FHIR-publisher-downloader.py

#!/usr/bin/env python

import requests
from bs4 import BeautifulSoup
import re
from requests import get
import subprocess as sp

# suppress 'InsecureRequestWarning' warning
# https://stackoverflow.com/a/28002687
from requests.packages.urllib3.exceptions import InsecureRequestWarning
requests.packages.urllib3.disable_warnings(InsecureRequestWarning)

# ======================================================================
URL     = "https://build.fhir.org/"

JAR_URL = "https://build.fhir.org/org.hl7.fhir.igpublisher.jar"

JAR_FILE_NAME = "org.hl7.fhir.igpublisher.jar"

GIT_CMD   = ['/usr/bin/git', 'log', '-1', '--pretty=%B']
#=======================================================================

def get_content():
    """

    """
    try:
        print(":: Getting FHIR current build version:")
        req = requests.get(URL, verify=False)
        return req

    except ConnectionError:
        message = "Could not connect"
        print(message)

def get_version_statement(data):
    # get the following footer text
    # ®© HL7.org 2011+. FHIR Release 4 Candidate (v3.1.0-12161) generated on Wed, Jun 21, 2017 19:17+0000.
    
    # get the content of the webpage
    content = get_content()

    soup = BeautifulSoup(content,"lxml")
    t = soup.find_all(string=re.compile("\d\.\d\.\d-\d{5}"))[0].lstrip().rstrip()
    return t


def get_version_number():
    # extract 'v3.1.0-12161' from
    # ®© HL7.org 2011+. FHIR Release 4 Candidate (v3.1.0-12161) generated on Wed, Jun 21, 2017 19:17+0000.
    data = get_version_statement()
    m = re.search(r"v\d\.\d\.\d-\d{5}", data)
    return m.group(0)

def get_ig_publisher_jar():
    # 
    print("Downloading IG publisher jar file ...")
    with open(JAR_FILE_NAME, "wb") as file:
        # get request
        response = get(JAR_URL)
        # write to file
        file.write(response.content)
        print("Download complete.")
        


def get_output(cmd):
    '''
    Runs a subprocess command to check its output
    Returns a byte, therefore needs to be converted to string
    and removal of the trailing '\n'
    '''
    output = sp.check_output(cmd)
    output = output.decode(encoding='UTF-8')                 # convert byte to string
    return output.strip()                                   # remove trailing '\n'


def get_last_commit_msg():
    # git log -1 --pretty=%B
    # need to split on line and retun first line only
    msg = get_output(GIT_CMD).split('\n')[0]
    return msg
    

# ======================================================================

if __name__ == '__main__':
    
    # get data
    version_statement = get_version_statement(content.text)
    current_version_number = get_version_number()
    last_commit_msg = get_last_commit_msg()
    
    # if last commit message contains the current version number
    if current_version_number in last_commit_msg:
        
        # then exit as the jar file is the latest
        print("The publisher jar is up-to-date.\nExiting.")
        exit(0)
    
    # the new jar file needs to be downloaded
    else:
        
        # download the publisher jar file
        print("get new jar")
        get_ig_publisher_jar()
        
        # run the publisher to check if no errors
        
        # when no errors
        # git add this file
        
        # git commit with the message of the full version statement
        
        # git push