|
|
Welcome to the Database Correction page. This page is for letting me or the other editors know of corrections that need to be made. Please read the posting instructions carefully.
|
|
| Thursday October 31 01:30:37 2019 Re: Crazy good resource on IMDb |
| Yep, another followup: I got extra-nerdy and extracted the information from James's lists into a spreadsheet format that will be more useful for cross-referencing. https://mega.nz/#!wW4lySZJ!X20vIn_qoBj2FiIG1fVRbQxtX_lhECuHVWialmqiaFM In case it isn't immediately obvious, the fields are as follows: * screencap URL * caption (usually includes actress name(s) and gag type) * IMDb actress ID(s) * actress name(s) * IMDb movie or series ID * movie or series title * IMDb episode ID * episode title A note about the actress name(s): this field will appear in parentheses if it is "suspect," which occurs if (a) none of the linked actress names appears in James's caption, or (b) there IS no caption and there's more than one linked actress name. These scenes should be checked by hand for extraneous/incorrect names or even putzes in the actress list. Fortunately, the OP is fairly meticulous, so there are only 29 such cases out of 2767 scenes (and only 4 actual mismatches between the caption and the name links). My next step will be to extract enough additional info from the title links -- release years for movies, season/episode numbers for TV -- to support reliable automated matching between this list, my clip archive, and Brian's DB. It might be a while before I get to that, so I figured this might be of interest in the meantime. |
| Raffish |
| Thursday October 31 09:20:10 2019 Re: Crazy good resource on IMDb |
| On October 31 2019 Raffish wrote: > Yep, another followup: I got extra-nerdy and extracted > the information from James's lists into a spreadsheet > format that will be more useful for cross-referencing. > https://mega.nz/#!wW4lySZJ!X20vIn_qoBj2FiIG1fVRbQxtX_lhECu > HVWialmqiaFM > In case it isn't immediately obvious, the fields are as > follows: > * screencap URL > * caption (usually includes actress name(s) and gag type) > * IMDb actress ID(s) > * actress name(s) > * IMDb movie or series ID > * movie or series title > * IMDb episode ID > * episode title > A note about the actress name(s): this field will appear > in parentheses if it is "suspect," which occurs > if (a) none of the linked actress names appears in > James's caption, or (b) there IS no caption and there's > more than one linked actress name. These scenes should be > checked by hand for extraneous/incorrect names or even > putzes in the actress list. Fortunately, the OP is fairly > meticulous, so there are only 29 such cases out of 2767 > scenes (and only 4 actual mismatches between the caption > and the name links). > My next step will be to extract enough additional info > from the title links -- release years for movies, > season/episode numbers for TV -- to support reliable > automated matching between this list, my clip archive, > and Brian's DB. It might be a while before I get to that, > so I figured this might be of interest in the meantime. I'm having a little trouble trying to understand what you're trying to accomplish here. I took a look at the file you extracted, and I can see it as a base for building entries in a new database, but I don't see where you can apply it to the current database. Do you want to double-check the existing entries for the overlap information (i.e., episode titles), or to add new entries for the pics that are missing? If you want to use it to add new entries for the pics that don't currently have a matching entry, we're going to have an issue that goes back to the original creation of the DB way back when. The original DB had dozens (hundreds?) of entries which had the title and actress (maybe), and just "bound and gagged" as a description. This version of the DB was useful, but it was also very frustrating. Creating more entries from the stub information in IMDB is bound to bring back the same problem. Whether its worth putting up with the skeleton entries or not to get the basic information I'll leave as open to discussion. |
| Gagster |
| Thursday October 31 19:13:09 2019 Re: Crazy good resource on IMDb |
| > I'm having a little trouble trying to understand what > you're trying to accomplish here. I took a look at the > file you extracted, and I can see it as a base for > building entries in a new database, but I don't see where > you can apply it to the current database. > Do you want to double-check the existing entries for the > overlap information (i.e., episode titles), or to add new > entries for the pics that are missing? Well, that depends on how many of the scenes don't have entries yet, and how many have entries but are missing actress names or episode titles. I don't really have a sense of that yet, but I have to figure that, given the number of scenes here, there are probably a decent number that fall into one of those categories. Of course, the same could be said of my clip archive, which has over six times that many scenes. But these do have the advantage of all being scenes with attractive gags, and in most cases the caption even lists the gag type (for turning into a tag). As for whether it's actually worth creating stub entries with only this minimal information, I remember bringing that up with Brian a few years ago (I think with respect to one of Guest123's massive scene lists), and I seem to recall that he was open to the idea. But yeah...I'm not actually expecting anyone to do anything to the DB manually based on this list. At some point, I'll throw together some code which will check it against what we've already got, and then we can decide if anything more is worth doing. |
| Raffish |
| Thursday October 31 19:29:36 2019 Re: Crazy good resource on IMDb |
| I do think that stub entries are worthwhile. Now that it's easier than ever before to find movies and TV episode online, the biggest obstacle to building more useful entries (and to editors making more clips for the community) is simply that people don't know that the scenes exist in the first place. Stub entries could go a long way toward fixing that. |
| Raffish |
|
|