photo

spiritfly

shared this problem
6 months ago

Employees Involved

photo

SCM

Admin

Statistics

8
Comments
1
Views

Share

Tags

12
votes

Found Critical Bug When Using Keywords From File

When using article creator to generate articles using "keywords from file" it behaves different than when using keywords entered in the field "combine keywords" regarding the writing of the output file. Here is how:

1. When trying to generate articles using the "keyword from file" and there is already file that has been previously generated, SEO CM says: "Skipping keyword 'thurs' cause output exists" and stops immediately without running the task. This is clearly a mistake because when you compare it to using the field "combine keywords" it simply creates a new text file while adding a number in case there is already an output file in the designated folder.

2. Again when trying to generate articles using the "keyword from file" and you check the "Overwrite" option, SEO CM will generate and overwrite the articles in the same run. For example if you have 3 keywords in the "keyword from file" and set "article count" to 6, SEO CM will generate 6 articles for keyword #1, then it will generate 6 articles for keyword #2 from the file and instead of saving them as new filenames(adding a number) it will OVERWRITE the previously generated 6 articles from the #1 keyword. And then it will overwrite them with the #3 keyword and so on.. Again this is a terrible bug in the software and gave me a lot of headaches until I found what it was.

I hope these will be resolved soon.

Official Answer
photo Employee
SCM Posted 6 months ago

Its not really a bug in so much as un-documented feature.

  1. When using keyword from file, it will skip over already existing files. The idea behind this is if you load a 100k keyword file, along the way some keywords will fail. So you can just re-run the task and SCM will re-process failed keywords.
  2. That's what overwrite option will do, instead of the numbering option, it will overwrite all content. The use case is. In GSA SER you load in static number of articles, then you re-run SCM and it will over-write those articles. You can't add new ones because GSA SER does not know about it.
Not sure if the documentation on those features is lacking.

Maybe you can suggest a way to make it clearer on the UI.

The article creator is a really complicated beast, we keep adding little extra features here and there and extra boxes, so it does complicate things.

Add Comment

Comments (8)

photo
11

The scenarios you have described seem to go well when using %keyword% as custom filename. However if you use a static placeholder such as %templatename% %projectname% %taskname% etc.. the party changes and weird thing start happening.

1. In the first case I described, SCM won't run at all. If there are 5 articles in the folder and the task is set to 5 articles count, it will not run no matter how many keywords there are in the filename OR if articles have been previously created or not by these keywords. There is no way SCM will know this because %keyword% as custom filename is how SCM knows which keywords have been processed already. So when used a static placeholder as custom filename it should run and write the new articles as new files instead in my opinion.

2. Again the same applies here and the issue becomes greater. I see how overwriting is useful especially when using GSA SER which is what I aim to do, but the thing is that SEO CM will overwrite the files multiple times DURING ONE RUN. This can't be useful. If there are 10 keywords in the file, SCM will process all of them, but will overwrite each result in the files and at the end only the last keyword generated articles will remain in the files.

Here is how SEO CM behaves during one run in the 2# scenario:

Settings:

keyword from file (10 keywords in file)

custom filename %template%

Articles count: 5

1. SCM starts scraping content

2. Creates 5 files and fills them with the content after it is spinned

3. Then begins to scrape content for second keyword

->> 4. Instead of creating new filenames and adding numbers, it overwrites the same 5 files that it created previously.

... it continues repeating step 3 and 4 for every keyword in the file.

Please note, it overwrites the files during ONE run. There is no time for GSA SER to sync the articles to GSA SER let alone SER read them.

The only exception to this is when using the placeholder %keyword% as custom filename. In this case SCM will behave as you described.

I hope I was clear what I was describing as it took me long time :)

I don't mind SCM being complicated at all, the more options the better. Your documentation on your website is great and I use it most of the time when I need to understand how things work in better details.

photo
8

Let me just add that using a static placeholder like %templatename% or %taskname% as filename is important as well and has it's uses. My personal usage is the following:

I have a batch script that will run at the end of each SCM task. This script will add a new set of keywords in the keyword list file I use in that task. So run the task gets new keywords from the file. So if I were to use %keyword% as filename I would get lots and lots of article files and they would never get overwritten. And if I use a static custom filename, I can easily replace my generated articles regardless of the keywords I have used to generate them.

The issue still remains though, how shold SCM behave in case %templatename% or %taskname% is used as filename..

photo Employee
8

What we need to do is decide if load from keyword file should ignore existing files or not.

What you described is not a bug. %taskname% will never create a unique filename by default.

It has to be decided if we let article creator add numbers to end of filename or to allow it to skip existing files.

photo
9

SCM wrote:

What we need to do is decide if load from keyword file should ignore existing files or not.

What you described is not a bug. %taskname% will never create a unique filename by default.

It has to be decided if we let article creator add numbers to end of filename or to allow it to skip existing files.

Hmm it's a rather complicated one. Can you tell me a bit more how both cases you suggested will behave in the following scenario:

1. keyword list file 150 keywords in file

2. set to 100 articles

3. there are 50 existing files in folder (templatename1-50.txt)

What will Article creator do if we decide to let it add numbers:

Create new files in the folder starting from 51?

- if so how many files per keyword will create?

Will not create articles for the first 50 keywords and start creating from 51 keyword onwards

Also how will it act if overwrite is checked?

I can see a number of options here of how it can be made, but I'm still not sure which one will fit best. I guess adding numbers would be the more productive one?

photo Employee
6

It would keep adding from 51 onwards.

It will restart from keyword 1 and do all 150 keywords.

It will always try to write to templatename.txt

Then appends the numbering.

Technically with overwrite checked, it will only write 1 file as it keeps overwriting itself due to the templatename always being the same. IE not very useful.

photo
8

The overwrite function could be used in the following way:

If there are 100 files with the templateN.txt (N being the numbers) it will overwrite all the 100 files and then the task would stop.

Otherwise numbering seems like a good option when not using the overwrite function.

photo Employee
7

spiritfly wrote:

The overwrite function could be used in the following way:

If there are 100 files with the templateN.txt (N being the numbers) it will overwrite all the 100 files and then the task would stop.

Otherwise numbering seems like a good option when not using the overwrite function.

For now I'm going to keep as is, cause I like that I can rerun task and it will only redo failed keywords in mass generate mode.

photo
1

SCM wrote:

spiritfly wrote:

The overwrite function could be used in the following way:

If there are 100 files with the templateN.txt (N being the numbers) it will overwrite all the 100 files and then the task would stop.

Otherwise numbering seems like a good option when not using the overwrite function.

For now I'm going to keep as is, cause I like that I can rerun task and it will only redo failed keywords in mass generate mode.
Yes! Keep as is. Perfect

Leave Comment

photo