datamule 2.1.0__tar.gz → 2.1.2__tar.gz

{datamule-2.1.0 → datamule-2.1.2}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: datamule
-Version: 2.1.0
+Version: 2.1.2
 Summary: Work with SEC submissions at scale.
 Home-page: https://github.com/john-friedman/datamule-python
 Author: John Friedman

{datamule-2.1.0 → datamule-2.1.2}/datamule/datamule/datamule_mysql_rds.py

@@ -71,6 +71,9 @@ class DatamuleMySQL:
             database = 'npx_db'
         elif table == 'information_table':
             database = '13fhr_db'
+        elif table in ['derivative_holding_ownership','derivative_transaction_ownership', 'metadata_ownership','non_derivative_holding_ownership',
+                       'non_derivative_transaction_ownership','owner_signature_ownership','reporting_owner_ownership']:
+            database = 'ownership_db'
         else:
             raise ValueError(f"Unsupported table: {table}")
         

{datamule-2.1.0 → datamule-2.1.2}/datamule/datamule/downloader.py

@@ -23,7 +23,11 @@ from ..utils.format_accession import format_accession
 # could be cleaned up
 
 # Set up logging
-logging.basicConfig(level=logging.INFO, format='%(asctime)s - %(levelname)s - %(message)s')
+logging.basicConfig(
+    level=logging.INFO,
+    format='%(asctime)s - %(levelname)s - %(message)s',
+    handlers=logging.getLogger().handlers,
+)
 logger = logging.getLogger(__name__)
 
 

{datamule-2.1.0 → datamule-2.1.2}/datamule/document/document.py

@@ -315,11 +315,11 @@ class Document:
         else:
             visualize_dict(self.data)
 
-    def get_section(self, title, format='dict'):
+    def get_section(self, title=None, title_regex=None,title_class=None, format='dict'):
         if not self.data:
             self.parse()
 
-        result = get_title(self.data,title)
+        result = get_title(self.data,title=title,title_regex=title_regex,title_class=title_class)
 
         if format == 'text':
             result = [item[1] for item in result]

{datamule-2.1.0 → datamule-2.1.2}/datamule/mapping_dicts/html_mapping_dicts.py

@@ -1,7 +1,7 @@
 dict_10k_html = {
     ('part',r'^part\s*([ivx]+)$') : 0,
     ('signatures',r'^signatures?\.*$') : 0,
-    ('item',r'^item\s*(\d+)\.?([a-z])?') : 1,
+    ('item',r'^item\s*(\d+)\.?([a-z])?(?![a-z])') : 1,
 }
 dict_10q_html = dict_10k_html
 
@@ -48,7 +48,7 @@ dict_10d_html = dict_10k_html
 
 dict_20f_html = {
     ('part',r'^part\s*([ivx]+)') : 0,
-    ('item',r'^item\s*(\d+)\.?([a-z])?') : 1,
+    ('item',r'^item\s*(\d+)\.?([a-z])?(?![a-z])') : 1,
     ('letter',r'\d*\.?([a-z])') : 2,
     ('signatures',r'^signatures?\.*$') : 0,
 }

{datamule-2.1.0 → datamule-2.1.2}/datamule/submission.py

@@ -163,8 +163,8 @@ class Submission:
                     content = zstd.ZstdDecompressor().decompress(content)
                 
                 # Decode text files
-                if extension in ['.htm', '.html', '.txt', '.xml']:
-                    content = content.decode('utf-8', errors='replace')
+                # if extension in ['.htm', '.html', '.txt', '.xml']:
+                #     content = content.decode('utf-8', errors='replace')
                 
                 document_path = f"{self.batch_tar_path}::{self.accession_prefix}/{filename}"
         
@@ -197,8 +197,8 @@ class Submission:
                         content = zstd.ZstdDecompressor().decompress(content)
                     
                     # Decode text files
-                    if extension in ['.htm', '.html', '.txt', '.xml']:
-                        content = content.decode('utf-8', errors='replace')
+                    # if extension in ['.htm', '.html', '.txt', '.xml']:
+                    #     content = content.decode('utf-8', errors='replace')
                     
                     document_path = f"{self.path}::{actual_filename}"
             
@@ -219,8 +219,8 @@ class Submission:
                     content = zstd.ZstdDecompressor().decompress(content)
                 
                 # Decode text files
-                if extension in ['.htm', '.html', '.txt', '.xml']:
-                    content = content.decode('utf-8', errors='replace')
+                # if extension in ['.htm', '.html', '.txt', '.xml']:
+                #     content = content.decode('utf-8', errors='replace')
 
         return Document(
             type=doc['type'], 

{datamule-2.1.0 → datamule-2.1.2}/datamule.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: datamule
-Version: 2.1.0
+Version: 2.1.2
 Summary: Work with SEC submissions at scale.
 Home-page: https://github.com/john-friedman/datamule-python
 Author: John Friedman

{datamule-2.1.0 → datamule-2.1.2}/setup.py

@@ -32,7 +32,7 @@ if not os.path.exists(file_path):
 setup(
     name="datamule",
     author="John Friedman",
-    version="2.1.0",
+    version="2.1.2",
     description="Work with SEC submissions at scale.",
     packages=find_packages(include=['datamule', 'datamule.*']),
     url="https://github.com/john-friedman/datamule-python",