Failing to get data for slideshow posts

[michaelgoodier]

Hi there - I'm failing to retrieve data for slideshow posts using tt_videos_hidden() - the post meta just doesn't seem to be in the json files. Thought i'd flag - it might need some individual node scraping of the html rather than parsing the json.

[JBGruber]

Can confirm. Here is a minimal example:

library(traktok)
df <- tt_videos_hidden("https://www.tiktok.com/@lemifn/photo/7470260357060136214")
#> ℹ Getting 1 unique link
#> ℹ Getting video 7470260357060136214
#> Warning: No video data found
#> ✔ Got video 7470260357060136214 (1/1).  [391ms]
#> 

^{Created on 2025-02-13 with reprex v2.1.1}

I'll look into it when I have time. They must have changed something.

[JBGruber]

The normal data does not contain the image links. It looks like the site makes another request to retrieve image data from "https://www.tiktok.com/api/item/detail/" with the post ID in a query parameter itemID. But it's unclear to me how these requests work. If I copy them from the browser to R, I get the image data the first time, but then it quickly breaks, presumably because my token (there is an X-Bogus and X-Gnarly token in the URL) expires.

TLDR: this is complicated and I probably won't be able to fix it

[JBGruber]

After writing the last comment, I played around with the same methods that fixed #14 and it works surprisingly okay. I drafted a function in a new branch now, but need to think how to integrate it into the package. Mainly, I'm unsure what to do when tt_videos_hidden encounters a slideshow. The approach is completely different, has more dependencies and is much 'heavier'. So maybe tt_videos_hidden should return the metadata and a message that tells the user how to use the separate tt_sldeshow to get missing image data, but not attempt to run anything automatically. Or it should check whether the requirements to run tt_sldeshow are satisfied and then just run it (which still has the issue that it would make tt_videos_hidden much slower.