Update site_info.json to add TCCON sites info

[alexdanjou]

Addition of TCCON site details

[rt17603]

I have a question around this - up until now we've generally included 3-letter site codes for each of our in-situ measurement sites. Was there a reason this included the 2-letter site code (e.g. was this specified somewhere else)? Or did we want to try and stick to the 3-letter codes?

[alexdanjou]

The reason is that they are referred to with a 2-letter site code in the TCCON data. But I agree that it is not really good given the amount of other sites that we have. I can put a 3-letter code, no pb on this

[rt17603]

The reason is that they are referred to with a 2-letter site code in the TCCON data. But I agree that it is not really good given the amount of other sites that we have. I can put a 3-letter code, no pb on this

Yes, that makes sense - 3-letter codes look good though.

[rt17603]

Looks good but a few suggested updates:

• For these specific cases, could we include a tccon_name variable so we still have a record of the original name this has in TCCON?
• Would we be able to include the new sites alphabetically? Hopefully it doesn't make it less practical to not have the TCCON sites grouped, but in the rest of the file the sites are alphabetical so its easy to locate them.

[alexdanjou]

Localisations seems to be completely wrong, I will double check the script to create them

[alexdanjou]

Should be good now. Some sites have a changing latitude/longitude (change of a few cents of degrees). Put them in an array style but not sure this is the right things to do.
Also, ordering the keys and formatting with an indent of 2 before storing changed the display of the other sites. Is it ok?

[rt17603]

Can I just check - has is been reformatted because you ran black (or something else) over it?

[alexdanjou]

It's because I converted my dict to json with json.dumps(d, indent=2, sort_keys=True).

[alexdanjou]

Are we good to merge or should I format this differently?

[rt17603]

Are we good to merge or should I format this differently?

I am a little concerned about the reformatting across the whole file - just in case anyone else has been using a derivative of this file they'd like to be able to merge back in at some point. Would be it be possible to ensure the updates only impact the added lines or would that be difficult to do?

[alexdanjou]

I've done as much as I could to keep the old formattimg. But the identation was not homogeneous and some sites where not stored at their place (in terms of allphabetical order). Reordering just the sites and not the keys inside (i.e. "height", "latitude",..) would have been too much of a pain, so I there is some changes there too.

I think we should go with this as it is. It will anyway be easier to add stuff in the future in an homogeneous format. Tell me if you're not ok @rt17603

[rt17603]

By the way, I'm not saying this was done perfectly but there was originally a logic to the order of the inlets, with the first generally being the default used. So I think we may need to a bit cautious with reordering the heights lists even if it seems logical to order with largest inlet first.

[alexdanjou]

The heights are strictly speaking not reordered as they are not keys to the json files. But the network order is changed some times, this changes the first inlet when all networks have not the same first inlet. (hope this is clear)
Anyhow I'm planning to delete this PR and go for yours