So, thanks to
@annchir and Lawrence over at the Daily Brick I've now, er, 'extracted' information for some 20,000 parts from LEGO.com and the result is here
http://www.brickset.com/parts/browse/All crowd-sourced and dodgy data from BrickLink has been ditched. This is all 100% official, with the following exceptions:
-Colour names are official LEGO names, but converted to proper case.
-Category names are official LEGO names, but I've created some new ones for Duplo, because for example, the Duplo Bricks and the 'System' Bricks categories are both called 'Bricks' (but with a different category number, which isn't exposed through the LEGO CS site). So if a part has 'Duplo' in its description its category is set to 'Duplo ' + category. This hasn't caught all Duplo parts but has most of them.
The entire list can be exported as a CSV at
http://www.brickset.com/export/parts/. Double-quotes are changed to single-quotes in the part name field to prevent the CSV getting screwed up.
At the moment it, and all other reports for that matter, only includes parts for which there is an image at LEGO.com.
Comments
Is it possible to have all similar parts (same category) together. Under reddish brown, I can see a lot of minifig heads together, quite a few torsos, then another load of torsos, etc. Why is it that not all torsos are together?
Lots of other categories have just one piece or a few pieces in, but the category is repeated many times. For example, there are loads of lines for plates, shouldn't all plates be together?
And what is up with Chewbacca?
http://www.brickset.com/parts/?part=4499888
Is it snowing again?
On the parts browser, can you add a mouseover text with the color name? Sometimes I can't tell the difference from the TLG CAD colors.
If you haven't given up on crowdsourcing yet, can we create linking/mapping/translation to BrickLink or "common" values? First I'd like to see translatoin of the color names. Second would be to translate part descriptions. There are some that appear to be just plain typos on Lego's entry. If we can get the linkage down, you could build the cross-reference hyperlinks. "Buy this part on BrickLink"...
For example the hat / cap feather ... 64647. Too many sets.
The spartan helmet ... 92158 ... 0 sets
Good idea about the mouse overs.
Not all sets have been inventoried. I think there's something funny about the CMF inventories at Lego.com so they probable haven't been added. I'll look into it.
Even if it cannot be corrected for CMF, this is still a very useful resource.
I suspect some means of filtering the search results by category and maybe colour might be useful.
The only possible improvement I could imagine making would be to list also those parts for which pictures do not appear in the service.LEGO.com database, but alas, I can see why that might be a challenge, since it would be hard to link those parts to sets' inventories.
I suspect that LEGO will get better at providing accurage inventories for current sets so maybe it won't be such an issue going forwards.