Merge pull request #989 from dufferzafar/har-extractor

Improve HAR Extractor Script
This commit is contained in:
Thomas Kriechbaumer 2016-03-04 19:55:52 +01:00
commit 428da2c4b1
4 changed files with 176 additions and 72 deletions

0
examples/__init__.py Normal file
View File

View File

@ -1,5 +1,4 @@
"""
This inline script utilizes harparser.HAR from
https://github.com/JustusW/harparser to generate a HAR log object.
"""
@ -17,7 +16,7 @@ class _HARLog(HAR.log):
__page_count__ = 0
__page_ref__ = {}
def __init__(self, page_list):
def __init__(self, page_list=[]):
self.__page_list__ = page_list
self.__page_count__ = 0
self.__page_ref__ = {}
@ -67,7 +66,7 @@ def start(context, argv):
'(- will output to stdout, filenames ending with .zhar '
'will result in compressed har)'
)
context.HARLog = _HARLog(['https://github.com'])
context.HARLog = _HARLog()
context.seen_server = set()
@ -83,8 +82,8 @@ def response(context, flow):
# Calculate the connect_time for this server_conn. Afterwards add it to
# seen list, in order to avoid the connect_time being present in entries
# that use an existing connection.
connect_time = flow.server_conn.timestamp_tcp_setup - \
flow.server_conn.timestamp_start
connect_time = (flow.server_conn.timestamp_tcp_setup -
flow.server_conn.timestamp_start)
context.seen_server.add(flow.server_conn)
if flow.server_conn.timestamp_ssl_setup is not None:
@ -92,8 +91,8 @@ def response(context, flow):
# the start of the successful tcp setup and the successful ssl
# setup. If no ssl setup has been made it is left as -1 since it
# doesn't apply to this connection.
ssl_time = flow.server_conn.timestamp_ssl_setup - \
flow.server_conn.timestamp_tcp_setup
ssl_time = (flow.server_conn.timestamp_ssl_setup -
flow.server_conn.timestamp_tcp_setup)
# Calculate the raw timings from the different timestamps present in the
# request and response object. For lack of a way to measure it dns timings
@ -112,80 +111,58 @@ def response(context, flow):
# HAR timings are integers in ms, so we have to re-encode the raw timings to
# that format.
timings = dict([(key, int(1000 * value))
for key, value in timings_raw.iteritems()])
timings = dict([(k, int(1000 * v)) for k, v in timings_raw.iteritems()])
# The full_time is the sum of all timings. Timings set to -1 will be ignored
# as per spec.
full_time = 0
for item in timings.values():
if item > -1:
full_time += item
# The full_time is the sum of all timings.
# Timings set to -1 will be ignored as per spec.
full_time = sum(v for v in timings.values() if v > -1)
started_date_time = datetime.fromtimestamp(
flow.request.timestamp_start,
tz=utc).isoformat()
started_date_time = datetime.utcfromtimestamp(
flow.request.timestamp_start).isoformat()
request_query_string = [{"name": k, "value": v}
for k, v in flow.request.query]
request_http_version = flow.request.http_version
# Cookies are shaped as tuples by MITMProxy.
request_cookies = [{"name": k.strip(), "value": v[0]}
for k, v in flow.request.cookies.items()]
request_headers = [{"name": k, "value": v} for k, v in flow.request.headers]
request_headers_size = len(str(flow.request.headers))
request_body_size = len(flow.request.content)
for k, v in flow.request.query or {}]
response_http_version = flow.response.http_version
# Cookies are shaped as tuples by MITMProxy.
response_cookies = [{"name": k.strip(), "value": v[0]}
for k, v in flow.response.cookies.items()]
response_headers = [{"name": k, "value": v}
for k, v in flow.response.headers]
response_headers_size = len(str(flow.response.headers))
response_body_size = len(flow.response.content)
response_body_decoded_size = len(flow.response.get_decoded_content())
response_body_compression = response_body_decoded_size - response_body_size
response_mime_type = flow.response.headers.get('Content-Type', '')
response_redirect_url = flow.response.headers.get('Location', '')
entry = HAR.entries(
{
entry = HAR.entries({
"startedDateTime": started_date_time,
"time": full_time,
"request": {
"method": flow.request.method,
"url": flow.request.url,
"httpVersion": request_http_version,
"cookies": request_cookies,
"headers": request_headers,
"httpVersion": flow.request.http_version,
"cookies": format_cookies(flow.request.cookies),
"headers": format_headers(flow.request.headers),
"queryString": request_query_string,
"headersSize": request_headers_size,
"bodySize": request_body_size,
"headersSize": len(str(flow.request.headers)),
"bodySize": len(flow.request.content),
},
"response": {
"status": flow.response.status_code,
"statusText": flow.response.msg,
"httpVersion": response_http_version,
"cookies": response_cookies,
"headers": response_headers,
"httpVersion": flow.response.http_version,
"cookies": format_cookies(flow.response.cookies),
"headers": format_headers(flow.response.headers),
"content": {
"size": response_body_size,
"compression": response_body_compression,
"mimeType": response_mime_type},
"redirectURL": response_redirect_url,
"headersSize": response_headers_size,
"mimeType": flow.response.headers.get('Content-Type', '')
},
"redirectURL": flow.response.headers.get('Location', ''),
"headersSize": len(str(flow.response.headers)),
"bodySize": response_body_size,
},
"cache": {},
"timings": timings,
})
# If the current url is in the page list of context.HARLog or does not have
# a referrer we add it as a new pages object.
if flow.request.url in context.HARLog.get_page_list() or flow.request.headers.get(
'Referer',
None) is None:
# If the current url is in the page list of context.HARLog or
# does not have a referrer, we add it as a new pages object.
if (flow.request.url in context.HARLog.get_page_list() or
flow.request.headers.get('Referer') is None):
page_id = context.HARLog.create_page_id()
context.HARLog.add(
HAR.pages({
@ -215,7 +192,7 @@ def done(context):
"""
Called once on script shutdown, after any other events.
"""
from pprint import pprint
import pprint
import json
json_dump = context.HARLog.json()
@ -239,6 +216,18 @@ def done(context):
)
def format_cookies(obj):
if obj:
return [{"name": k.strip(), "value": v[0]} for k, v in obj.items()]
return ""
def format_headers(obj):
if obj:
return [{"name": k, "value": v} for k, v in obj.fields]
return ""
def print_attributes(obj, filter_string=None, hide_privates=False):
"""
Useful helper method to quickly get all attributes of an object and its

