Scan folders/file scanner confusion

[patch]

I do not think movie file tagging is widely used or standardised yet...

That seems to be the case. I just spent hours trying to get image tag information into MC—it seems hopeless. I recall abcAVI Tag Editor saves IMDb information in AVI's,

To get MC to use tag information it both needs to be written to the movie file, and MC needs the code to recognise it. I suspect you would have problems with both steps ATM.

Master databases worry me...

I was assuming this would be a one-time thing—just for getting started. I meant "master" only in the sense it would include all movies on the server, and users would start their own database using a copy of it.

I had assumed that it was proposed a server version on PVD would need to be run on the file server and client versions run on other computers or accounts. For this to work I need
1) Server implementation of VDB (Master database implementation)
2) Computers sharing movie files to be switched on
3) Server VDB to be running once somewhere
4) Client version of VDB to be run by users

My preferred set-up is to just reliably re-scan my movie collection from each computer / account. For this to work I need
1) Version of VDB with 100% accuracy on re-scanning files (after it has been done once manually & results exported)
2) Computers sharing files switched on (running operating system file sharing)
3) Stand alone version of VDB on client computer (& intermittently rescanning shared file directories)

Yes the files would need to be copied together if movie files were moved, but I already do that with subtitled movies.

I have no doubt you can do it—you're the one running an industrial network in your home! Other users don't where or how there files are stored. For them, separate files increases the risk of things going wrong.

You over estimate what I'm doing.
"File server" is a windows XP machine the "share folder" enabled the the directories containing my movies.

I find some movies have more than one file (trailer, subtitles). So for consistency I put each movie in it's own directory, so adding another file to that directory would be a trivial change. For those that have multiple movies in one folder they could either move the whole directory also or sort by file name and move the movie file & associated tag file. If not manual re-scan would be require. Either way I believe it is workable until tags within movie file are better supported.

In summary
What I'm really trying to achieve is 100% rescan accuracy without human intervention. Independent of path to the movie, user account I'm scanning from or computer I'm using. This ability can then be used for many applications including a multi-user network environment, movie storage updates, holiday set up etc.

The only way I can see this being achieved is if the manual selections made during the initial scan are stored with the movie file. The simplest implementation being a file containing the URL to the movie description (eg imdb) and cover image (eg Amazon).

The rest of the discussion is really just trying to predict what direction the industry is likely to go in the future by comparing how it is done with music files (a more mature technology).

[rick.ca]

I had assumed that it was proposed a server version on PVD would need to be run...

No, no, no! I know it's confusing, but this is a three-way conversation. My recommendations are never the same as nostra's. 

I understand why you would prefer not to run a server version of PVD. I was agreeing with you in recommending each user run PVD stand-alone, using their own separate database. I was merely pointing out it would not be as much work to create those separate databases as you seemed to think it would be.

The only way I can see this being achieved is if the manual selections made during the initial scan are stored with the movie file.

Again, I'm not questioning the means, I'm questioning the objective. Your scheme is only necessary in the event one needs to rebuild a database starting with nothing but the movie files. This is not necessary if you are just changing the location of files—the rescan function works with 100% accuracy (I tested it since my last post). It is not necessary if you are creating a new database that is a subset of an existing one (your vacation scenario).

I understand you want any user to be able to build a database by selecting files and letting PVD do the rest. I've already suggested your request may be easy to implement. Please don't think I'm arguing this because I don't think you should have the feature you want. Nostra is perfectly capable of weighing the pros and cons and the amount of effort required. I'm countering your implication there is something fundamentally wrong with PVD when it's actually performing extremely well—according to its intended purpose. If any further effort is required in this area, I have to wonder if it should not be directed to making the initial scan more accurate. Everyone would benefit from that.

[patch]

While I still think (in the short term) a separate file containing tag information is an efficient way of supporting all video formats and avoiding player compatibility issues.

It is interesting several programs support writing tags within movie files such as
IMDB Movie DB http://www.dancemammal.com/moviedb.htm
abcAVI Tag Editor http://www.exefind.com/abcavi-tag-editor-P21011.html

I note the movie file reader PVD uses supports reading some movie data, pity it didn't support writing as well
http://mediainfo.sourceforge.net/en

[seabass]

Master databases worry me...

First of all sorry for my bad english..
I don’t know if I my suggestion is to primitive, but wouldn’t it be the easiest way to solve all the problems if you would create a simple importfilter which can access a normal pvd databasefile then you just have to include somewhere in the preferences the option to enter a filepath to the central pvd file on the server and choose this plugin for standard import and you would always have 100% accuracy.

