slarson
commented
11 years ago Great start! Your general structure looks good. A few comments below.
#OpenWorm CSV to JSON
#author: Gaston Gentile
import gspread
import csv
import json
import getpass
'''Para inicio de sesion automatico borrar variables user & pwd
y escibrilos entre comillas en la variable gc.
To automatic sign in, remove user & pwd variables, and write
them between quotes in gc variable respectively'''
#FIRST STEP
def start():
user = raw_input("Gmail user: ")
pwd = getpass.getpass("Gmail Password: ")
gc = gspread.login(user,pwd)
docurl = raw_input("Spreadsheet url: ")
sh = gc.open_by_url(docurl)#apertura del documento/Open Doc.
worksheet = sh.sheet1#Seleccion de la hoja/Sheet selection.Here it looks like you may not be selecting the correct sheet yet but I think you know this.
allsheet = worksheet.get_all_values()#Toma todos los valores de la hoja/Take all the values of the sheet.
parse = str(allsheet)#Convierte los datos a string/Parse data to string.
#Guarda el CSV en Celegans.csv/Save the CSV in Celegans.csv
csv = open('Celegans.csv', 'w')
escribir = csv.write(parse)
csv.close()
#SECOND STEP
#it's permited add more fieldnames, but i dont know which are the positions (neurotransmitter, neuron names, etc.)
def convert():
f = open('Celegans.csv','r')
reader = csv.DictReader(f,fieldnames=('fieldname1','fieldname2','fieldname3'),delimiter=';')
lineas = [l for l in reader]
salida = json.dumps(lineas)
For this part, I'm not sure this will format the JSON to have the required structure as we see in flare.json. Not all json is created alike. But you have correctly found the right libraries to use so bravo for that. For a next step, I would try using Celegans.json in place of flare.json in the pack-hierarchy example in your browser and see if it looks at all correct.
open('Celegans.json','w').write(salida)
start()
convert()
GasGen
We'd like to create a map of the completeness of the model and it involves a few different steps. We've found the following viz example that we'd like to try to build off of:
http://mbostock.github.com/d3/talk/20111116/pack-hierarchy.html
this uses the following library (http://d3js.org/)
The pack-hierarchy example is driven by the following JSON:
http://mbostock.github.com/d3/talk/20111116/flare.json
We figure if we can write JSON out in this manner that comes from the data sets that we have then it will be easy to create this visualization without having to do much d3js programming. So the problem is basically just converting a spreadsheet into this JSON.
For iteration 0, we'd like to have a Python script that can take this sheet that lists the c. elegans neurons and creates JSON like flare.json that can drive a visualization.
The simplest way to do this is to download the sheet as CSV and parse as text. A better way would be to use the GSpread library to access the spreadsheet programmatically on Google Drive.
The top level should be the neuron names in Column B, then as children from that top level should be the neurotransmitter in column D and the Receptor information in Column F. From here there are a lot of bells and whistles to add but this is a good place to start. We are interested in having a Python script rather than a program in another language because we'd like to put it under version control and evolve the visualization as we have additional data to put into it (we already do but merging the other data sets is yet another project).
The neurons send chemical signals using different molecules -- those are the neurotransmitters. The receptors are the "locks" that the "key" of the neurotransmitters fit into. The receptors then have their own dynamics that influences how the neurons work.