search

PESTools / pestools

PESTools
12 stars 10 forks source link

MF proceedings paper #29

Closed aleaf closed 9 years ago

aleaf commented 9 years ago

Evan, how about doing this in Markdown? See the readme file. It looks like we can meet all of their formatting requirements this way (I am using Mou locally, which can render to a PDF). Dataframes can be put out to html using the .to_html() method, and then pasted into the markdown (I had to first strip out the line endings and unicode stuff though).

My goal before I leave today is to get a spatial plot of identifiability implemented.

aleaf commented 9 years ago

Also, Mike his planning to implement his observation weighting routine sometime this week.

aleaf commented 9 years ago

turns out nothing is easy, and the PDF rendering by Mou looks great except for the figures, which are either pixelated (PDFs aren't maintained in vector format), or not sized right (if pngs are used). But an export of html can be opened in word and properly formatted without too much work. Maybe there's a markdown editor that is better at exporting, I don't know.

echristi commented 9 years ago

I think we can keep it in the notebook with markdown and code cells. I can convert to a format that will work at the end via MS word with figures. I'll mock something up to show what I'm thinking.

echristi commented 9 years ago

I tried a notebook here in my fork: https://github.com/echristi/pest_tools-1/blob/master/examples/MFandM_Test.ipynb

I'm not sure it's much better. We are going to have to decide how to show what it can do without blowing the page limit. Quickly generates too many figures.

aleaf commented 9 years ago

Evan, can you push that again? I'm trying to open it on my local copy of your fork (not sure if that is the best way to do it), but I'm getting a "bad request" message.

echristi commented 9 years ago

OK. Try now. I'm probably doing something wrong

aleaf commented 9 years ago

No, it was my bad. I hadn't updated ipython in awhile! This looks good. I say we just limit the demo to a few key tables and whatever figures will fit. We could use .head() to only display the first 5 lines.

aleaf commented 9 years ago

Evan, check out the notebook MFpaper_suggestions. Just some ideas on what we could include. I implemented one of Mike's rainbow bar plots, histogram grid, and a quick plot of pilot point identifiabilities. I did not make it to getting shapefile overlays included :(

These new methods could use a bit of work, but hopefully are OK for demo purposes. Let me know if you run into any issues.

I am out on vacation tomorrow thru Sunday, but can return to this when I get back (although I will be in Chicago helping out with a short course Mon-Wed next week). You maybe don't care about our bureaucratic problems, but I have two USGS reviewers lined up. I was hoping we could get it to them by early next week.

echristi commented 9 years ago

I'll take this and get it cleaned up into paper form by Sun. We can worry about code clean up later. This will be fine for the demo write up.

aleaf commented 9 years ago

Ok, obviously feel free to add whatever else you think is important- you’re the first author!

It might also be good to include one of those plots of objective fn. contribution (or something similar) by PEST iteration, to demonstrate the rei class.

On Mar 31, 2015, at 6:58 PM, echristi notifications@github.com wrote:

I'll take this and get it cleaned up into paper form by Sun. We can worry about code clean up later. This will be fine for the demo write up.

— Reply to this email directly or view it on GitHub.