlimit*100); }
Please change the line
if(maxLength - minLength < 5)
to
if(maxLength - minLength < 5 && matchLimit < 95)
and let me know if it works for you.
cheers, Christian
Am 04.04.2013 20:04, schrieb Teemu Suikki:
Hi!
I have vdr 1.7.27 and epgsearch 1.0.1.beta2, from yavdr..
I'm recording "Game of Thrones" from Canal Digital Nordic. The third season is starting in a few days..
The episodes have no episode name, and the description field simply says: "(1:10/s3) 3. kausi maailman suosituimmasta HBO-sarjasta."
The text part is always the same, just the episode number varies. So just a single character difference. But epgsearch thinks these are repeats! Even if I set required match to 100%, it still thinks they are repeats.. shouldn't 100% mean exact match?
vdr mailing listvdr@linuxtv.orghttp://www.linuxtv.org/cgi-bin/mailman/listinfo/vdr
Ok, I fixed it..
In file epgsearchext.c, function CheckRepeatTimers, there is a line:
if (EventsMatch(pEvent, pEventP, compareTitle, compareSubtitle, compareSummary, compareDate, catvaluesAvoidRepeat))
There is no match limit given! So I added it like this:
if (EventsMatch(pEvent, pEventP, compareTitle, compareSubtitle, compareSummary, compareDate, catvaluesAvoidRepeat, compareSummaryMatchInPercent))
Seems to work now. Still needs your earlier patch in epgsearchtools.c of course.
5.4.2013 5.48 "Teemu Suikki" tsuikki@zuik.org kirjoitti:
I tried this, no change..
Then I added some debug. For some reason, matchLimit is 90 even if I have set it to 100! There must be some other bug that resets matchLimit..
If I set matchLimit=100 inside the function, it works as expected if your patch is installed. 5.4.2013 0.32 "Christian Wieninger" cwieninger@gmx.de kirjoitti:
Hi,
probably the following code is the reason for this behaviour:
epgsearchtools.c:
// last try with Levenshtein Distance, only compare the first 1000 chars double fMatch = FuzzyMatch(eDescr, rDescr, 1000); double tmp_matchlimit = matchLimit/100.0; if(maxLength - minLength < 5) { tmp_matchlimit = 0.95; LogFile.Log(2,"difference between both descriptions is < 5 setting matchlimit to: %.2f %%", tmp_matchlimit*100); }
Please change the line
if(maxLength - minLength < 5)
to
if(maxLength - minLength < 5 && matchLimit < 95)
and let me know if it works for you.
cheers, Christian
Am 04.04.2013 20:04, schrieb Teemu Suikki:
Hi!
I have vdr 1.7.27 and epgsearch 1.0.1.beta2, from yavdr..
I'm recording "Game of Thrones" from Canal Digital Nordic. The third season is starting in a few days..
The episodes have no episode name, and the description field simply says: "(1:10/s3) 3. kausi maailman suosituimmasta HBO-sarjasta."
The text part is always the same, just the episode number varies. So just a single character difference. But epgsearch thinks these are repeats! Even if I set required match to 100%, it still thinks they are repeats.. shouldn't 100% mean exact match?
vdr mailing listvdr@linuxtv.orghttp://www.linuxtv.org/cgi-bin/mailman/listinfo/vdr
Am 05.04.2013 05:21, schrieb Teemu Suikki:
Ok, I fixed it..
In file epgsearchext.c, function CheckRepeatTimers, there is a line:
if (EventsMatch(pEvent, pEventP, compareTitle, compareSubtitle, compareSummary, compareDate, catvaluesAvoidRepeat))
There is no match limit given! So I added it like this:
if (EventsMatch(pEvent, pEventP, compareTitle, compareSubtitle, compareSummary, compareDate, catvaluesAvoidRepeat, compareSummaryMatchInPercent))
Seems to work now. Still needs your earlier patch in epgsearchtools.c of course.
thanks for the feedback. After testing it in my environment I will push it to the git.
cheers, Christian
5.4.2013 5.48 "Teemu Suikki" <tsuikki@zuik.org mailto:tsuikki@zuik.org> kirjoitti:
I tried this, no change.. Then I added some debug. For some reason, matchLimit is 90 even if I have set it to 100! There must be some other bug that resets matchLimit.. If I set matchLimit=100 inside the function, it works as expected if your patch is installed. 5.4.2013 0.32 "Christian Wieninger" <cwieninger@gmx.de <mailto:cwieninger@gmx.de>> kirjoitti: Hi, probably the following code is the reason for this behaviour: epgsearchtools.c: // last try with Levenshtein Distance, only compare the first 1000 chars double fMatch = FuzzyMatch(eDescr, rDescr, 1000); double tmp_matchlimit = matchLimit/100.0; if(maxLength - minLength < 5) { tmp_matchlimit = 0.95; LogFile.Log(2,"difference between both descriptions is < 5 setting matchlimit to: %.2f %%", tmp_matchlimit*100); } Please change the line if(maxLength - minLength < 5) to if(maxLength - minLength < 5 && matchLimit < 95) and let me know if it works for you. cheers, Christian Am 04.04.2013 20:04, schrieb Teemu Suikki:
Hi! I have vdr 1.7.27 and epgsearch 1.0.1.beta2, from yavdr.. I'm recording "Game of Thrones" from Canal Digital Nordic. The third season is starting in a few days.. The episodes have no episode name, and the description field simply says: "(1:10/s3) 3. kausi maailman suosituimmasta HBO-sarjasta." The text part is always the same, just the episode number varies. So just a single character difference. But epgsearch thinks these are repeats! Even if I set required match to 100%, it still thinks they are repeats.. shouldn't 100% mean exact match? _______________________________________________ vdr mailing list vdr@linuxtv.org <mailto:vdr@linuxtv.org> http://www.linuxtv.org/cgi-bin/mailman/listinfo/vdr
vdr mailing list vdr@linuxtv.org http://www.linuxtv.org/cgi-bin/mailman/listinfo/vdr