Personal Video Database

I'll try to clarify as much as possible, considernig language barriers.

Thanks. So it really has nothing to do with individual fields, but the window itself. Based on your description, what seems to be required is a scroll control that can be applied to a section of the skin. But what would be more straightforward and useful to more users is a splitter for the Panel. That would result in a scroll bar for each of the top and bottom splits, allowing any part of the panel to be placed in either split. The splitter could be dragged to any position to set the size of the splits, and simply dragged to the top or bottom to return to one view.

But why stop there? A split List would be very useful too. The currently selected movie could be kept in view in one split, while the other is used to scroll to a different movie. Having two movies in view in the list would provide an easy way to switch between the two. And maybe an Alt-Click of a movie in a split list would instead open that movie in the corresponding panel split—so the two movies could be compared directly.

Nope, not finished. This reminds me of something suggested long ago. That's another panel to display the person record last selected in the movie panel. In other words, clicking on a person link in a movie record would show their record in a panel below the movie panel, rather than switching to People View. Imagine that being the bottom split of the movie panel rather than a separate panel. That would allow the space used by each to be controlled with the splitter in exactly the same manner. The List could change between movies and people depending on which panel is selected. Or if split, would show movies in the top and people in the bottom split.

I'll stop now, before this gets moved from the TODO list to the BUCKET list.

So, my OP application still stands. Any ideas?

Maybe you could clarify what it is you are asking for or suggesting. The design reflects the decision to address the need with collapsible fields rather than scroll-able ones. The answer to 'How?' is therefore it's not possible, by design. If the suggestion is scroll bars would be better than collapsible fields, I think most users would disagree. It's fair to ask if scroll-able fields might be offered as an alternative, but I wouldn't hold my breath. There are surely worthy items on our collective wish list which have no existing alternative. Even if it were easy to do (I have no idea), nostra would have to consider whether providing two mechanisms to achieve one purpose is worthwhile, or just adds to the overall complexity of the program (i.e., consider how such a choice would be explained to a new user).

Does anyone know what the problem may be?

I'm getting 404 errors and no results as well. Something probably changed at IMDb that will require the plugins be updated.

Can I automatically use the filename (for ex "Terminator.avi") to fill the "Original Title" field (Terminator) ?

I think so—by using <origtitle> instead of <title> in the relevant regex. I believe the same value will be used for [Title] if it's not otherwise provided. But this will only work for files being added to the database via the scanner. We really need the plugin(s) to be working properly in retrieving and updating [Original Title].

Pity, nobody can help me.

Welcome, and sorry you haven't yet found help. Your question is very specific, and that normally increases the chances of getting a good answer. Unfortunately, it also makes is easier for the reader to think, "I don't know," and therefore not respond.

In this case, I can tell you there probably are ways to import XML data, but not directly. I've never found a way to easily convert XML files into a plain text, CSV or XML files that can be imported. If there were a tool that would batch convert the NFOs to the particular TXT format PVD will import (when found "beside" he video file), that, of course, would be a solution.

But it might be better to question your objective. Your original question implies it's important the XBMC produced NFO's be the source of the data. But PVD's primary purpose is provide the tools necessary to gather and maintain the data for a video database. As such, it would normally be the source of data for XBMC, and it seems that might what you actually want to do. If the data in the existing NFO files is the same as what PVD will get anyway (e.g., basic data from IMDb), then there's no need to import that. If there's some personal data (e.g., date viewed, rating), there's probably a more straightforward way to do a one-time export/import of that data using a list.

Could that be a solution?? Can PVD match with the IMDB ID the movie to 100%?

Yes. Unfortunately, PVD will not use the IMDb tt number to find the movie. (I'm not sure if even the IMDb web interface will.) But with an accurate [Title] and [Year] you should get the correct movie about 90% of the time. That's not bad, since it means a Silent Mode import will get 90% in one go. Each of the remaining 10% with then require a few seconds of your time to resolve whatever the ambiguity is. Even if there's a 100 of them, you'll be done faster than anyone can dream up a better solution.

