ryanamannion / pcgs_scraper Goto Github PK

Would be nice if the user was able to specify the location that the data was downloaded to. Perhaps in a config file, or maybe as part of the CLI

May be useful to fold references to things like

Red = RD
Brown = BN
Full Bands = FB
etc.

and use that to make search more accurate

if there are multiple results use the remaining parts of the string with edit distance to rank

When there is only one price listed in an e.g. MS and MS+ category listing, if only the + price is listed it will be the first item in the price tuple, indicating that it is the price for MS

it may be worthwhile to make this an object

It could:

• keep track of the different structures if it is downloading it all in one go
• have methods for loading from pkl, json
• be a bit easier to understand how all the parts work together
• have unified parameters for prices and number scraper (e.g. delay_s)

maybe a milestone for v0.1.0

Would be nice to scrape images, or at least the URLs where you can find the images

Right now, only MS prices are pulled and stored. I want to pull e.g. Proof prices also in the future

PCGS# behavior is weird, documenting cases here

1909-S VDB:
Price Guide: has PCGS numbers 2426, 2427, 2428
PCGS number lookup: only 2428, 2426 and 2427 say they are invalid US coin numbers

RIght now how the system works is it uses the PCGS number guide to scrape detailed descriptions, and matches that up with the price guide. Any matches have the description added, any non-matches (i.e. only in price guide or only in numbers) are discarded.

That may need to be fixed...

would be nice to have URLs for coin facts and other things in the coin's dict

e.g. PCGS#82650 has a coinfacts URL of https://www.pcgs.com/coinfacts/coin/82650

this would be as simple as taking the PCGS# and adding it to the end of the standard URL https://www.pcgs.com/coinfacts/coin/