datamule 1.6.4__py3-none-any.whl → 1.6.6__py3-none-any.whl

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
@@ -12,9 +12,8 @@ class SecConnector:
12
12
 
13
13
  self.quiet = quiet
14
14
  self.auth_url = "https://sec-websocket-auth-worker.jgfriedman99.workers.dev/"
15
- self.websocket_url = "ws://3.80.249.191:8080/ws"
16
15
 
17
- def _get_jwt_token(self):
16
+ def _get_jwt_token_and_ip(self):
18
17
  if not self.quiet:
19
18
  print("Getting JWT token...")
20
19
 
@@ -33,11 +32,12 @@ class SecConnector:
33
32
  if not self.quiet:
34
33
  print("JWT token obtained")
35
34
 
36
- return data['token']
35
+ return data['token'], data['websocket_ip']
37
36
 
38
37
  def connect(self, data_callback=None):
39
- token = self._get_jwt_token()
40
- ws_url = f"{self.websocket_url}?token={token}"
38
+ token,websocket_ip = self._get_jwt_token_and_ip()
39
+ ws_url = f"ws://{websocket_ip}/ws?token={token}"
40
+ print(ws_url)
41
41
 
42
42
  if not self.quiet:
43
43
  print("Connecting to WebSocket...")
@@ -65,6 +65,7 @@ def clean_efts_hits(hits):
65
65
  class Monitor():
66
66
  def __init__(self):
67
67
  self.accessions = deque(maxlen=50000)
68
+ self.accessions_set = set()
68
69
  self.ratelimiters = {'sec.gov': PreciseRateLimiter(rate=5)}
69
70
  self.efts_query = EFTSQuery(quiet=True)
70
71
  self.efts_query.limiter = self.ratelimiters['sec.gov']
@@ -250,7 +251,14 @@ class Monitor():
250
251
  new_items = []
251
252
  for item in items:
252
253
  accession = item['accession']
253
- if accession not in self.accessions:
254
+ if accession not in self.accessions_set: # O(1) lookup instead of O(n)
255
+ # Check if deque is at capacity and will evict an item
256
+ if len(self.accessions) == self.accessions.maxlen:
257
+ evicted_accession = self.accessions[0] # Get leftmost item that will be evicted
258
+ self.accessions_set.discard(evicted_accession) # Remove from set
259
+
260
+ # Add to both structures
254
261
  self.accessions.append(accession)
262
+ self.accessions_set.add(accession)
255
263
  new_items.append(item)
256
264
  return new_items
@@ -1,6 +1,6 @@
1
1
  Metadata-Version: 2.1
2
2
  Name: datamule
3
- Version: 1.6.4
3
+ Version: 1.6.6
4
4
  Summary: Work with SEC submissions at scale.
5
5
  Home-page: https://github.com/john-friedman/datamule-python
6
6
  Author: John Friedman
@@ -9,7 +9,7 @@ datamule/sheet.py,sha256=TvFqK9eAYuVoJ2uWdAlx5EN6vS9lke-aZf7FqtUiDBc,22304
9
9
  datamule/submission.py,sha256=ooLsesZ5HkgSWyEFID4u08CobTxdo35eAUHSCB6fw2k,10332
10
10
  datamule/data/listed_filer_metadata.csv,sha256=dT9fQ8AC5P1-Udf_UF0ZkdXJ88jNxJb_tuhi5YYL1rc,2426827
11
11
  datamule/datamule/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
12
- datamule/datamule/sec_connector.py,sha256=T3edE7I-d4oHysqj7zYlIOxH3Fuauj9tfw39UdFWvB8,2393
12
+ datamule/datamule/sec_connector.py,sha256=sr80z_43xXJe3hS8Wfir8LOsNgAAuFIXsFq7xeixc-I,2409
13
13
  datamule/document/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
14
14
  datamule/document/document.py,sha256=YGo-Iz_sBXekUeKEAoNJV0BiLDtSOgD9OXFo2FocYq8,14439
15
15
  datamule/document/processing.py,sha256=jDCEzBFDSQtq7nQxRScIsbALnFcvMPOkNkMUCa7mFxg,31921
@@ -52,7 +52,7 @@ datamule/sec/infrastructure/submissions_metadata.py,sha256=f1KarzFSryKm0EV8DCDNs
52
52
  datamule/sec/submissions/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
53
53
  datamule/sec/submissions/downloader.py,sha256=zGS0oJJI8tVF_GnVpZm20MymdYxnjrEjQioSVggw7Ck,1486
54
54
  datamule/sec/submissions/eftsquery.py,sha256=mSZon8rlW8dxma7M49ZW5V02Fn-ENOdt9TNO6elBrhE,27983
55
- datamule/sec/submissions/monitor.py,sha256=1JUMRYsTqtd31hX3UrUA_aXFUmZN6n-V7h0i1gavNOs,11395
55
+ datamule/sec/submissions/monitor.py,sha256=6mE0NZFdPId69t4V53GwBb9sqtRN7HE54sU3WpU0bnY,11900
56
56
  datamule/sec/submissions/streamer.py,sha256=Qydj40CmWB_wsPv2dibefRohmCokegG2pR7iZ9C3xLQ,11584
57
57
  datamule/sec/submissions/textsearch.py,sha256=MKDXEz_VI_0ljl73_aw2lx4MVzJW5uDt8KxjvJBwPwM,5794
58
58
  datamule/sec/xbrl/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
@@ -66,7 +66,7 @@ datamule/seclibrary/downloader.py,sha256=3jEy67oiEg8BF20KcKCx2KC0UjHzhiepdu29TOa
66
66
  datamule/seclibrary/query.py,sha256=qGuursTERRbOGfoDcYcpo4oWkW3PCBW6x1Qf1Puiak4,7352
67
67
  datamule/utils/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
68
68
  datamule/utils/construct_submissions_data.py,sha256=aX7ZaAp3zXHLcv4TFk_rGwjb8r7yNDQDFVg4nPf60kM,5934
69
- datamule-1.6.4.dist-info/METADATA,sha256=4v85X90MyUpvQC37rMyiEA_3UA-2scIHz4tTF1xH2e4,524
70
- datamule-1.6.4.dist-info/WHEEL,sha256=oiQVh_5PnQM0E3gPdiz09WCNmwiHDMaGer_elqB3coM,92
71
- datamule-1.6.4.dist-info/top_level.txt,sha256=iOfgmtSMFVyr7JGl_bYSTDry79JbmsG4p8zKq89ktKk,9
72
- datamule-1.6.4.dist-info/RECORD,,
69
+ datamule-1.6.6.dist-info/METADATA,sha256=qpBESqH9BTqVYWawc44WBq5ki5rUF2XO5Gh4h7Olx8o,524
70
+ datamule-1.6.6.dist-info/WHEEL,sha256=oiQVh_5PnQM0E3gPdiz09WCNmwiHDMaGer_elqB3coM,92
71
+ datamule-1.6.6.dist-info/top_level.txt,sha256=iOfgmtSMFVyr7JGl_bYSTDry79JbmsG4p8zKq89ktKk,9
72
+ datamule-1.6.6.dist-info/RECORD,,