Rewrite jupyter_parse_api.py to be object oriented

[somerandomsequence]

I think the API helper code will be better organized in an object oriented manner. Let's start with these classes:

• Library (equates to 'sample' in our current API)
• Sample (equates to 'position' in our current API)

And this initial specification:

• Library
  • init(id=None): initialization function that can take a single ID and performs a query to pull basic metadata into this object for that sample
  • search_by_ids([1,2,3]): static function that returns a list of Sample objects, one for each matching id
    • search_by_composition(only=[],not=[],any=[]): static function that returns a list of Sample objects, one for each library that matches the given element constraints
    • spectra(which=''): returns a pandas data frame containing all spectra data for a given sample
    • properties(): returns a pandas data frame containing all properties for a given library
    • samples(): returns a list of Sample objects for this library, each initialized with their properties and metadata
• Sample
  • init(id=None): initialization function that can take a single ID and perform a query to pull basic metadata for this position
  • properties(): identical to Library.properties() but just for this sample
  • spectra(which=''): ...

In addition, let's put these definitions in two files, 'library.py' and 'sample.py' in the 'lib' directory.

[meschw04]

Made the "lib" directory with files "library.py" and "sample.py". These are coming along pretty well so far, some library.py has a general structure while sample.py is nearly finished. You're right, the class structure is far more convenient. Let me know what you think of the "sample.py" file so I can apply the same concepts to the "library.py" file. Thanks!

[somerandomsequence]

This looks like a great start. I made one issue related to method variable names. In general, let's try to follow the PEP 8 style guidelines:

https://www.python.org/dev/peps/pep-0008/

You may want to try out IntelliJ Idea, an IDE that I've been using. Among other things, with the python module installed, it will highlight areas of code that deviate from PEP8, which is handy.

https://www.jetbrains.com/idea/specials/idea/idea.html

[meschw04]

Hi Caleb, There are currently functions within "library.py" called "search_by_ids" and "search_by_composition". You state that these are static functions that list "Sample objects". I'm a bit confused. Do you want these functions to look at positions that match certain constraints? Or do you want the functions to look at all samples that match these constraints? What exactly do you want the outputs of these functions to be? Thanks, Marcus

[somerandomsequence]

These functions are static because they don't use the state of the library object. I'm imagining they would return a list of Library objects that match the search criteria. For instance, I could do:

for lib in Library.search_by_composition(["Zn","Ni"]):
  prop = lib.properties()
  print "Library %d" % (lib.id,)
  for t in prop["thickness"]:
    print t

This would find all libraries having Zn or Ni and then print out the unique ID for each library and all the measured thickness values (44 per library).

Does that make sense?

[meschw04]

Pretty sure I got it right this time. It's being used pretty effectively in the notebooks now, let me know if you think a change is necessary.