EBTST - EBT stats tool

[jrnsndrs]

(...)
@jrnsndrs: all your dutch notes (no exception—congrats!) have the full post code (9999 ZZ). This isn't matched by EBTST now, so you end up with no notes in the Netherlands as far as regions is concerned . No need to ask your fellow trackers, the problem is identified .

Due to the headache, runny nose and continuous sneezing I'm having, I'm afraid the solution to this will have to wait. At least I already have a solution in mind (which should also fix giulcenc's issue with ZBG 1000, ATD 4000 and BKR 9013).

I had already posted when I found the above message. I will report this back.

And get well soon.

[dserrano5]

All bugs squashed (in development version):

• Portuguese locations shouldn't be repeated (removed pt.txt).
• Regarding the regions section, some postal codes are given special treatment:
  • All countries: all spaces are removed.
  • Malta: if no matches for "AAA-1234" are found, then try "AAA".
  • Netherlands: if no matches for "1234 AA" are found, then try "1234".
  • Poland: if no matches for "12-345" are found, then try "12345" (I did this just for fun).
  • Portugal: if no matches for "1234-123" are found, then try "1234".
  • United Kingdom: if no matches for "AB12 3CD" (or "A12 3BC" or "AB1 2CD" or… ) are found, then try "AB" (or "A1" in that case).
  • United States: if no matches for "12345-1234" (or "AB, 12345") are found, then try "12345" (just for fun again).
• Fix the sorting of entries in regions, locations and hit locations regarding non-ASCII letters. Improved hit locations while I was at it (was being sorted only by number of hits, and location names were shuffled).
• Use a different separator for an internal data structure (fix to giulcenc's empty subgroup name).
• Region names were being only read only up to the first comma.

[jrnsndrs]

Regions are back here, thanks.

One little problem remains (that most likely must be solved in nl.txt):
The dots numbered 1-9 appear at the wrong places.

Regions_wrongly_placed_dots.PNG (15.16 KiB) Viewed 1372 times

I think you must renumber them to 01 - 09. At least such numbers work correctly with the biggest and smallest municipalities.

[lmviterbo]

Great job!

This made me detect two errors on the Cidades_e_Vilas txt file.

• The line

  Code: Select all

  2495;Fátima=Fátima

  is duplicated (so I get 4 notes in Fátima instead of the real 2 I registered there).
• The line

  Code: Select all

  1890;Moscavide=Moscavide

  should read

  Code: Select all

  1990;Moscavide=Moscavide

  the only change needed being 1990 instead of 1890.

If this kind of errors is bound to happen again, is it better for you that I report it here or download the txt file and send you a corrected version (or upload it)?


Now three requests. The first one is simple, the second one (about Locations, not Regions) most likely not, the third one (about Regions again) probably even less.

1. It would be great to have the totals in brackets after the titles of each group, as it's already done on the Locations page. Instead of Vilas de Portugal, I'd like to read Vilas de Portugal (72). This could be applied to all the groups and subgroups, for every country. (The perfect solution would be something like Vilas de Portugal (72/582), showing the total number of vilas in Portugal — although I believe this might be a little harder to implement.)
2. The list of Locations is showing 7 notes in Calheta. In fact, I have 4 notes in one relatively big town named Calheta (a "vila") and 3 other in another small town with the same name. So my total number of locations in Portugal is in fact not correct. EBT knows it's two different towns, because they're linked to different city-profiles. Is it possible to change *everything* in the program so that it reads city-profiles links (and its correct names) instead of the names of locations where we registered our notes? This would have an extra advantage: city managers would still need to manually subsume new mistyped names to city-profiles as they already do on the citymanager, but this tedious procedure would no longer be needed to be duplicated on EBTST txt files.
3. It would be amazing if we had a link to a note for every line on the Regions page, just like the ones on the Locations page.

[dserrano5]

The dots numbered 1-9 appear at the wrong places. […] I think you must renumber them to 01 - 09. At least such numbers work correctly with the biggest and smallest municipalities.

Some groups in Austria, Belgium, Germany, Spain and Italy have this problem. It was easier to change the sorting algorithm instead of going through all the configuration files.

If this kind of errors is bound to happen again, is it better for you that I report it here or download the txt file and send you a corrected version (or upload it)?

Here, or in a PM, is fine. Unless there are tons of changes, in which case I'd prefer someone else to do the work for me .