..or if a pvd file cant be accessed on a server because of the restrictions of firebird just put an exported csv file and a folder with the pictures there.

[rick.ca]

I'm not sure I understand your idea, but it gives me one... The ability to export to, and import from another PVD database might be a useful "utility" feature. It would provide a convenient means for creating a database that is a subset of another, and to share records with other databases. I'm assuming the export of all movie information, including all images and associated people. It probably should include the option to exclude "personal" information like date seen, rating, comments, etc. The advantage of this over a CSV/XLS export/import is the resulting movie records would not need to be updated.

Neither idea satisfies patch's need. If I understand correctly, he has multiple users running independent versions of PVD, but sharing a common movie collection. Your feature or mine could be used to establish a new database for a new user, but won't help on an ongoing basis. A problem remains with updating new movies (i.e., downloading information). Some movies cannot be identified by scripts without some intervention by the user. Once this has been done by one user, it would be helpful if the identifying information (e.g., IMDb record number, suitable poster URL, etc.) was automatically available to the other users—so their updates would be fully automatic.

Nice ideas, but I think the priority should be the improvement of the program's ability to correctly identify movies from file names. If were successful 99% of the time, there may be no point in adding more features to deal with the 1%.

[nostra]

There will be a possibility to append another PVD database and copy/paste records from one database to another in 0.9.9.

For the patсh's problem: it is already possible to write a script that would generate a txt file with any information needed for each movie placing it near video file. Then a script like this http://www.videodb.info/forum_en/index.php?topic=904.0 could be used to retrieve the information from txt files.

[rick.ca]

There will be a possibility to append another PVD database and copy/paste records from one database to another in 0.9.9.

Just to clarify... The append function would include all information, including custom fields not in the destination database? Will there be the option of excluding "personal" information like date seen, rating, comments, etc? If I wanted to share a subset of my database with another, I would delete unwanted records from a copy of my database, and then the copy would be appended to the other.

I presume copy/paste would only include the record copied, not any associated records—so the pasted records would need to be updated.

For the path's problem...

Let me outline this to see it I understand correctly...

• An export script would create a text file for each selected movie containing the necessary identifying information.
• An import script would import any information found a text file with a name identical to the movie file, for each selected movie.
• The import script would be added to the beginning, and the export script to the end, of of a batch file for importing movie information.
• The batch file would be selected for updating movie information in the Scan Results dialog.

So, if a new movie had been previously updated (by another user in a shared collection environment), the information necessary to ensure a fully automatic update would be found in the text files. If not previously updated, a text file would be created after the update was successfully completed.

It's not clear to me if this will work for posters as well. If a value (a URL, I presume) for a poster is read from the text file and saved to the database, will the existing poster plugins use this information to download the poster?

In the circumstances, this seems to be the most practical solution. All we need are some kind volunteers to write those scripts. 

[patch]

For the patсh's problem: it is already possible to write a script that would generate a txt file with any information needed for each movie placing it near video file. Then a script like this http://www.videodb.info/forum_en/index.php?topic=904.0 could be used to retrieve the information from txt files.

Excellent.
This should achieve the functionality I was looking for.
Thanks for implementing it.

[patch]

a script like this http://www.videodb.info/forum_en/index.php?topic=904.0 could be used to retrieve the information from txt files.

Might be some value in extending this to extract the imdb link out of .nfo if present and not previously defined in database or tag file (.txt)

[nostra]

I do not have time now for improving the scripts. You can do this if you wish...

[patch]

For the patсh's problem: it is already possible to write a script that would generate a txt file with any information needed for each movie placing it near video file. Then a script like this http://www.videodb.info/forum_en/index.php?topic=904.0 could be used to retrieve the information from txt files.

Has anyone had any luck with this?

I had a look at scripts and they all seam to relate to importing data into PVD, typically via a url reference, parsing html page, and filling relevant database fields.

Plugin templates seam more relevant to me. The closest I could find is the "Web Pages(Index&Movies)" template which produces one (html) file per movie.
I'm guessing what is need is to:

1) Set "recordfilename" to an absolute address based on "File Path" rather than an address relative to the index file. Not sure how to handle multiple entries in path, probably need to process them the same as PVD does when grouping files, or txtinfo_en.psf http://www.videodb.info/forum_en/index.php?topic=904.0 does in function GetDownloadURL : String; Had a go at the latter but couldn't get it working

2) Produce a pages/movie.ptm to export the text file contents, probably based on "XML (Movies).ptm". Looks like it wouldn't be too bad to adapt to what txtinfo_en.psf wants

3) Run the export function to write tags to all known movie files.

