generate sqlite db from csv's in repo... could offer more config opti…

[whatever]

Include Makefile and python script to generate db (#1)

• generate sqlite db from csv's in repo... could offer more config options fwiw

• ❌ remove trailing newline

Aside: I am using a sqlite db based on these CSV's to make a torch dataset:
https://github.com/whatever/nude-2.0/blob/matt/view-met-and-tag/src/nude2/data.py#L143-L152

[gregsadetsky]

hey I'm finally able to look into this! thanks again for your contribution!

few notes:

• I ran the Python code and was surprised that the resulting sqlite .db file did not have any rows in the chicago_images tables -- turns out there's a small mistake on this line i.e. it should be inserting into chicago_images, not met_images
• the included 2.sqlify/open-access-is-great-but-where-are-the-images.db file has this same issue i.e. there are no rows in the chicago_images tables

if you don't mind fixing that & regenerating the sqlite3 database, that'd be great! I'll be happy to merge after that

---

also I just opened an issue (#6) for a sqlite-related side idea if you're curious/interested! :-)

[whatever]

@gregsadetsky - ah this only includes the met_images by design! It would need another minor, separate tweak to accommodate chicago, since the ID type is VARCHAR(*) not NUMBER

[gregsadetsky]

@gregsadetsky - ah this only includes the met_images by design! It would need another minor, separate tweak to accommodate chicago, since the ID type is VARCHAR(*) not NUMBER

Just to make sure I understand, you do read from chicago-images.csv though, right?

https://github.com/gregsadetsky/open-access-is-great-but-where-are-the-images/pull/2/files#diff-2878a1dcb938b7463f42f6e736e2512f59dead0a92da93ae133d13d6a6eeb3cfR68

That's what surprised me -- i.e. reading from chicago but inserting into the met table

[whatever]

That's what surprised me -- i.e. reading from chicago but inserting into the met table

Ah I see now - I made the corresponding change!