The list of Locations is showing 7 notes in Calheta. […] EBT knows it's two different towns, because they're linked to different city-profiles. Is it possible to change *everything* in the program so that it reads city-profiles links

The city profile information isn't included in the CSV file so there's no reliable way to know whether Calheta 9370 and Calheta 9400 are different places or simply different post codes within a large city. Calheta is actually a good example given that 9370 and 9400 aren't far away numbers.

The tally is doable, the linking to notes is in fact easiest of your 3 items .

[giulcenc]

• Regarding the regions section, some postal codes are given special treatment:
  • Malta: if no matches for "AAA-1234" are found, then try "AAA".

I notice you are using three files for Maltese data in Regions: mt.txt, my_giulcenc.txt and mt_yli.txt. The first and the third was made by yli, the second by me.

The first file (mt.txt) is based on locations only, and give stats about Local Councils generating the table named Local Councils / Kunsilli Lokali of Malta. It's not complete, because it counts locations only if written exactly as in the listing. All my Maltese notes are not counted by this file, due to the different location names.

The second file (mt_giulcenc.txt) is not updated since July 2008, and generated the tables Kunsilli Lokali ta' Malta / Local Councils of Malta and Islands of Malta, both based on the three letters of postal codes. It's not complete, and it doesn't count three of my 13 Maltese notes.

The third file (mt_yli.txt) is the most updated and the most complete, and generates the tables Local Councils / Kunsilli Lokali of Malta and Island of Malta, both based on the three letters of postal codes. The second table doesn't count three of my 13 Maltese notes. The first table counts all my notes, but the 3 notes inserted from ZBG1011 / ZBG 1011 and the 7 inserted from ZBG 1000 are listed separately.

To avoid duplicate tables and to have the most complete data, I think you can delete the first two files and preserve the mt_yli.txt one.

Malta.jpg (103.38 KiB) Viewed 1326 times

• Use a different separator for an internal data structure (fix to giulcenc's empty subgroup name).

Great works, it's working fine, now

In the last weeks, some italian users suggested me some little corrections and additions to italian file, I can send you (in the next days) via PM the updated version of the file or do you prefer corrections&additions only?

[lmviterbo]

The city profile information isn't included in the CSV file so there's no reliable way to know whether Calheta 9370 and Calheta 9400 are different places or simply different post codes within a large city. Calheta is actually a good example given that 9370 and 9400 aren't far away numbers.

So I guess this would be easily solved if the csv file included city profile information. I have already sent a request to the Development Working Group.

The tally is doable, the linking to notes is in fact easiest of your 3 items .

Great! (I will refrain from guessing which is easy or not )

[jrnsndrs]

The dots numbered 1-9 appear at the wrong places. […] I think you must renumber them to 01 - 09. At least such numbers work correctly with the biggest and smallest municipalities.

Some groups in Austria, Belgium, Germany, Spain and Italy have this problem. It was easier to change the sorting algorithm instead of going through all the configuration files.

They are now in the correct order.

I didn't know the same problem was at loads of other places. Of course, in that case it is indeed better to try to solve it in one time instead of doing the same thing over and over and over and (yawn) over again.

[jrnsndrs]

I can't see any things that are working incorrectly.
But I can't see the regions of countries where I have no notes at all.

[dserrano5]

To avoid duplicate tables and to have the most complete data, I think you can delete the first two files and preserve the mt_yli.txt one.

Thanks for this. NIG has duplicated tables then, since stock 1.84 has all 3 files.

In the last weeks, some italian users suggested me some little corrections and additions to italian file, I can send you (in the next days) via PM the updated version of the file or do you prefer corrections&additions only?

Corrections only, please. The new configuration files have different URLs to images (and some images remove due to them no longer existing in imageshack) and I would rather work on them instead of starting from a NIG file again.

May I suggest diff? (available for windows e.g. here along a bunch of other UNIX utilities). So you don't have to work on translating those changes into english, then I don't have to translate that english back into changes . Maybe there's some kind of visual diff program for windows, but that's a route I haven't explored at all . Basic usage:

Code: Select all

diff.exe -u oldfile.txt newfile.txt >diff.txt

Then you would send me diff.txt and I'd apply the changes automatically with the patch program.

But I can't see the regions of countries where I have no notes at all.

