Merge lp:~rodsmith/hwcert-tools/show-hw into lp:~hardware-certification/hwcert-tools/reporting-tools

[Rod Smith]

Added new show_hw script, which pulls down dmidecode data from C3. This script is a bit hacky, but is adequate for my immediate needs. Requesting merge in case it might be useful to others in the future, either as-is or as a base for further refinement.

Resubmission: Addressed Jeff's comment.

[Jeff Lane  (bladernr)]

Other than one minor nitpick that really is a semantics issue, I'm fine iwth this. If you don't want to change the shebang line, that's fine too, just say so and I'll approve.

review: Needs Information

[Jeff Lane  (bladernr)]

review: Approve

=== added file 'certification_reports/show_hw.py'
--- certification_reports/show_hw.py	1970-01-01 00:00:00 +0000
+++ certification_reports/show_hw.py	2014-11-20 17:04:44 +0000
@@ -0,0 +1,179 @@
+#!/usr/bin/python2
+#
+# Script to extract dmidecode data from C3. Unfortunately, security issues
+# require using both a C3 API key AND SSO credentials. The former are entered
+# on the command line and the script prompts for the latter. Results are
+# stored in a sequence of files, called dmidecode.#.{CID}.txt, where "#"
+# is the sequence number in which the dmidecode data was retrieved and
+# "{CID}" is the machine's CID number. Note that multiple dmidecode files
+# may be retrieved for some computers; I believe this depends on how many
+# were stored in C3.
+#
+# This script uses code provided to me by Daniel Manrique and written by
+# somebody else, whose name I don't recall.
+#
+# This script cannot retrieve dmidecode data by CID value, and some machines
+# which have stored data are NOT being retrieved. This is clearly a bug.
+#
+# The "maximum number of files to be retrieved" value is approximate; the
+# script may return one or two more or less than this.
+#
+# Required parameters:
+# - C3 username
+# - C3 API key (obtained from https://certification.canonical.com/me/)
+# - Maximum number of files to be retrieved (a run during development
+#   produced 7759 files, so specifying at least 10000 is advisable,
+#   unless you want just a few)
+#
+# By Rod Smith, 2014-10-23
+
+from api_utils import APIQuery, QueryError
+from argparse import ArgumentParser
+
+import os
+import sys
+import getpass
+import mechanize
+
+C3_URL = "https://certification.canonical.com"
+
+api = APIQuery(C3_URL)
+
+
+def get_browser():
+    """
+    Return a mechanize.Browser configured appropriately.
+    """
+    browser = mechanize.Browser()
+    browser.set_handle_robots(False)
+    return browser
+
+
+def prompt_for_credentials():
+    """
+    Return username and password collected from stdin.
+    """
+    print >> sys.stderr, "\rUbuntu SSO username (Canonical e-mail): ",
+    username = raw_input()
+    password = getpass.getpass()
+    return username, password
+
+
+def prompt_for_code():
+    """
+    Return code collected from stdin.
+    """
+    print >> sys.stderr, "\r2-factor authentication code: ",
+    return raw_input()
+
+
+def get_session_cookie(browser=get_browser()):
+    """
+    Fetch initial url, login with the supplied credentials, and return the
+    session id for the authenticated session.
+    """
+    bad_creds = False
+    bad_code = False
+
+    print >> sys.stderr, "Authenticating with Ubuntu SSO..."
+    browser.open("https://certification.canonical.com/openid/login")
+    while True:
+        # There are four possibilities
+        # 1- I'm already authenticated, so bapireak out and fetch the
+        # delicious cookie.
+        if "currently logged in " in browser.response().read():
+            break
+        # 2- I'm in an openid "Continue" page. I can tell because
+        # I have a "Continue" button.
+        browser.select_form(nr=0)
+        try:
+            browser.form.find_control(predicate=lambda c: 'value' in
+                                      getattr(c, 'attrs', '') and
+                                      c.attrs['value'] == 'Continue')
+            browser.submit()
+        except mechanize.ControlNotFoundError:
+            #3- I'm in a page asking for login/password *or* auth token.
+            try:
+                browser.form.find_control("oath_token")
+            except mechanize.ControlNotFoundError:
+                if bad_creds:
+                    print >> sys.stderr, "\rBad credentials"
+                bad_creds = True
+                username, password = prompt_for_credentials()
+                browser["email"] = username
+                browser["password"] = password
+                browser.submit()
+            else:
+                if bad_code:
+                    print >> sys.stderr, "\rBad code"
+                bad_code = True
+                code = prompt_for_code()
+                browser["oath_token"] = code
+                browser.submit()
+    # Get session id from cookiejar
+    cookiejar = browser._ua_handlers["_cookies"].cookiejar
+    oemshare_cookies = cookiejar._cookies["certification.canonical.com"]
+    return oemshare_cookies["/"]["sessionid"].value, browser
+
+
+def extract_cid(path):
+    head = path
+    for i in range(6):
+        head, tail = os.path.split(head)
+    return tail
+
+
+def get_dmi_data(username, api_key, max_records):
+    cookie, browser = get_session_cookie()
+    request_params = {"username": username,
+                      "api_key": api_key,
+                      "limit": 20,
+                      "offset": 0}
+    keep_going = True
+    i = 1
+
+    while keep_going:
+        try:
+            results = api.single_query(C3_URL + '/api/v1/attachment/',
+                                       params=request_params)
+        except QueryError:
+            print("Failed to retrieve metadata!")
+            sys.exit(1)
+
+        for k, v in results.iteritems():
+            if k == 'meta':
+                total_count = v['total_count']
+            if k == 'objects':
+                for att in v:
+                    if att['name'] == 'dmidecode_attachment':
+                        try:
+                            cid = extract_cid(att['url'])
+                            dmidec = browser.open(C3_URL + att['url'])
+                            filename = "dmidecode." + str(i) + "." + str(cid) + ".txt"
+                            print("Saving to " + filename)
+                            f = open(filename, "w")
+                            f.write(dmidec.read())
+                            f.close()
+                            i = i + 1
+                        except:
+                            print("Could not retrieve data")
+
+        if i >= max_records or request_params['offset'] > total_count:
+            keep_going = False
+        request_params['offset'] = request_params['offset'] + request_params['limit']
+
+
+def main():
+    parser = ArgumentParser("Extract dmidecode data from C3.")
+    parser.add_argument('username',
+                        help="Launchpad username used to access C3.")
+    parser.add_argument('api_key',
+                        help="API key used to access C3.")
+    parser.add_argument('max_records',
+                        help="Limit on number of attachments to save")
+    args = parser.parse_args()
+
+    get_dmi_data(args.username, args.api_key, int(args.max_records))
+
+if __name__ == "__main__":
+    sys.exit(main())