TR rules improvements (function, context, model names and descriptions)

[alicekwak]

TR rules improvements (function, context, model names and descriptions):

• TR rules for extracting model names and descriptions added
• More function rules and context rules added
• Cross sentence mention was added (to be used in combining function fragments)
• A method for model coreference resolution was added
• Expansion handler was edited to make adjustments for model description mentions
• Cosmos document, Json processor, and data loader were revised to filter out section headers

[codecov[bot]]

Codecov Report

Merging #259 (a64b512) into master (509e1dd) will increase coverage by 0.01%. The diff coverage is n/a.

@@            Coverage Diff             @@
##           master     #259      +/-   ##
==========================================
+ Coverage   67.19%   67.20%   +0.01%     
==========================================
  Files          94       94              
  Lines       17821    17821              
==========================================
+ Hits        11974    11976       +2     
+ Misses       5847     5845       -2     

Impacted Files	Coverage Δ
automates/model_analysis/graph_manipulation.py	70.75% <0.00%> (+0.55%)	:arrow_up:

---

Continue to review full report at Codecov.

Legend - Click here to learn more Δ = absolute <relative> (impact), ø = not affected, ? = missing data Powered by Codecov. Last update 509e1dd...a64b512. Read the comment docs.

[alicekwak]

Hi @alicekwak and @maxaalexeeva : Thank you for all of the work on this -- there is quite a lot. I cannot speak to the details, unfortunately, but I did read through everything and tried to follow what I can. T

he main question that this raises is: Do we have an "entry point" to how someone new coming onto this project can get to understanding the overall structure of all of the rules? I do like the detailed comments that explain how some rules are being suppressed b/c of interactions, that is helpful. But I'm not sure how to see the "big picture".

I suggest added one comment to include the summary interpretation that Masha provided.

Other than that, I think this is good and I approve.

Thanks!

Hi @cl4yton, thank you for the review! I'll add the comment as a note as you suggested. And as for your question, I don't think we have an "entry point" to the rules for now, except for the Odin manual. We might need to write a manual for our own TR rules and methods (maybe at the wiki page?) that can work as an entry point to someone newly joining our project. I think that will help them grasp the big picture of how the rules / methods work.

[maxaalexeeva]

@cl4yton if by entry points you mean methods to run the rules, there are a few places I can point you to. I'll try to start updating the wiki this week with the TR developments from the past year. @alicekwak I might request some write-ups from you as well for the stuff you've been working on.

[maxaalexeeva]

Ah I see @alicekwak also mentioned the wiki. Yeah, we should start updating that.

[alicekwak]

@cl4yton if by entry points you mean methods to run the rules, there are a few places I can point you to. I'll try to start updating the wiki this week with the TR developments from the past year. @alicekwak I might request some write-ups from you as well for the stuff you've been working on.

Sure! Just let me know which parts I should work on. I can write some explanations on function & context & model-related extractions (+ related methods).