Fix broken unicode

[rosnfeld]

Some of the Spanish descriptions have odd characters in them - is this a problem in the source data, the "downcasting" to utf-8 from utf-16, or something in the display?

Repro case: look at project number "SCR.550651.1.B2007".

[rosnfeld]

Hmm, looking at the raw unicode for the repro case I suspect it is the source data that is wrong. Emacs displays the same characters as Chrome does, when reading a freshly-unzipped CSV - so python doesn't even enter into it.

Also: I can see from the HTTP headers that django is sending UTF-8, and pasting various "unicode tests" into my html renders properly.

[rosnfeld]

Still, perhaps there is a way to "fix" the source data in these cases.

[rosnfeld]

I am changing the milestone on it and removing the "bug" label as I think it's now (yet another) project to fix CRS data.