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
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
|
import sys
import json
import base64
import requests
from http.server import BaseHTTPRequestHandler, HTTPServer
from collections import namedtuple
from sandcrawler.ia import SavePageNowClient, CdxApiClient, WaybackClient, WaybackError, SavePageNowError, CdxApiError, PetaboxError, cdx_to_dict
from sandcrawler.grobid import GrobidClient
from sandcrawler.misc import gen_file_metadata
from sandcrawler.html import extract_fulltext_url
from sandcrawler.workers import SandcrawlerWorker
class IngestFileWorker(SandcrawlerWorker):
"""
High level flow is to look in history first, then go to live web if
resource not found. Following redirects is treated as "fetching a
resource". Current version fetches a single resource; if it isn't a hit
but is an HTML 200, treats it as a landing page, tries to extract
fulltext link, then fetches that resource.
process(request) -> response
Does all the things!
Check existing processing (short circuit):
check_existing_ingest(base_url) -> ingest_file_result or none
process_existing(result) -> response
try fetching all the rows we want. if any don't exist, fetch the resource itself and call process_hit()
Fetch resource:
find_resource(url) -> ResourceResult
Process resource:
process_hit(ResourceResult) -> response
process_grobid(ResourceResult)
"""
def __init__(self, sink=None, **kwargs):
super().__init__()
self.sink = sink
self.wayback_client = kwargs.get('wayback_client')
if not self.wayback_client:
self.wayback_client = WaybackClient()
self.spn_client = kwargs.get('spn_client')
if not self.spn_client:
self.spn_client = SavePageNowClient()
self.grobid_client = kwargs.get('grobid_client')
if not self.grobid_client:
self.grobid_client = GrobidClient()
self.try_existing_ingest = False
self.try_wayback = True
self.try_spn2 = True
def check_existing_ingest(self, base_url):
"""
Check in sandcrawler-db (postgres) to see if we have already ingested
this URL (ingest file result table).
Returns existing row *if* found *and* we should use it, otherwise None.
Looks at existing ingest results and makes a decision based on, eg,
status and timestamp.
"""
if not self.try_existing_ingest:
return None
raise NotImplementedError
# this "return True" is just here to make pylint happy
return True
def find_resource(self, url, best_mimetype=None):
"""
Looks in wayback for a resource starting at the URL, following any
redirects. If a hit isn't found, try crawling with SPN.
"""
via = "none"
resource = None
if self.try_wayback:
via = "wayback"
resource = self.wayback_client.lookup_resource(url, best_mimetype)
if self.try_spn2 and (not resource or not resource.hit):
via = "spn2"
resource = self.spn_client.crawl_resource(url, self.wayback_client)
print("[FETCH {}\t] {}\turl:{}".format(
via,
resource.status,
url),
file=sys.stderr)
return resource
def process_existing(self, request, result_row):
"""
If we have an existing ingest file result, do any database fetches or
additional processing necessary to return a result.
"""
result = {
'hit': result_row.hit,
'status': result_row.status,
'request': request,
}
# TODO: fetch file_meta
# TODO: fetch grobid
return result
def process_hit(self, resource, file_meta):
"""
Run all the necessary processing for a new/fresh ingest hit.
"""
return {
'grobid': self.process_grobid(resource),
}
def process_grobid(self, resource):
"""
Submits to resource body to GROBID for processing.
TODO: By default checks sandcrawler-db for an existing row first, then
decide if we should re-process
TODO: Code to push to Kafka might also go here?
"""
result = self.grobid_client.process_fulltext(resource.body)
if result['status'] == "success":
metadata = self.grobid_client.metadata(result)
if metadata:
result.update(metadata)
result.pop('tei_xml', None)
return result
def process(self, request):
# for now, only pdf ingest is implemented
assert request.get('ingest_type') == "pdf"
ingest_type = request.get('ingest_type')
base_url = request['base_url']
best_mimetype = None
if ingest_type == "pdf":
best_mimetype = "application/pdf"
existing = self.check_existing_ingest(base_url)
if existing:
return self.process_existing(request, existing)
result = dict(request=request, hit=False)
try:
# first hop
resource = self.find_resource(base_url, best_mimetype)
if not resource.hit:
result['status'] = resource.status
return result
file_meta = gen_file_metadata(resource.body)
if "html" in file_meta['mimetype']:
# got landing page, try another hop
fulltext_url = extract_fulltext_url(resource.terminal_url, resource.body)
result['html'] = fulltext_url
if not fulltext_url or not 'pdf_url' in fulltext_url:
result['status'] = 'no-pdf-link'
return result
print("\tlanding page URL extracted ({}): {}".format(
fulltext_url.get('technique'),
fulltext_url['pdf_url'],
),
file=sys.stderr)
resource = self.find_resource(fulltext_url['pdf_url'], best_mimetype)
if not resource.hit:
result['status'] = resource.status
return result
file_meta = gen_file_metadata(resource.body)
except SavePageNowError as e:
result['status'] = 'spn-error'
result['error_message'] = str(e)
return result
except PetaboxError as e:
result['status'] = 'petabox-error'
result['error_message'] = str(e)
return result
except CdxApiError as e:
result['status'] = 'cdx-error'
result['error_message'] = str(e)
return result
except WaybackError as e:
result['status'] = 'wayback-error'
result['error_message'] = str(e)
return result
if resource.terminal_dt:
result['terminal'] = {
"terminal_url": resource.terminal_url,
"terminal_dt": resource.terminal_dt,
"terminal_status_code": resource.terminal_status_code,
}
# must be a hit if we got this far
assert resource.hit == True
assert resource.terminal_status_code == 200
result['file_meta'] = file_meta
# other failure cases
if not resource.body or file_meta['size_bytes'] == 0:
result['status'] = 'null-body'
return result
if not (resource.hit and file_meta['mimetype'] == "application/pdf"):
result['status'] = "wrong-mimetype" # formerly: "other-mimetype"
return result
info = self.process_hit(resource, file_meta)
result.update(info)
result['status'] = "success"
result['hit'] = True
result['cdx'] = cdx_to_dict(resource.cdx)
return result
class IngestFileRequestHandler(BaseHTTPRequestHandler):
def do_POST(self):
if self.path != "/ingest":
self.send_response(404)
self.end_headers()
self.wfile.write("404: Not Found")
return
length = int(self.headers.get('content-length'))
request = json.loads(self.rfile.read(length).decode('utf-8'))
print("Got request: {}".format(request))
ingester = IngestFileWorker()
result = ingester.process(request)
self.send_response(200)
self.end_headers()
self.wfile.write(json.dumps(result))
|