You describe a file-naming scheme in which [Title] and [Year] would always be found in the folder name, consistently delimited by the "_" (or by the "tt000000" separating them). This should do the trick:

(?i)^.+\\(?P<title>[^\\]+)_tt\d{7}_\((?P<year>(19|20)\d{2})\)

The attached screen shot of Expresso's 'regex analyzer' window shows what this expression means/does.

[attachment deleted by admin]

Well Dr. nostra, it's like this...I'm not suggesting this is any quantum physics phenomenon, but it's clearly the sort of behaviour that vanishes quickly once such questions are asked. Before responding that it still happens frequently, I decided to test to see if I could elaborate (although I've never seen a pattern). I had no difficulty finding a few (I might have tried 10 records to find them), then decided to take before and after screen shots to illustrate. Naturally, it didn't happen again (I checked about 50 records).

I use only IMDb plugins to update people. They're both up-to-date.

I'll try to remember to update this if things change. Maybe if I can do the test and get the screen shots without looking I'll capture the elusive behaviour.

Welcome.

I  like to import my imdb-watchlist into database from PVD.

The first row of the CSV file should be a header indicating the IMDb data element included. Configure the CSV Import plugin, matching each of those to the PVD field you want to put it in. Note that it's the order in which those fields are placed that determines which fields the data is imported to. Also ensure the plugin options are appropriate to the circumstances. Then run the plugin, select the file and answer the prompts (if any).

The Excel Import plugin is essentially the same. You may find it easier to open the CSV in Excel, see the data in columns, make any necessary changes (e.g., deleting unwanted columns), and then import the file (saved as an XLS or CSV).

I just don't get it how to import series, seasons and episodes.

Perhaps you have so, but the first step is to download the series, including episodes. You may, but don't have to download the episode data at the same time. All you need are properly numbered episode records for each of the video files you have on hand.

Do they have to be in a special format (naming)?

Yes. Like any file added via the scanner, the files must be named in a consistent manner, and there must be one or more regex in the configuration that match that pattern. If your episodes are named Series Name S00E00.xxx, you should have no problem.

I'm sure the issue is not being ignored because it's deemed unimportant. It is important, and probably too complex to be dealt with easily. It's also the sort of thing where no one solution will be optimal for all situations and make everyone happy.

Having a more or less established collection, I don't find this to be a huge problem. Nevertheless, I do find it frustrating when the behaviour is confusing, or simply won't do what I need it to do. One thing I would dearly love it to do is add trailers to [File Path]. That requires the new file literally be added to the field, not changing the value already recorded or adding a new movie record. Should this be handled within the current design, or should there be other means of detecting and adding trailers? [Hint: There will be no agreement on how that should work.]

This issue can't be separated from that of duplicate recognition and handling, which has always been problematic. Many don't seem to appreciate that for every movie in their database, there might be 100 invisible ones included in filmographies, awards, etc. When a movie is added, it's necessary to determine (or at least try) to find the invisible record that very likely already exists.

If and when nostra has time to address these things, I would suggest the scan results dialog needs to be very clear about whatever it's able to do with a high degree of assurance. For things that are ambiguous, it's okay to allow the user to make changes (as it currently does), but it would be much more effective to provide some interactive tools. In the time I spend wondering what it will do when I make some change, I could select an existing movie from a list (or "add new movie") and tell it exactly what to do (i.e., change existing path or add new one). Because the window is modal, it would also be helpful if there were a "bookmark for follow-up" option for those situations that will remain ambiguous until the user can examine the results. It's often easier to guess and then correct the results.

I noticed that my personalized pvdconf.ini becomes corrupted...

I'm stumped, but it might be helpful to know exactly how the file is 'corrupted'.

And a last straw to grasp...You might verify that your AppData location hasn't changed (I can't imagine why it would) by checking HKEY_CURRENT_USER\Software\Microsoft\Windows\CurrentVersion\Explorer\User Shell Folders to see if AppData = %USERPROFILE%\AppData\Roaming.