1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
|
# XXX: some broken MRO thing going on in here due to python3 object wrangling
# in `wayback` library. Means we can't run pylint.
# pylint: skip-file
import os, sys, time
import requests
import datetime
import wayback.exception
from http.client import IncompleteRead
from wayback.resourcestore import ResourceStore
from gwb.loader import CDXLoaderFactory
from .misc import b32_hex, requests_retry_session
class CdxApiError(Exception):
pass
class CdxApiClient:
def __init__(self, host_url="https://web.archive.org/cdx/search/cdx"):
self.host_url = host_url
def lookup_latest(self, url, recent_only=True, follow_redirects=False):
"""
Looks up most recent HTTP 200 record for the given URL.
Returns a CDX dict, or None if not found.
XXX: should do authorized lookup using cookie to get all fields
"""
WAYBACK_ENDPOINT = "https://web.archive.org/web/"
since = datetime.date.today() - datetime.timedelta(weeks=4)
params = {
'url': url,
'matchType': 'exact',
'limit': -1,
'output': 'json',
}
if recent_only:
params['from'] = '%04d%02d%02d' % (since.year, since.month, since.day),
if not follow_redirects:
params['filter'] = 'statuscode:200'
resp = requests.get(self.host_url, params=params)
if resp.status_code != 200:
raise CdxApiError(resp.text)
rj = resp.json()
if len(rj) <= 1:
return None
cdx = rj[1]
assert len(cdx) == 7 # JSON is short
cdx = dict(
surt=cdx[0],
datetime=cdx[1],
url=cdx[2],
mimetype=cdx[3],
http_status=int(cdx[4]),
sha1b32=cdx[5],
sha1hex=b32_hex(cdx[5]),
)
if follow_redirects and cdx['http_status'] in (301, 302):
resp = requests.get(WAYBACK_ENDPOINT + cdx['datetime'] + "id_/" + cdx['url'])
assert resp.status_code == 200
next_url = '/'.join(resp.url.split('/')[5:])
assert next_url != url
return self.lookup_latest(next_url)
return cdx
class WaybackError(Exception):
pass
class WaybackClient:
def __init__(self, cdx_client=None, **kwargs):
if cdx_client:
self.cdx_client = cdx_client
else:
self.cdx_client = CdxApiClient()
# /serve/ instead of /download/ doesn't record view count
self.petabox_base_url = kwargs.get('petabox_base_url', 'http://archive.org/serve/')
# gwb library will fall back to reading from /opt/.petabox/webdata.secret
self.petabox_webdata_secret = kwargs.get('petabox_webdata_secret', os.environ.get('PETABOX_WEBDATA_SECRET'))
self.warc_uri_prefix = kwargs.get('warc_uri_prefix', 'https://archive.org/serve/')
self.rstore = None
def fetch_warc_content(self, warc_path, offset, c_size):
warc_uri = self.warc_uri_prefix + warc_path
if not self.rstore:
self.rstore = ResourceStore(loaderfactory=CDXLoaderFactory(
webdata_secret=self.petabox_webdata_secret,
download_base_url=self.petabox_base_url))
try:
gwb_record = self.rstore.load_resource(warc_uri, offset, c_size)
except wayback.exception.ResourceUnavailable:
raise WaybackError("failed to load file contents from wayback/petabox (ResourceUnavailable)")
except ValueError as ve:
raise WaybackError("failed to load file contents from wayback/petabox (ValueError: {})".format(ve))
except EOFError as eofe:
raise WaybackError("failed to load file contents from wayback/petabox (EOFError: {})".format(eofe))
except TypeError as te:
raise WaybackError("failed to load file contents from wayback/petabox (TypeError: {}; likely a bug in wayback python code)".format(te))
# Note: could consider a generic "except Exception" here, as we get so
# many petabox errors. Do want jobs to fail loud and clear when the
# whole cluster is down though.
if gwb_record.get_status()[0] != 200:
raise WaybackError("archived HTTP response (WARC) was not 200: {}".format(gwb_record.get_status()[0]))
try:
raw_content = gwb_record.open_raw_content().read()
except IncompleteRead as ire:
raise WaybackError("failed to read actual file contents from wayback/petabox (IncompleteRead: {})".format(ire))
return raw_content
def fetch_url_datetime(self, url, datetime):
cdx_row = self.cdx_client.lookup(url, datetime)
return self.fetch_warc_content(
cdx_row['warc_path'],
cdx_row['warc_offset'],
cdx_row['warc_csize'])
class SavePageNowError(Exception):
pass
class SavePageNowClient:
def __init__(self, cdx_client=None,
v1endpoint="https://web.archive.org/save/",
v2endpoint="https://web.archive.org/save"):
if cdx_client:
self.cdx_client = cdx_client
else:
self.cdx_client = CdxApiClient()
self.ia_access_key = os.environ.get('IA_ACCESS_KEY')
self.ia_secret_key = os.environ.get('IA_SECRET_KEY')
self.v1endpoint = v1endpoint
self.v2endpoint = v2endpoint
self.http_session = requests_retry_session(retries=5, backoff_factor=3)
self.http_session.headers.update({
'User-Agent': 'Mozilla/5.0 sandcrawler.SavePageNowClient',
})
self.v2_session = requests_retry_session(retries=5, backoff_factor=3)
self.v2_session.headers.update({
'User-Agent': 'Mozilla/5.0 sandcrawler.SavePageNowClient',
'Accept': 'application/json',
'Authorization': 'LOW {}:{}'.format(self.ia_access_key, self.ia_secret_key),
})
def save_url_now_v1(self, url):
"""
Returns a tuple (cdx, blob) on success of single fetch, or raises an
error on non-success.
"""
resp = self.http_session.get(self.v1endpoint + url)
if resp.status_code != 200:
raise SavePageNowError("HTTP status: {}, url: {}".format(resp.status_code, url))
terminal_url = '/'.join(resp.url.split('/')[5:])
body = resp.content
cdx = self.cdx_client.lookup_latest(terminal_url)
if not cdx:
raise SavePageNowError("SPN was successful, but CDX lookup then failed")
return (cdx, body)
def save_url_now_v2(self, url):
"""
Returns a list of cdx objects, or raises an error on non-success.
"""
if not (self.ia_access_key and self.ia_secret_key):
raise Exception("SPNv2 requires authentication (IA_ACCESS_KEY/IA_SECRET_KEY)")
resp = self.v2_session.post(
self.v2endpoint,
data={
'url': url,
'capture_all': 1,
'if_not_archived_within': '1d',
},
)
if resp.status_code != 200:
raise SavePageNowError("HTTP status: {}, url: {}".format(resp.status_code, url))
resp_json = resp.json()
assert resp_json
# poll until complete
for i in range(90):
resp = self.v2_session.get("{}/status/{}".format(self.v2endpoint, resp_json['job_id']))
resp.raise_for_status()
status = resp.json()['status']
if status == 'success':
resp = resp.json()
if resp.get('message', '').startswith('The same snapshot had been made'):
raise SavePageNowError("SPN2 re-snapshot withing short time window")
break
elif status == 'pending':
time.sleep(1.0)
else:
raise SavePageNowError("SPN2 status:{} url:{}".format(status, url))
#print(resp)
return resp['resources']
|