But that's ok, isn't it? What's the use of a ton of empty tables for Brazil, China, India, Iraq, Oman, Mongolia…?

--

I've just added a section "Hit regions" to the development EBTST instance. You'll see a potentially incomplete output until you re-upload your hits CSV. Let's hope the numbers make some sense . Translations appreciated for "Hit partners" and "Hit notes", with some emphasis on this differentiation.

[lmviterbo]

After many hours of checking and comparison of several docs and sources, I have found a dozen errors and some 30 omissions in either Cidades or Vilas. I'm writing the errors here (and only those for the moment) but I can do it elsewhere if you prefer.

Cidades
9001;Horta=Horta should be 9901;Horta=Horta
8400;Lagoa=Lagoa should be 8400;Lagoa=Lagoa (Faro)
9560;Lagoa=Lagoa (Açores) should be added

Vilas
2405;Macieira=Macieira should be 2405;Maceira=Maceira
3660;Santa Cruz da Tapa=Santa Cruz da Tapa should be 3660;Santa Cruz da Trapa=Santa Cruz da Trapa
4500;Silvade=Silvade should be 4500;Silvalde=Silvalde
5445;Carrazeda de Montenegro=Carrazeda de Montenegro should be 5445;Carrazedo de Montenegro=Carrazedo de Montenegro
6420;Vila Franca das Neves=Vila Franca das Neves should be 6420;Vila Franca das Naves=Vila Franca das Naves
7577;Cercal do Alentejo=Cercal do Alentejo should be 7555;Cercal do Alentejo=Cercal do Alentejo
9560;Lagoa=Lagoa shouldn't be here, because this Lagoa is now a cidade

After the corrections, we should have a total of 159 cidades and 581 vilas.

[jrnsndrs]

I've just added a section "Hit regions" to the development EBTST instance. You'll see a potentially incomplete output until you re-upload your hits CSV. Let's hope the numbers make some sense . Translations appreciated for "Hit partners" and "Hit notes", with some emphasis on this differentiation.

You seem to disagree with NIG about how hit partners must be counted. For this section I agree with NIG: The number of notes found in (say) Groningen (which is 7 for me) I think this is more interesting than the number of people who did that. (which is only 2 for me)

Hitprovincies_EBTST.PNG (11.29 KiB) Viewed 1232 times

Hitprovincies_NIG.PNG (12.49 KiB) Viewed 1232 times

---

Hit regions=Hitregio's
Hit partners=Hitpartners
Hit notes=Hitbiljetten

But I can't see the regions of countries where I have no notes at all.

But that's ok, isn't it? What's the use of a ton of empty tables for Brazil, China, India, Iraq, Oman, Mongolia…?

That's indeed quite obvious. I wanted to say that I couldn't help you with those. My only excuse for this poor message is that I had my attention somewhere else: There was Big News of Great National Importance on TV (announcment of the abdication of HM Queen Beatrix)

[tralla]

Hello,
another small glitch: You're obviously sourcing two configs for "Biggest Cities of Germany". Be sure to use the one I sent you; the other is *very* much outdated.

On the "Look and Feel" side, I'd love a header for each country (my "Regions" page starts with the header "Gemeindebezirke of Vienna"; I'd very much prefer "Austria").

And (again), a TOC would make navigation much easier.

Cheers, tralla

[Eureka.72]

There is a bug in my "general information" stats...

[dserrano5]

After the corrections, we should have a total of 159 cidades and 581 vilas.

Strange, I get 160 cidades (changes still not published).

You seem to disagree with NIG about how hit partners must be counted. For this section I agree with NIG: The number of notes found in (say) Groningen (which is 7 for me) I think this is more interesting than the number of people who did that. (which is only 2 for me)

In fact my 14 shitty hits preclude me from having a good idea of what NIG's numbers are supposed to mean so I'm not surprised about getting it wrong. I'll take a look to your hits and see if I can figure something out.

On the "Look and Feel" side, I'd love a header for each country (my "Regions" page starts with the header "Gemeindebezirke of Vienna"; I'd very much prefer "Austria").

And (again), a TOC would make navigation much easier.

Thanks for the "biggest cities" remark .

Both things are coming soon (read: already done in my laptop). Won't be exactly pretty but I'm feeling lazy regarding the HTML. I'm not a designer after all.

There is a bug in my "general information" stats...

Sorry but I'm unable to see the bug . What's wrong about "president" there?