View File

@ -0,0 +1,78 @@
{
"test_response": {
"log": {
"__page_count__": 1,
"version": "1.2",
"creator": {
"comment": "",
"version": "0.1",
"name": "MITMPROXY HARExtractor"
},
"pages": [
{
"startedDateTime": "1993-08-24T14:41:12",
"id": "autopage_1",
"title": "http://address:22/path"
}
],
"entries": [
{
"pageref": "autopage_1",
"startedDateTime": "1993-08-24T14:41:12",
"cache": {},
"request": {
"cookies": [],
"url": "http://address:22/path",
"queryString": [],
"headers": [
{
"name": "header",
"value": "qvalue"
},
{
"name": "content-length",
"value": "7"
}
],
"headersSize": 35,
"httpVersion": "HTTP/1.1",
"method": "GET",
"bodySize": 7
},
"timings": {
"receive": 0,
"ssl": 1000,
"connect": 1000,
"send": 0,
"wait": 0
},
"time": 2000,
"response": {
"status": 200,
"cookies": [],
"statusText": "OK",
"content": {
"mimeType": "",
"compression": 0,
"size": 7
},
"headers": [
{
"name": "content-length",
"value": "7"
},
{
"name": "header-response",
"value": "svalue"
}
],
"headersSize": 44,
"redirectURL": "",
"httpVersion": "HTTP/1.1",
"bodySize": 7
}
}
]
}
}
}

View File

@ -0,0 +1,37 @@
import json
import netlib.tutils
from . import tutils
from examples import har_extractor
class Context(object):
pass
trequest = netlib.tutils.treq(
timestamp_start=746203272,
timestamp_end=746203272,
)
tresponse = netlib.tutils.tresp(
timestamp_start=746203272,
timestamp_end=746203272,
)
def test_start():
tutils.raises(ValueError, har_extractor.start, Context(), [])
def test_response():
ctx = Context()
ctx.HARLog = har_extractor._HARLog([])
ctx.seen_server = set()
fl = tutils.tflow(req=trequest, resp=tresponse)
har_extractor.response(ctx, fl)
with open(tutils.test_data.path("data/har_extractor.har")) as fp:
test_data = json.load(fp)
assert json.loads(ctx.HARLog.json()) == test_data["test_response"]