[linkup]

Hello,
Long time movie jukebox guy on a search for a new media manager to use after switching to Kodi.

My movies have always been in their own folders with the poster, both named the movie name. I found a script to rename the posters from moviename.jpg to moviename-poster.jpg.

It seems TMM doesn't recognize that there is existing posters. The first folder I gave TMM to scrape contains 183 movies. TMM did a great job of recognizing the movies. Only about 5 needed me to identify the movie with it's IMDb ID.

However, of all of the movies, there are only check marks for 22 posters and although I didn't have a lot of fanart, none of them were found.

I understand the program is set up to scrape everything, but why didn't it find more posters and fan art?

Is there an option I am missing, or is it possible to enter a small piece of code that would do a search for an existing poster and use it. Even better, an option on the scrape screen with a couple choices. 1. scrape first, use mine second 2. using existing first, then scrape for missing.

When I view the movies, it seems TMM found them to put the thumbnail in the corner, but the movie is still marked as not having found the poster. I can't tell where there is actually a missing poster.

Thanks,
Lew

[mlaggner]

tmm detects a wide range of artwork names; <moviename>-poster.ext will be detected (<moviename>.ext would have worked too ).
If tmm detects any of these artwork files, it will show that inside the program (did you see the media files tab on the lower right - there you see all files which tmm detected for your movie).

the indicator in the list is not only for poster - it is for artwork in common (which consists of poster AND fanart). If any of these is missing (fanart or poster), there won't be a cross.

[linkup]

Hello sir,

1. Are you saying if a moviename.jpg or moviename-poster.jpg are in the folder, it won't scrape and overwrite?

2. Does the -poster version have precedence over the regular jpg?

I frankly don't care about fanart, if I have it, great, but don't go out of my way to find it. I won't make a movie decision based on it, lets put it that way. But, for cleaning up my collection, I do care if poster is available. A previous program I used with a similar column style distinguished posters from fanart.

3. if I put a poster or fanart file in the folder, how do I get TMM to update them? how and when does the TMM info get updated to reflect changes?

4. Why do many of the titles have a bogus 12/31/69 release date, even when it knows the release year and it is a different year?

Sorry for all the newbie questions, but so far, your program is my choice to use and support!

Lew

[linkup]

Found something else I can't seem to deal with. I have an improperly scraped movie. I manually edited the details page to correct the information. However the movie is not in IMDb and it has the ID of the wrong video. I edit the field, I delete the ID, I click "OK", and the value is right back again?

[mlaggner]

1. well, if you tell tmm to scrape and get artwork, it will overwrite existing ones. if you uncheck the option "Artwork" in the scrape dialog, tmm will get all meta data except artwork (thus not overwriting it)
2. no; -poster is just another valid name pattern for a poster
2a. as said before, tmm uses poster and fanart as an indicator for artwork (others like logos, banner, ... are optional). At this time we won't split up artwork into separate columns for poster and fanart. You can easily download fanart for all movies which have a valid tmdb/imdb id
3. if you modify the file structure for your movies (e.g. copying/renaming artwork), just do an "update data sources" to let tmm rescan the file structure
4. the release date is read out of the NFO file. If there is no parseable release date, an initial (technical) date is shown
5. set the cursor in the line if the imdb id and press the remove button - works for me

[linkup]

1. Where is the option to scrape and get artwork and ADD it to what is there? I worked hard to find most of my posters as I have them for titles IMDb doesn't.
2. guess I didn't ask the question right. If in a folder is a file called poster.jpg, and one called moviename-poster, and if one doesn't have precedence over the other, then it will grab one one time, and another another time? No way of knowing and nothing in the code to give priority, so what does it know to do?
2a. What is the easy way to get fanart if the scrape process the first time didn't work?

After doing a test folder, I submitted about 2900 movies to TMM. Out of the 2900, a full 2/3 didn't find fanart. ALL of them have IMDB numbers...well, maybe 10 out of the 2900 don't, but the other 2590 movies do. Of those 2890 movies, in almost all cases it matched the movie to the right IMDb number, but did NOT find fanart. I wouldn't care about the combined poster/fanart feature if it was better at finding artwork.

That is why I did a rescrape. Whatever I did found 1/3 more fanarts as only 1/3 are missing them. That is still 1000 videos to go through to find out what is missing.

