Additional info in roles

[GeorgeFive]

I noticed this in the changelog:

Fix cast() parsing of role. The role field no longer contains anything other than the name of the role played

So, am I reading it right in that this is intentional, or is this an unfixed bug? I really think that this is pretty important information...

[IzzySoft]

See the full details in #10 George:

• Yes, this was intentional
• Yes, there is "pretty important information"
• No, that got not lost. You just need to investigate the array again, there are additional fields now :)

Also see bcc1e22 (especially the commit message: "Add extra fields to cast() results. Tidy up role parsing. Fixes #10 ")

[GeorgeFive]

Ok, I found the name_alias field, so that fixed the one problem I was having. However, I'm not seeing anything for the additional info about a role... example, "archive footage", "voice", "uncredited", etc. Am I missing something else here?

Example: http://www.imdb.com/title/tt2015381/fullcredits

[IzzySoft]

Afraid not: Looking at the changes of bcc1e22 it seems @tboothman didn't account for that kind of details (and I didn't think of them to mention).

@tboothman Could you please take a look at that (line 1203 ff)? Not sure if we can come up with a complete list of possibilities. But maybe one possible path is using a temp string with the full "role_cell", and always removing what's matched (episodes etc.). Then trim() the temp_string and check if something remained, and put it into a separate field (e.g. role_others). This way nothing gets lost.

[GeorgeFive]

Here's a complete list of every attribute in use as of right now. This includes some incredibly rare ones and typos, but all are included for the sake of completion. The ones marked with an asterisk are the ones that I'd personally add clauses for, if you want to go that route.

*voice
*uncredited
*archive footage
*credit only
*also archive footage
*rumored
*unconfirmed
*scenes deleted
*archive sound
*singing voice

--this is used 1,422 times, but shouldn't be needed with the existing clause--
episode

--everything below here is used less than 500 times on the site--
attached
additional voices
performers

--everything below here is used less than 100 times on the site--
song
songs
director
archive voice
stunts
sound supervisor
uncredited voice
special makeup effects artist
executive producer
in talks
original music by
production assistant
sound mixer
archieve footage
costume assistant
segment
uncredtied
camera operator
helicopter co-pilot
key craft service
post-production coordinator
producer
writer
aerial ground coordinator
archive audio
costumer
intern
paint artist
story editor

[tboothman]

Let me know what you think of this.
I don't think a ton of fields is the answer, so I've only added credited as a separate field as that feels like something someone would want to filter on. The others are just extra information that is just nice to have and not really important enough to warrant a field.

[IzzySoft]

I'd put that "extra" stuff into an extra field (just one field for "everything else"). Either "as-is" (remaining string), or as structured data (e.g. json). What do you think? Sounds reasonable?

[tboothman]

I'm not sure what you mean. What/why JSON? Which is the 'extra' stuff?

[IzzySoft]

Just scroll up a little, before your last comment. George points out a lot of "extra stuff" sometimes popping up. Before that, I've explained about the "string of remaining stuff".

Why JSON? Because that way, a single field could hold structured data. Hm, but taking a look at the last commit referenced above, it seems that's exactly what you've done – just using another array instead of JSON. Even easier to access that way. If that eats whatever might be left, that should be more than fine 😸