recurring: Unexpected Data Format during capturing Instagram Data

[bodilvreugdenhil]

Hi!

It seems as though this exact problem has come up again:

#27

I experienced the same error as this user when trying to capture data from Instagram.
I did make sure that 4CAT and Zeeschuimer were completely up to date. Perhaps Instagram changed their data structure again recently.

Thanks in advance!

[dale-wahl]

Could you check the dataset log in 4CAT. Is the error the same as in this part: Unable to map item: KeyError-'full_name'? Or are you receiving the unexpected data format warning for a different key name?

[bodilvreugdenhil]

Yes, it is exactly that error. Unable to map item: KeyError-'full_name'

[dale-wahl]

Ah, the fix did not make the 1.40 release, but is available in the master version of 4CAT (digitalmethodsinitiative/4cat@fa3be93).

You can use the “latest” Docker tag in the .env file to use the most current version of 4cat which I checked and includes the above fix. After updating 4cat, the status message of the Dataset will not update (unless you re-import the data), but all processors will use the author_fullname field appropriately.

[namansnghl]

Hi, was this resolved? I see this problem came back and even the latest version does not solve it. Am I missing something?

"MapItemException (item 1): Unable to map item: KeyError-'like_count'"

[dale-wahl]

@namansnghl the original error with full_name was resolved and released.

Your error appears to be a different data issue which means Instagram reformatted their data again and someone will have to take a look.

[namansnghl]

@dale-wahl I will raise an issue on 4cat repo. Correct me if I am wrong - this is to be corrected on there and not on Zeeschuimer

[namansnghl]

Here is the URL to the issue I created
digitalmethodsinitiative/4cat#471