English > Support

IMDB Rating not filling-In automatically

<< < (7/11) > >>

buah:

--- Quote from: rick.ca on January 12, 2010, 01:55:47 am ---They can no longer be selected in the plugin configuration, and previous settings are lost. Do you see the same thing?
--- End quote ---

If the question is for me, I'm not sure what "previous settings" are, cause I've tested it only since last two days on a single entry - "Shawshank", trying to retrieve any following data step-by-step, before I hit "Silent-import~5500-movie-entries-imported-from-MS Excel", and until now that was the only issue I couldn't resolve by myself.

Actually, there is another issue I noticed, but I assume it's not a proper topic for it (it's about media type not retrieving info when importing data from a network share, or from a mapped network drive). I'll post it later. It's not of such an importance at the moment, I guess.


Cheers

rick.ca:

--- Quote from: nostra on January 12, 2010, 02:07:58 am ---Can't confirm that. Could you check at what point exactly does it go wrong on your system?
--- End quote ---

Sorry. I must have had the wrong plugin selected when I checked the configuration. :-[

But I can confirm what buah has observed. To be a little more specific, the plugin is saving NULL to the selected rank field—any existing value is deleted.

buah:


--- Quote ---You should assign a Memo field for Tech Specs to make it display properly. For Votes and Rank the Number type is optimal as you gain a possibility to sort and filter by these fields in a wright way.
--- End quote ---

Thank you.

To rick: Now I know what you meant. Unfortunately (or it's not?:)) I never had any value other than NULL, since I started to use it from scratch :)

mgpw4me@yahoo.com:

--- Quote from: rick.ca on January 11, 2010, 11:26:31 pm ---I'm surprised the "Please God, Now" didn't work, but it really does seem to be broken. Maybe by the same website change that broke the rating, but wasn't noticed. I wonder if we can devise a test to determine whether the plugin is fully functional at any point in time.

--- End quote ---

Quick sanity check for rating:

if ( Rating = FloatToStr(StrToFloat(Rating)) then
 we have a usable rating
else
 the script is broken or you need to check if there is a reason there is no rating...not enough user ratings submitted for example

Test:
  Rating = '', conversion results in '0.0' ... not equal, bad rating
  Rating = 'some text value', conversion results in '0.0'...not equal, bad rating
  Rating = '1.0', conversion results in '1.0', equal, good rating

With a bad rating value, the script should not even attempt to add one to the database.  In silent mode there isn't much that can be done, but a showmessage('rating not acquired...' + reason ) could be displayed in 'user' mode. 

Any time a valid value is 'missed', it should be considered that the parsing may need to be done at a lower level.  For example:
<td>rating</td>
changes to
<td class='rating'>rating</td>
If you were using <td> as the start of text indicator, then you'd consider using <td, then parse until you found the matching '>', at which point you'd have the start of text position, you could look for the text termination character sequence.

Writing this type of code for every value is possible, but very time consuming, and since code (to my knowledge) is not sharable between program (ie. #include common_routines.psf) you'd have to change every program that used it when changes were made.  When you consider the shear volume of code involved in this project, it's easy to understand why this hasn't been done.

rick.ca:
I was thinking along the lines of a brute-force user test. Like downloading a set of reasonably static records that covers all data types, exporting everything to text, and then comparing that to the previous test to identify any new errors or omissions.

Navigation

[0] Message Index

[#] Next page

[*] Previous page