Missing posts from wordpress export

[dreeves]

### Desiderata
- [x] Get them back
- [ ] Diagnose how we missed them in the initial export
- [ ] Diagnose missing Disqus ID for /timecarrot post

The source doc URLs for these posts were missing in posts.json (called wp-posts.csv at the time):

• /pareto
• /copenhagen

I've now fixed that but am anxious to understand what happened in case it points to deeper problems.

Possibly relatedly, I'm seeing at least one other problem in wp-posts.csv: blog.bmndr.co/timecarrot has no disqus ID in the csv file but does have Disqus comments.

Cognata

• 318

• http://beeminder.com/changelog#4615 (THIS UVI)

Verbata: wordpress export, mendoza, missing posts,

[narthur]

@dreeves I don't think I'm using the disqus IDs in the csv for legacy posts. I'm generating those on the fly.

[dreeves]

generating those on the fly

Ah, how are they generated? Don't we want to include them explicitly in the metadata (aka frontmatter) for consistency? (This may be a dumb question and I may be making a lot of assumptions I don't realize I'm making about how this works.)

[I had a PS here that I'm moving to a separate comment]

[narthur]

@dreeves I just checked my source spreadsheet and it looks like those are the only two posts missing a source url that we need to worry about. There are two others--a test post and one that looks like a spam post with this title:

How Do You Get A Ton Of Likes On Instagram Ig-Up.com

[narthur]

@dreeves Hmm, I'm not sure. I guess I could make an argument either way for legacy posts. But yes, seems reasonable to go ahead and make those explicit, too, maybe at the same time we get rid of the csvs and go to a single json file.

[dreeves]

To review/clarify, there are perhaps 2 distinct issues: (1) diagnose how we missed /pareto and /copenhagen in the wordpress export, and (2) disqus IDs for wordpress-exported posts.

[narthur]

@dreeves We didn't miss them per se. They were in the CSVs. They were just missing the raw markdown source URLs.

[narthur]

So I guess the question would be why didn't the builds fail given there were no source urls for those posts.

[narthur]

Oh, nvm. I see what you meant. They weren't in sources.txt. I'm guessing I saw that those entries were missing source urls and assumed those weren't valid posts so manually removed them from sources.txt. So I'm guessing that was a bad assumption on my part.

[dreeves]

But how were the source URLs missing for those two in the first place? I can't imagine anything different about the way we originally published those on wordpress.

[narthur]

@dreeves that I don't know. I could give you the raw exported data I have, but beyond that I wouldn't know how to debug it without access to the old blog.

[dreeves]

We do have the full backup of the old wordpress site if necessary but can we start by checking what's in the raw exported data? Are /pareto and /copenhagen just mysteriously missing the source doc URLs? Nothing else different about them?