4) Produce a txtinfo_en+imdb+allmovie.batch to import movie data into new user accounts. Maybe straight forward but I don't really understand how the various importing options relate to Scripts and Plugins

[patch]

Plugin templates seam more relevant to me. The closest I could find is the "Web Pages(Index&Movies)" template which produces one (html) file per movie.
I'm guessing what is need is to:

1) Set "recordfilename" to an absolute address based on "File Path" rather than an address relative to the index file. Not sure how to handle multiple entries in path, probably need to process them the same as PVD does when grouping files, or txtinfo_en.psf http://www.videodb.info/forum_en/index.php?topic=904.0 does in function GetDownloadURL : String; Had a go at the latter but couldn't get it working

I now suspect this is not possible.
Looking at http://www.videodb.info/forum_en/index.php?topic=766.0 and experimenting with "Templates for plane text export" I suspect templates for export plugin do not support functions and recordfilename does not support absolute addressing (sub-files must go in the same directory as user selects to save the index).

Looks like I will have to hope nostra adds tagging in 0.9.9 or learn to write my own export script.

[nostra]

Looking at http://www.videodb.info/forum_en/index.php?topic=766.0 and experimenting with "Templates for plane text export" I suspect templates for export plugin do not support functions and recordfilename does not support absolute addressing (sub-files must go in the same directory as user selects to save the index).

Yes, it is not possible with a standard export plugin, but can be solved pretty easily with a script. I have no problem with writing a script for you, but only after 0.9.9...

[patch]

I have no problem with writing a script for you

That is great news as with txtinfo_en.psf that will implement movie tagging functionality

but only after 0.9.9...

Sounds like sensible prioritisation to me. I'm looking forward to seeing 0.9.9 also

[patch]

Sounds like sensible prioritisation to me. I'm looking forward to seeing 0.9.9 also

This is frustrating.
Just finished backing up my movies, updated PVD data then copied it to the backup drive.
Or that was what was suppose to happen. Unfortunately I rushed the last step and actually copied the old copy over the new one.

Now I'm back to on old copy of my PVD data which cannot be opened by a non-administrator. So will need to export to CVS, manually import pictures, & redo last 2 months cataloguing.

I'm really really looking forward to being able to export tags. Doing this stuff twice because of my own mistakes is very frustrating.

[rick.ca]

Ouch! Sorry to hear that.

You don't want to hear this, but for my own sake I need to repeat my mantra: "All my backup routines shall be automated, multi-versioned and otherwise idiot-proof. The biggest risk to my data is ME!"

[patch]

The biggest risk to my data is ME!"

Yep, I agree. Was just a bit annoyed with myself last night. Oh well, today is another day.

[patch]

How do you name your files? Could you post some examples that produce different results?

Usually I name my files as follows:
...\Title\Title.S02E01.episodename.avi

Some of the files are named:
\Title\Season\Title - 1x02 - Episodename - details.avi

Same however naming is not as consistent as it should be. All TV series files have season & episode indicated but it varies for example season 1 episode 2 may be specified by any of the following
s1e02
s01e02
01x02
1x02
[01x02]
[1x02]
102
In most cases this numbering is proceeded by the series name however the punctuation & captilisation varies. In some the series name is omitted altogether. In some a "series description" follows the numbering, in others just encoding information.

What is consistent is I put all files pertaining to a series in the same directory.
...\Title\s1\
...\Title\s01\
...\Title\Season 1\
Less commonly
...\Title\

IMO
For PVD to have any real chance of picking up all these files it would have to assume once one episode of a TV show has been found in a directory, then the other files in the same directory tree also pertain to the same TV show. This would enable much less stringent matching criterion without confusion from movies (which may spans multiple CD or DVD). Not unlike how vob & similar files from a dvd rip are handled.

This would make adding episode to and existing TV series in PVD much more reliable.
How to respond when PVD first encounters a news TV series is more difficult, as after one is confirmed then other files become additions to an existing series. Might be easiest it users just add one episode first then re-run the file scanner to add all other episodes (unless someone can think of a more elegant solution).

[rick.ca]

For PVD to have any real chance of picking up all these files it would have to assume once one episode of a TV show has been found in a directory, then the other files in the same directory tree also pertain to the same TV show.

I wouldn't be very happy with that. Most of my episodes go into one directory until viewed, then deleted. If it's too difficult to code a scanner that can understand a variety of different file naming conventions, I'd rather one be imposed. Considering any of the possible sequences of characters that denote season and episode are very unlikely to mean anything else, would it not suffice that the title appear first, followed by the season/episode, and then anything after that is ignored?

[nostra]

I think I can manage to make the scanner understand multiple naming formats, not 100% precision but still pretty precise.
Just stay tuned for the next beta