4. What do you mean "initial technical date" Is that when they invented film? I would rather have an empty field than one filled with bogus information that would be misleading.

5. ahh....hitting the negative button.... Seems editing the value to "blank" and saving would do the same thing.

6. Of those 2900 movies, 7 show no year and a couple hundred have 1901 and others have early dates I know are bogus. However ALL of them have the "i" button checked. I assume the "i" button meant a successful match with a movie identity, Seems the "info" button should not be checked, and like with release date, I would rather have no year listed if it doesn't know what the year is as it didn't match the movie.

I don't know what else to do to get a better scrape. When all of the movie names and folder names match the IMDb titles, what more can I do other than manually matching up all of them? I read in another program that I could add the IMDb ID to the NFO, but that would take hours and hours if not days.

So far I am liking TMM better than the other options. I wish I could get paid back for my tedious work of carefully choosing folder and file names by being rewarded by a program that could be counted on to match ABC with ABC. Other than some movies that there are more than one version of, the others I would expect 99.9% matches on, and for those with more than one year, the choice to pick the right one.

Thanks for everything!

Lew

[mlaggner]

we need to clarify some things (I think we speak a different language )

a) tmm tries to detect as much as possible from importing your exisiting data. This is the file/folder name for detecting the movie name (and year). The existing artwork and any existing meta data from NFO files. All found data is stored in the internal tmm database and used for further operations.

b) when you scrape (getting meta data and artwork from the internet) you may do a search to find the right movie on the net (if you did not get any database id imported from filename/NFO) OR directly get any meta data using the detected id. You can only get meta data from the internet when there is any valid id from the meta data provider (like imdb or tmdb). Same is for artwork.

c) our results is just as good as the detected values (from file names and/or NFO import) and the results from the data providers in the internet

back to your points..

1.) there is no explicit option for that. If you don't want any poster get overwritten, I'd suppose you to (temporary) deactivate poster scraping in the movie options (just uncheck all poster file name options). If you scrape artwork right now, it will only download anything else than poster

2.) if you have two different poster with two different file names, tmm will find them both upon "update data sources" in the order the file system listing reports them (mostly in alphabetical order). tmm will take the first found for all operations (attention - renaming would drop the other ones, since there is/should only one poster per movie!)

2a.) right click on the movies, and choose "get meta data of selected movies" and uncheck anything but artwork (remember the hint from 1.) )
also be sure you've activated tmdb and fanart.tv as artwork providers in the settings - these are the best sources for artwork (there should be artwork for all popular movies, really old movies and/or local movies won't have any fanart in any source...)

4.) java (and most other IT systems) count time from 01.01.1970 in milliseconds. If no date is found, the object is created with the date 01.01.1970. This shouldn't be any problem, since it is only visible in the edit movie dialog??

6.) the i column indicates if there is meta data present for the movie (i.e. has a valid NFO file)

well, having bad scrape results can have different sources. But if you have valid imdb ids detected by tmm the only source for bad results would be the scraper itself. But without any examples I cannot do more - for my library at home there are only 2 movies which did not get detected by the scrapers..

[linkup]

Hmmmmm.... I wrote a reply and it is gone Wonder if I didn't save it after writing it...

I fed TMM about 2900 videos. quite a few had no year present so I know they didn't match and another couple hundred that showed a year of 1901 it didn't match. I fixed all of those and now I am wading through the rest. Have worked my way up to "C" so far.

1. regarding "i" for info. A blank NFO does not equal info. Info implies there is some info. If your "i" doesn't indicate a successful match, and it only means the movie has been added to the list, then why checkmark the box, or even have the box if it doesn't give meaningful data?

2. I would still like posters to have their own column but not a huge deal as I have learned to ignore that column as it has no value to me

I have found a bug....

If TMM scrapes the wrong movie and stores the poster, artwork, and genre, and then you scrape the correct movie, but the correct movie doesn't have fanart, it leaves the wrong fanart and genre buttons there. If it has been determined the movie is wrong, and another one is chosen, it ought to remove the incorrect artwork and everything else it "matched" from the incorrect scrape.

