Finds Docs PRs, updates labels and opens in browser

find_docs.py

import webbrowser
import sys
import requests
import json

mytoken = 'secret_token'
url = 'https://api.github.com/repos/dotcloud/docker'
hurl = 'https://github.com/dotcloud/docker'
addlabels = {}
labelarray = [ '/project/doc' ] #start with the docs label, and then any others for code
addlabels['docs'] = json.dumps( labelarray )
# We're not going to add a second lable for code anymore
# since the core team will be watching all PRs.
addlabels['both'] = addlabels['docs']

headers = { 'Authorization' : 'token ' + mytoken }

# pull requests that contain docs (and possibly code)
docprs = {}
docprs['docs'] = set()
docprs['both'] = set()
docprs['new'] = set()

# dictionary of documents in the PRs, with a list of the PRs they occur in.
prdocs = {}

def postlabels(pr, labels):
    print "Adding %s to PR %s response=" % (addlabels[labels], pr),
    docurl = "%s/issues/%s/labels" % (url, pr)
    r = requests.post(docurl, headers=headers, data=addlabels[labels])
    print r.status_code

#-------- MAIN ----------

# Get a list of all the open pull requests
listresp=requests.get('%s/pulls' % url, headers=headers)
# And make a list of their IDs. Aren't list comprehensions fun?
ids=[p['number'] for p in listresp.json()]

# And get any additional pages!
while "next" in listresp.links and "url" in listresp.links["next"] and len(listresp.links["next"]["url"]):
    listresp = requests.get(listresp.links["next"]["url"], headers=headers)
    ids.extend([p['number'] for p in listresp.json()])

totalPRs = len(ids)
countPRs = 1
print "Found %s open PRs." % totalPRs
print "Getting files for each..."

# Go through each PR and get the files, then check for docs.
for i in ids:

    # Special case for a PR that has like a zillion files in it
    if i == 3408:
        print "Ignoring 3408 because it is too noisy."
        continue

    prString = "\r[{count}/{total}] {pr}".format(count=countPRs, total=totalPRs, pr=i)
    countPRs = countPRs + 1
    sys.stdout.write(prString)
    sys.stdout.flush()

    filesresp = requests.get('%s/pulls/%s/files' % (url, i), headers=headers)
    bHasCode = False
    for f in filesresp.json():
        fname = f['filename']
        if fname.startswith('docs/'):
            docprs['docs'].add(i)
            if fname in prdocs:
                prdocs[fname].add(i)
            else:
                prdocs[fname] = set()
                prdocs[fname].add(i)
        else:
            bHasCode = True

    if bHasCode and i in docprs['docs']:
        docprs['both'].add(i)

print "\nDone."

docsonly = docprs['docs'] - docprs['both']

print "Checking labels for the %s doc PRs:" % len(docprs['docs'])

# Add the doc label to doc PRs. PRs numbers are also issues.
for d in docprs['docs']:
    labelresp = requests.get('%s/issues/%s/labels' % (url, d), headers=headers)
    issuelabels = [i['name'] for i in labelresp.json()]
    hasDocLabel = labelarray[0] in issuelabels
    hasCodeLabel = True # HACK to remove second 'code' in issuelabels
    if d in docsonly:
        if hasDocLabel:
            print "Issue %s already labeled as %s." % (d, labelarray[0])
        else:
            postlabels(d, 'docs')
            docprs['new'].add(d)
    else: # Needs both labels
        if hasDocLabel and hasCodeLabel:
            print "Issue %s already has labels: %s" % (d, labelarray)
        else:
            postlabels(d, 'both')
            docprs['new'].add(d)

for d in docprs['docs']:
    dest = '%s/issues/%s' % (hurl, d)
    webbrowser.open(dest)

for file in prdocs:
    print file, prdocs[file]

print "docprs: ", docprs['docs']
print "Docs only: ", docsonly
print "Both: ", docprs['both']
print "NEW: ", docprs['new']