eric-czech

import requests
from urllib.parse import quote_plus

def get_info(disease):
    url_fmt = 'http://www.ebi.ac.uk/spot/zooma/v2/api/services/annotate?propertyValue={disease}&propertyType=disease&filter=ontologies:[efo]'
    url = url_fmt.format(disease=quote_plus(disease))
    print(url)
    res = requests.get(url).json()
    if len(res) == 0:
        return None

eric-czech

def search_efo(disease):
    import requests
    from urllib.parse import quote_plus
    res = requests.get(f"https://www.ebi.ac.uk/ols/api/select?q={quote_plus(disease)}&ontology=efo")
    res = res.json()
    docs = res['response']['docs']
    if len(docs) == 0:
        return None
    return docs[0]

eric-czech

eric-czech

library(ggplot2)

# For figure 2 of paper https://www.semanticscholar.org/paper/Dealing-with-Logs-and-Zeros-in-Regression-Models-Bellégo-Pape/546a6f45433b413721f9a60f0be8e3e2b69fe103
set.seed(1)
beta <- 1
x_max <- 1
x <- x_max * runif(10000)
y <- sapply(x, function(x) rpois(1, lambda=exp(beta * x + 1)))
df <- do.call(rbind, lapply(seq(.01, 1, by=.01), function(delta){
  beta_hat <- unname(lm(log(y + delta) ~ x)$coefficients['x'])

eric-czech

def get_ols(term):
    import requests
    url = 'http://www.ebi.ac.uk/ols/api/ontologies/efo/terms?obo_id=' + term
    res = requests.get(url)
    terms = res.json()['_embedded']['terms']
    assert len(terms) == 1
    return terms[0]

eric-czech

eric-czech

eric-czech

eric-czech

eric-czech

# Dask Cloud Provider REPL
# 
# This is useful for creating clusters indepedent of the code that runs on them
# Example: python scripts/cloudprovider.py -- --interactive
#
from dask_cloudprovider.gcp.instances import GCPCluster
import fire
import os
import json