Second thing maybe isn't a bug, but affects how the program works. In previous programs I have had, you could put things either in the folder name or movie name that were informational. Second title, year, etc. I found that some of my misscrapes were in those situations. Like with other punctuation, the brackets or parens are stripped out. So instead of doing the scrape on ABC or [DEF], it is scraping ABCDEF so it doesn't match. Either name individually scraped fine. Seems to work properly, bracketed or parenthesed values should be ignored for that step.

Third thing I noticed....is SD the lowest quality logo you have to display? I found a movie I had two of...one was 320 wide (CGA), one was 640 wide, both labeled SD. 640/480i is the lowest resolution that should be labeled SD. I think rather than having a label for lower res dimension, CGA, EGA, and HVGA you might have just one symbol to indicate less than SD quality/resolution. Whether you call it Sub-SD, Low Res, or CRAP, it doesn't matter, but there should be a way to symbolize sub-SD resolutions.

One thing I like I should mention. Since I have 3000 of these to go through, one-by-one, it is nice being able to shrink the browser window down so at the bottom of the window I can see the file path, which tells me the movie rating (how I file) and then the movie name, i.e the folder. Then above that I can see the name TMM thinks it is so I can make a quick comparison between title and what I think the movie is.

I have been testing and comparing four programs plus I revisited a couple older programs I had used previously and so far I think TMM is the winner. One of the other programs I had been using and I thought was abandoned is showing some life, and that is Ember. Another top contender was Media Companion but due some restrictions from Google, he had to re-write some code. He wrote me this morning to say a new version is working better. He made me two custom versions to solve problems before discovering the Google issue which he says is now fixed.

In my ideal world, I would get all three of you in a room to combine the best of the three products as each has feature I like, and they all have problems.

Thanks!

[linkup]

Just remembered something else which didn't seem to work right, and that is runtime. Runtime doesn't seem to be calculated from the actual run time, so it is getting the value elsewhere. Sometimes when comparing movies of the same name I need to evaluate language, resolution, and runtime. Some films are censored, so there are shorter versions, or longer extended versions, etc. When there are two of the same movie, they will be next to each other in the list, then I toggle over to "Media Information" and toggle between them to compare quality, language, and length.

I remember the other day have two normal movie length videos and it said one was 31 and one was 34 minutes. I went out to Explorer to further compare. I can't remember if I was comparing runtime on the main screen or media screen.

Don't know if it is round up or? but the video on my screen right now says 57min on detail page, 56 min on media page. The video was 56:50. Different calculation for the two pages?

Thanks again!

[myron]

Detecting a clear moviename out of the filename(folder is always hard.
We try our best, to remove known words, and reassemble a useful title.
But with a second titles appending it for sure will parse to much.
(we do not strip out everything in brackets, we have a finer grained logic)
But year will be skipped if it is the last good token in filename (there should be no 1901 though)
If you could share some of these names, we even could improve this
(Personally i really would see the complete logfiles of your movie names)

"i"
should be only ticked, if we have at least some basic information like title, year, plot, genres, ...
Have to check, might be a bug...

SD logo.
Well, yes, guess there is noone yet.
We display 480/540/576/720/1080/ and up (quite the same ways as Kodi)
SD is all below 480

Shrinking the windows should be possible;
There are always dragable sliders between the list, the visual fanart, and the tabs

We have 2 runtimes.
One is calculated from every file (vie mediafile) and summed up,
the other one comes from the scraper.
Depending on your settings (prefer mediainformation or not), we should use one or the other....

[linkup]

Found a good example of the glitch with runtime....I had three copies of "The Disappearance of Alice Creed", All in the Rated R folder, one with and one without the "The" in front. Two with the "the" in front, one had the year in parens at the end, 3 file folders, same movie.

The folder and movie without the "the" has the right IMDb ID, and no TMDB id. It says runtime of 136 on the details page, nothing on the media info page.

The folder with the leading "the" and without the year has the same correct IMDb ID, but it also has a TMDB id. This one shows 93 min runtime on the details page and 97 min on the media info page

The third one with "the" in front and (2009) at the end has an IMDb ID, does NOT have a TMDb ID. It shows the movie runtime as 100 minutes on the details page and 97 minutes on the media page.

