Log In
Home
(0 votes) 2:31 pm
March 1, 2023
Offline
So I'm seeing a whole bunch of these in my log:
[13-03-23 12:03:04] Processing a new post...
[13-03-23 12:03:04] Generating OpenAI GPT (turbo) content for: KBS Large 17-in-1 Bread Machine
[13-03-23 12:03:35] Success
[13-03-23 12:03:35] Executing user-defined PHP code
[13-03-23 12:03:35] Done
[13-03-23 12:03:35] The post will be skipped ($post = false).
I understand if the post exists already it will be skipped.
My question is, it seems like it's generating an article first from openAI and then check for duplicates?
Then it skips posting if it exists?
When I look at my openAI tokens it it's been depleted really fast but not too many articles being created...
If you can shed some light on this I would appreciate it.
Thanks
2:36 pm
July 2, 2009
Offline
I think you are asking about this: 
Login to see this link
Actually it depends how exactly the posts are being generated. From some source or from a dummy feed. Do they have post titles and source links before a request to OpenAI API? What are your duplicate check settings?
7:09 pm
March 1, 2023
Offline
Ok. thanks I'll take a look at it. I check both the title and URL for duplicates
8:36 pm
July 2, 2009
Offline
nevil7 said
Ok. thanks I'll take a look at it. I check both the title and URL for duplicates
According to that, your post gets sorted out by the custom PHP code. If that code returns false, the post will be skipped.
1:30 pm
March 1, 2023
Offline
As per our conversation via email. Just wanted to recap what we've talked about. correct me if I'm wrong.
We came to the conclusion based on the logs that in fact the article was being generated first then it checks for duplicates. It should be the other way around. see email below:
Login to see the quote
6:43 pm
July 2, 2009
Offline
I've checked your RSS file and found a problem there. Every time it gets pulled the same product URL's there are different. E.g.:
https://www.amazon.com/Ninja-AF101-Fryer-Black-gray/dp/B07FDJMC9Q/ref=sr_1_1?keywords=air+fryer&qid=1679160510&sr=8-1
https://www.amazon.com/Ninja-AF101-Fryer-Black-gray/dp/B07FDJMC9Q/ref=sr_1_1?keywords=air+fryer&qid=1679159700&sr=8-1
https://www.amazon.com/Ninja-AF101-Fryer-Black-gray/dp/B07FDJMC9Q/ref=sr_1_1?keywords=air+fryer&qid=1679159616&sr=8-1
...
So when plugin compares their links towards the ones stored in WordPress database, it considers them different items (as actually they are), makes an OpenAI GPT request and generates a duplicated post every time you pull the feed.
So the plugin is absolutely ok, while your feed is not, because every time it's get pulled, the same products there have different URL's...
When plugin compares their links towards the ones stored in WordPress database, it considers them different items (as actually they are), so it proceeds with an OpenAI GPT request and generates a duplicated post.
The plugin is absolutely ok, while your feed is not, because every time it's get pulled, the same products there have different URL's. Please fix it.
7:49 pm
March 1, 2023
Offline
But I'm filtering based on title only now, not url.
8:08 pm
July 2, 2009
Offline
nevil7 said
But I'm filtering based on title only now, not url.
Yes, you are. So there are no dupes anymore. Only unnecessary OpenAI API requests...
All the posts get checked for their source links before paid OpenAI API requests, 3rd-party spinners, translators etc. This is necessary to prevent their unnecessary use. The second tome they get checked for titles (in case if it's required by your duplicate check rules). I believe it's obvious that we can't check the post titles for uniqueness prior their modification by spinners, rewrites and translators.
8:12 pm
March 1, 2023
Offline
I'm not rewriting the title either with OpenAI. I leave the tiltle area blank
8:13 pm
July 2, 2009
Offline
Your are not. Someone else does it.
8:22 pm
March 1, 2023
Offline
So whats my solution? should change the php code to have it check for duplicate title or url first and if it exists skip it? I hate to throw away my money every time OpenAI creates and article and then finds out the post already exists and skips it...
not sure how to proceed...Please help
8:23 pm
July 2, 2009
Offline
Your solution is described above. Just fix the feed to prevent it from generation of random qid's every time it gets pulled. I don't know how necessary they are. If they have no sense, you may ask the plugin to remove them at all...
8:35 pm
March 1, 2023
Offline
ok. Don't think that's the issue though if I'm filtering based on title.
I'll try a different feed to see if I get the same results
based on the these logs:
13-03-23 12:03:04] Processing a new post…
[13-03-23 12:03:04] Generating OpenAI GPT (turbo) content for: KBS Large 17-in-1 Bread Machine
[13-03-23 12:03:35] Success
[13-03-23 12:03:35] Executing user-defined PHP code
[13-03-23 12:03:35] Done
[13-03-23 12:03:35] The post will be skipped ($post = false).
It creates an article first with OpenAI then it checks for duplicates. If it exists it skips it.
My understanding, It should check for duplicates first based on title or url (doesn't matter) then create an article.
Maybe I'm wrong but I'll find out with a new feed source soon.
8:41 pm
July 2, 2009
Offline
Unfortunately, your assumption is wrong. The title can not be checked before its possible modification by spinners, re-writers, translators and course your PHP snippets that may alter just everything...
As about the article/product source URL, so it's being checked before 3rd-part service requests to save your money. This is an obvious and only one possible solution according to a simple logic.
The easiest solution is to correct the broken feed.
8:47 pm
March 1, 2023
Offline
So how do I ask the plugin to remove the qid’s?
All RSSMost Users Ever Online: 541
Currently Online: air111003-air
9 Guest(s)
Currently Browsing this Page:
1 Guest(s)
Top Posters:
ninja321: 84
s.baryshev.aoasp: 68
Freedom: 61
Pandermos: 54
MediFormatica: 49
B8europe: 48
Member Stats:
Guest Posters: 337
Members: 2852
Moderators: 0
Admins: 1
Forum Stats:
Groups: 1
Forums: 5
Topics: 1640
Posts: 8352
Newest Members:
torontomark48, info.ckmedianetwork, contact.mybeautystar, samuelbodde, john.prush, creightonnick0Administrators: CyberSEO: 3947