Import Hatena Diaries to Scrapbox with past dates

Objective.

When I search for Scrapbox, I don't get any hits on past blog posts, I think I should get a hit.

It's automatically converted to Hatena blog, but I'm not paying for it, so I get ads.

I feel bad when I link them.

I want to bracket past articles as well.

2021-12-20

Q: It's been six months.

A: Very good

In the past (2018) I had even created a "script to convert exported XML to JSON in Scrapbox format".

Why haven't you imported it into this project?

Formatting is not fully supported.

I didn't want the top of the page to be filled with mechanical articles.

→ I noticed that if you put the update date in the past, the top page won't be filled.

(computer) format

I don't think that's possible, but something lost came up and I'm not sure how to solve it.

It's not a big deal, so I parsed it on my own.

No conversion from Hatena notation to Scrapbox notation

I put it all together and put it in Scrapbox code notation.

If it hits the search and you can read it, that's all that matters.

Do `html.unescape

I was going to use the title of the blog as the page title, but decided against it because I felt it wasn't necessary.

All titles will be machine generated.

Easy mechanical removal if you decide it's not good enough after import

Q: Wouldn't it be better to make a backup before importing?

A: If I want to undo the import after running the import for a while and editing various pages, I can't restore from a backup.

Safe and secure override operation

It's sad when a script is updated and reconverted and then overwritten and the human-written text is lost.

This anxiety prevents us from "keeping it updated".

Don't change the machine-generated page, just create a page with a good title that can be duplicated when you want to change it.

https://gyazo.com/64bf2d44ad71f4b8515f6e6016d98f67

The creation date and time are correctly set to three years ago.

But, well, it might have been enough to just say "not on the top page" without having to match the exact date and time of creation.

code:python

def timestamp(*args):

return datetime.datetime(*map(int, args)).timestamp()

2021-06-29 PS: I still prefer the current "date and time the article was actually written" because I'm comfortable with the search results being in chronological order.

search (e.g. for someone using a search engine)

https://gyazo.com/e3626a3f075a86a8838aa67c3723f9d4

I made a heading page but deleted it because I knew I would never use it.

scale

120,000 lines of XML

220,000 lines of JSON

1500 pages

superscription

Schedule: Can be overwritten if the change date/time has not changed.

Schedule: Check for conflicts on import

Actual: Always overwritable

Try it on an empty project

It has to be admin.

Will bots be able to script their participation in the project as well?

code::

POST

---