To the right, my scraper code, to the left, some of the results scraped from Mr. Polt’s collection page. Sadly, not useable.
It won’t matter what these fields are called in your application. Here is a list, in order, of what fields should be exported. “Required” fields, if left blank, will result in that record not being imported:
1) Manufacturer/Marque [required](brand) Note: This field will be matched against typewriterdatabase.com’s table of manufacturers. if an exact non-case sensitive match occurs, the record will be accepted, otherwise, it will not be imported.
2) Year of Manufacture [required](should be an integer year like “1955”. if the collector is not sure about the exact date, use best guess and place astrisk * after date, like “1955*”)
3) Description (description of machine, about a paragraph.)
4) Model [required](model name, like “Quiet De Luxe” or “#10” or “Lettera 32”)
5) Serial Number [required](unique serial number engraved in frame)
and optionally, if your application supports it:
6) Collection (if machine is owned or not) Note: acceptable values are “My Collection” and “Sightings”. defaults to “My Collection” if value is out of range.
7) Photo URL (web-accessible URL of front-facing machine photo. JPG-only)
8) Typesample URL (web-accessible URL of type sample image. JPG-only)
The USB Typwriter guy gave away this neat Royal finger placement guide on his Facebook page. Tori claimed dibs and it came in the mail today. Thanks, Mr USB Typewriter guy! (:
Great looing typeface. It looks like Hermes Petit Pica. I think Herems called this font Small Pica Classic in the USA.
Speaking of Hermes. I will be adding another one to my collection. A 1963 elite script.
Gee, these things are habit forming.
Too bad the “scraping” didn’t work. Feel free to keep trying to exploit me! :D
Herman Price, like the accountant he is, has spreadsheets of actual serial numbers that he’s gathered for many unusual models. Could be useful for you. They don’t allow you to date a machine, usually, but they give a useful indication of the serial number range, which you could add to your pages on little-known makes.
Indeed! That would be useful. Would he be interested in having that information uploaded to the database?
I’ll ask Herman.
Just a thought about scraping the data from Richard P’s site: Since this will be a one-time data load, why not scrape the data as you show, then import it to excel and manually separate the manufacturer/model? With only 300 records, this shouldn’t be a big task. Then you could import the data to your database. I wouldn’t want to do this on an on-going basis, but once wouldn’t be too hard, if this is the only issue.
mainly because the HTML is irregularly formatted enough that my regex matches only catch about 50 of those 300+ machines, so the import to excel would be woefully incomplete.
My cruddy coding is what happens when you learn basic HTML in 1995 and type it in one bit at a time …
An Irishman I know refers to this process as ‘data-munging’. Imagine him trying to explain a strategy for exporting content from one site to another was especially side-splitting, what with his accent, bad phone lines and my woeful ignorance. Being of a more visual bent, I reckon the biggest joy your venture could be putting pictures to machines. I’ll aim to knock up a tab delimited file to your schema as soon as I can.