import sys
if sys.hexversion < 0x030000F0:
    raise RuntimeError("This script requires Python3")
"""
   This script runs a couple of Bugzilla searches:
   - all bugs that were created in the last 3 months  
   - all RESOLVED bugs that changed to RESOLVED in the last 3 months
   and creates an output file:
   { "pmc1" : [created, resolved],
     "pmc2: ...
   }

    Output:
    ../data/bugzillastats.json
"""
import errtee  # this is used, even though it is not referenced
import json
import requests
from collections import defaultdict

# There are 3 Bugzilla instances
BASE = "https://bz.apache.org/bugzilla/rest.cgi/bug?include_fields=product"
BASEOO = "https://bz.apache.org/ooo/rest.cgi/bug?include_fields=product"
BASESA = "https://bz.apache.org/SpamAssassin/rest.cgi/bug?include_fields=product"

CREATED = "&f1=creation_ts&o1=greaterthaneq&v1=-3m"
RESOLVED = "&chfield=bug_status&chfieldvalue=RESOLVED&chfieldfrom=-3m&chfieldto=Now"

# Maintained in Puppet; contain api key(s)
BUGZILLA_CREDFILE = '/usr/local/etc/tokens/bugzilla%s.txt'

__BUGZILLA_STATS = "../data/bugzillastats.json"

def getapikey(suffix=''):
    file = BUGZILLA_CREDFILE % suffix
    try:
        with open(file, "r") as f:
            apikey = f.read().strip()
    except Exception:
        raise RuntimeError(f"Failed to read/parse Bugzilla credentials {file}")
    if apikey == '*': # for local testing
        import getpass
        apikey = getpass.getpass(f"API key #{suffix}: ")
    return apikey

def getCSV(url, apikey):
    csv = defaultdict(int)
    rsp = requests.get(f"{url}&api_key={apikey}")
    # We don't use raise_for_status() because that would reveal the API key
    if not rsp.status_code == 200:
        raise RuntimeError(f"Code {rsp.status_code} getting {url}")
    text = rsp.text
    bugs = json.loads(text)['bugs']
    for k in bugs:
        csv[k['product']] += 1
    return csv

# test data:
# created = {'JMeter': 57, 'Ant': 15, 'APR': 10, 'Log4j': 3, 'WebSH': 1, 'Tomcat 9': 4, 'Apache httpd-2': 64, 
#          'Tomcat 7': 25, 'Tomcat Connectors': 8, 'Tomcat 6': 4, 'Taglibs': 1, 'Tomcat Native': 6, 'Tomcat 8': 85, 'POI': 90}
# resolved = {'JMeter': 33, 'Ant': 8, 'APR': 2, 'Log4j': 2, 'WebSH': 1, 'Tomcat 9': 3, 'Apache httpd-2': 35, 
#           'Tomcat 7': 26, 'Tomcat Modules': 1, 'Tomcat Connectors': 7, 'Tomcat 6': 5, 'Taglibs': 2, 'Tomcat Native': 1, 'Tomcat 8': 88, 'POI': 115, 'Rivet': 0}

print("Getting Bugzilla API keys")
apikey = getapikey()
apikeyoo = getapikey('oo')
apikeysa = getapikey('sa')

stats = {}

# Other bugzilla users are Ant, APR, JMeter, POI
prod2pmc = {
    'log4j': 'logging',
    'websh': 'tcl',
    'rivet': 'tcl',
    'taglibs': 'tomcat',
    }

def getPMC(product):
    if product.startswith("Tomcat "):
        return 'tomcat'
    if product.startswith('Apache httpd'):
        return 'httpd'
    low = product.lower()
    if low in prod2pmc:
        return prod2pmc[low]
    return low

def addCount(product, index, count, pmcoverride=None):
    pmc = pmcoverride or getPMC(product)
    if not pmc in stats:
        stats[pmc]=[0,0,{}]
    stats[pmc][index] += count
    try:
        stats[pmc][2][product][index] += count
    except KeyError:
        stats[pmc][2][product] = [0,0]
        stats[pmc][2][product][index] = count


def getCounts(base, apikey, override=None):
    print(f"Getting list of {override or ''} bugs created in the last 3 months")
    created = getCSV(base + CREATED, apikey)
    for product in created:
        addCount(product, 0, created[product], override)

    print(f"Getting list of {override or ''} bugs resolved in the last 3 months")
    resolved = getCSV(base + RESOLVED, apikey)
    for product in resolved:
        addCount(product, 1, resolved[product], override)

getCounts(BASE, apikey)
getCounts(BASEOO, apikeyoo, 'openoffice')
getCounts(BASESA, apikeysa, 'spamassassin')

print("Writing " + __BUGZILLA_STATS)
with open(__BUGZILLA_STATS,"w") as f:
    json.dump(stats, f, indent=1, sort_keys=True)
    f.close()
print("All done")