datamule 1.4.0__py3-none-any.whl → 1.4.2__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.
@@ -3,7 +3,7 @@ import csv
3
3
  import re
4
4
  from doc2dict import xml2dict, txt2dict, dict2dict
5
5
  from doc2dict.mapping import flatten_hierarchy
6
- from doc2dict import html2dict, visualize_dict, get_title, unnest_dict
6
+ from doc2dict import html2dict, visualize_dict, get_title, unnest_dict, pdf2dict
7
7
  from ..mapping_dicts.txt_mapping_dicts import dict_10k, dict_10q, dict_8k, dict_13d, dict_13g
8
8
  from ..mapping_dicts.xml_mapping_dicts import dict_345
9
9
  from ..mapping_dicts.html_mapping_dicts import dict_10k_html, dict_10q_html, dict_8k_html
@@ -134,6 +134,8 @@ class Document:
134
134
  mapping_dict = dict_345
135
135
 
136
136
  self.data = xml2dict(content=self.content, mapping_dict=mapping_dict)
137
+ elif self.extension == '.pdf':
138
+ self.data = pdf2dict(content=self.content, mapping_dict=mapping_dict)
137
139
  else:
138
140
  pass
139
141
 
@@ -235,8 +237,7 @@ class Document:
235
237
  return result
236
238
 
237
239
 
238
- # this will all have to be changed. default will be to flatten everything
239
- # candidate for deletion
240
+ # TODO CHANGE THIS
240
241
  def __iter__(self):
241
242
  self.parse()
242
243
 
@@ -1,6 +1,6 @@
1
1
  Metadata-Version: 2.1
2
2
  Name: datamule
3
- Version: 1.4.0
3
+ Version: 1.4.2
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
@@ -8,7 +8,7 @@ datamule/sheet.py,sha256=TvFqK9eAYuVoJ2uWdAlx5EN6vS9lke-aZf7FqtUiDBc,22304
8
8
  datamule/submission.py,sha256=Yh5nG3ioumhl6z30wJdIEmKjDDNSuo0r2xycZSIaeIg,11035
9
9
  datamule/data/listed_filer_metadata.csv,sha256=dT9fQ8AC5P1-Udf_UF0ZkdXJ88jNxJb_tuhi5YYL1rc,2426827
10
10
  datamule/document/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
11
- datamule/document/document.py,sha256=fQ7woCklxFwzABvjdFknFR8AFUwPuEEfuXZ_VE5q-7Y,10954
11
+ datamule/document/document.py,sha256=uHpkLyl6er4MdZDDJNzxSdtwkfDMtMVeG01uRCRblfA,11002
12
12
  datamule/document/processing.py,sha256=jDCEzBFDSQtq7nQxRScIsbALnFcvMPOkNkMUCa7mFxg,31921
13
13
  datamule/document/table.py,sha256=73yUJKY82ap32jhLmZeTti-jQ_lyhcJGlGwyxLtgYOg,12944
14
14
  datamule/document/mappings/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
@@ -58,7 +58,7 @@ datamule/seclibrary/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSu
58
58
  datamule/seclibrary/bq.py,sha256=C8sb_rpXTvchprrFLcbRar4Qi0XWW25tnv1YsHSS5o4,18025
59
59
  datamule/seclibrary/downloader.py,sha256=PIgz_7ASUTZOHcUZGcD1SmLaGSbq7xe7EiJT0Z7HU4M,13653
60
60
  datamule/seclibrary/query.py,sha256=qGuursTERRbOGfoDcYcpo4oWkW3PCBW6x1Qf1Puiak4,7352
61
- datamule-1.4.0.dist-info/METADATA,sha256=6H4Zzj37KnA56z-_XNAEv2NNpW3qKiEV_sul921KshM,469
62
- datamule-1.4.0.dist-info/WHEEL,sha256=oiQVh_5PnQM0E3gPdiz09WCNmwiHDMaGer_elqB3coM,92
63
- datamule-1.4.0.dist-info/top_level.txt,sha256=iOfgmtSMFVyr7JGl_bYSTDry79JbmsG4p8zKq89ktKk,9
64
- datamule-1.4.0.dist-info/RECORD,,
61
+ datamule-1.4.2.dist-info/METADATA,sha256=mUujGnhNFWQFCH8qg2aJ4_bS4Zp-XhEkD_Y6x9qAh9Q,469
62
+ datamule-1.4.2.dist-info/WHEEL,sha256=oiQVh_5PnQM0E3gPdiz09WCNmwiHDMaGer_elqB3coM,92
63
+ datamule-1.4.2.dist-info/top_level.txt,sha256=iOfgmtSMFVyr7JGl_bYSTDry79JbmsG4p8zKq89ktKk,9
64
+ datamule-1.4.2.dist-info/RECORD,,