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
|
from typing import Any, Dict, Optional
import requests
from grobid2json import teixml2json
from .ia import WaybackClient
from .misc import gen_file_metadata
from .workers import SandcrawlerFetchWorker, SandcrawlerWorker
class GrobidClient(object):
def __init__(self, host_url: str = "http://grobid.qa.fatcat.wiki", **kwargs):
self.host_url = host_url
self.consolidate_mode = int(kwargs.get('consolidate_mode', 0))
def process_fulltext(self, blob: bytes, consolidate_mode: Optional[int] = None) -> Dict[str, Any]:
"""
Returns dict with keys:
- status_code
- status (slug)
- error_msg (if status == 'error')
- tei_xml (if status is 200)
TODO: persist connection for performance?
"""
assert blob
if consolidate_mode is None:
consolidate_mode = self.consolidate_mode
assert consolidate_mode is not None
try:
grobid_response = requests.post(
self.host_url + "/api/processFulltextDocument",
files={
'input': blob,
'consolidateHeader': consolidate_mode,
'consolidateCitations': 0, # too expensive for now
'includeRawCitations': 1,
},
timeout=180.0,
)
except requests.Timeout:
return {
'status': 'error-timeout',
'status_code': -4, # heritrix3 "HTTP timeout" code
'error_msg': 'GROBID request (HTTP POST) timeout',
}
info: Dict[str, Any] = dict(status_code=grobid_response.status_code)
if grobid_response.status_code == 200:
info['status'] = 'success'
info['tei_xml'] = grobid_response.text
if len(info['tei_xml']) > 12000000:
# XML is larger than Kafka message size, and much larger than
# an article in general; bail out
info['status'] = 'error'
info['error_msg'] = "response XML too large: {} bytes".format(
len(info['tei_xml']))
info.pop('tei_xml')
else:
# response.text is .content decoded as utf-8
info['status'] = 'error'
info['error_msg'] = grobid_response.text[:10000]
return info
def metadata(self, result: Dict[str, Any]) -> Optional[Dict[str, Any]]:
if result['status'] != 'success':
return None
tei_json = teixml2json(result['tei_xml'], encumbered=False)
meta = dict()
biblio = dict()
for k in (
'title',
'authors',
'journal',
'date',
'doi',
):
if tei_json.get(k):
biblio[k] = tei_json[k]
meta['biblio'] = biblio
for k in ('grobid_version', 'grobid_timestamp', 'fatcat_release', 'language_code'):
if tei_json.get(k):
meta[k] = tei_json[k]
return meta
class GrobidWorker(SandcrawlerFetchWorker):
def __init__(self,
grobid_client: GrobidClient,
wayback_client: Optional[WaybackClient] = None,
sink: Optional[SandcrawlerWorker] = None,
**kwargs):
super().__init__(wayback_client=wayback_client)
self.grobid_client = grobid_client
self.sink = sink
self.consolidate_mode = 0
def timeout_response(self, task: Any) -> Any:
default_key = task['sha1hex']
return dict(
status="error-timeout",
error_msg="internal GROBID worker timeout",
source=task,
key=default_key,
)
def process(self, record: Any, key: Optional[str] = None) -> Any:
fetch_result = self.fetch_blob(record)
if fetch_result['status'] != 'success':
return fetch_result
blob = fetch_result['blob']
result = self.grobid_client.process_fulltext(blob,
consolidate_mode=self.consolidate_mode)
result['file_meta'] = gen_file_metadata(blob)
result['source'] = record
result['key'] = result['file_meta']['sha1hex']
return result
class GrobidBlobWorker(SandcrawlerWorker):
"""
This is sort of like GrobidWorker, except it receives blobs directly,
instead of fetching blobs from some remote store.
"""
def __init__(self,
grobid_client: GrobidClient,
sink: Optional[SandcrawlerWorker] = None,
**kwargs):
super().__init__()
self.grobid_client = grobid_client
self.sink = sink
self.consolidate_mode = 0
def process(self, blob: Any, key: Optional[str] = None) -> Any:
if not blob:
return None
result = self.grobid_client.process_fulltext(blob,
consolidate_mode=self.consolidate_mode)
result['file_meta'] = gen_file_metadata(blob)
result['key'] = result['file_meta']['sha1hex']
return result
|