It turned out the movie that didn't show a runtime on the details page was corrupt, so it couldn't be determined. The key here is that for one movie, all with the correct IMDb ID all show different runtime lengths in the details, 136, 93, and 100, none of them being the correct 97 minutes my version is. I noticed IMDb says it is 93 minutes, so that is where the middle of the three got it's info from, but for whatever reason, the other two got their info from elsewhere and none of the three was for my version of the movie.

I really do appreciate all the coding and testing that has gone into TMM. Thanks!

[linkup]

Did it again....wrote the message above this but didn't save it so you had written a reply while I was still writing things about my previous message, so things are out of sync.

Regarding find a name....frankly, this is where I think these programs all s**k. I have a movie called "The Babysitter". Simple...Do you know what it scraped?

"Don't Tell Mom the Babysitter's Dead"

First of all, we have the KISS rule. If the folder, the movie, and everything says babysitter, why would it check for something else and accept it?

The matching errors fall into a couple categories and I wish on the get-go I would have had some choices.

1. Only accepting exact name matches (other than The, a, l', etc.(others who are more "sloppy" with their naming should be fine as is, and they can deal with the issues as of a result)
2.. If there are two or more exact name matches, prompt me to figure it out.
3. Ignore values in brackets and parenthesis. Like "comments" in programs you write for yourself. They were a part of earlier naming standards.

Think I better save this now as I am distracted and unable to be coherent for long at a time...am I getting ADD?

[linkup]

Respond now to your response

The 1901s are substituted when there is no match. Here again, I prefer blank (which is what 10% of the no-matches showed) vs. created data, i.e. the release date.

Sorry, didn't ask for shrinking windows. I was complimenting you that I could make the window smaller and get the file title and movie title fairly near each other.

In that some of my obscure content is old, a fair amount is lower than SD standards so a way to indicate it would be nice.

Are you saying runtime adds up all video content, i.e. the trailer is included in the runtime? I assume you mean the one on the media info page? In my perfect world, the scraper would identify the main video, i.e. the longest and display that runtime value. Even better, rather than going to the other tab, there could be two values next to each other. Expected RT based on what it scraped, next to it actual. Even when I fully minimize the view, there is a lot of space for cool new things, hint, hint...

Let me guess, at one time was there one less row in the data, on the screen? Meaning, when I minimize the screen to it's smallest, it cuts off the bottom row, the movie file. As if the minimum height was set prior to the current data height?

Going back to my three versions of the one movie. You mentioned the info displayed on the Details screen is either what it summed up or what it scraped? Is that why they are inconsistent? I only have IMDb marked as a movie scraper. It shouldn't be grabbing info from different places for the identical movie in 3 different folders.

As mentioned, there are two values we should care about, if only for curiosity sake. The runtime of the movie and what the scraper (IMDb) thinks it should be. Would the average person care, when looking for a movie to watch, what else is in the folder, i.e. bonus material, etc. I only want to be up for one more hour, I want to find a one hour movie

and one more anomaly I found.... A movie with the fanart checkmark that didn't have a poster or fanart. blank, nada, zilch, zero,

Just finished figuring out my quarterly sales tax...what a headache... thanks for putting up with this old curmudgeon

[myron]

This thread is getting too big on questions.
Feel free to add feature requests / bugs to our issue tracker on github (one for each), else we cannot handle nor discuss this accordingly.
And please provide LOGS and/or screenshots/examples - discussion is nice, but hard to read/understand...

Regarding matching:
IIRC, we always take the foldername, get a "clean" name out of it, and save it as movie title.
If this movie is in an directory with other movies - we take the filename instead (also cleaned)
Now we have the title
This is sent (quite 1:1) to the scraper search
If one scraper can't find the correct "the babysitter", another would! (not our problem)
You can play around with the title match confidence slider in settings - if set to 100%, autoscrape will only take the perfect matches.


How do you get the 1901?
Cannot reproduce this...

[myron]

RUNTIME:
we have 3 pages displaying the runtime.
- on details tab we display the runtime according to your settings (either the value entered on movie edit == scraped runtime OR the runtime of all type=video files)
- on media information tab (runtime of all type=video files)
- on media files tab individually for every file

When you change your preferences, the details tab does change runtime, the others not.
And NO, we're not taking trailers & other files into account - just the main video (type=video)

So for my POV, this works as expected... (beside some minor rounding differences, since we only display minutes)