datamule 2.2.6__tar.gz → 2.2.8__tar.gz

{datamule-2.2.6 → datamule-2.2.8}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: datamule
-Version: 2.2.6
+Version: 2.2.8
 Summary: Work with SEC submissions at scale.
 Home-page: https://github.com/john-friedman/datamule-python
 Author: John Friedman

{datamule-2.2.6 → datamule-2.2.8}/datamule/document/document.py

@@ -16,6 +16,7 @@ import tempfile
 from .tables.tables import Tables
 
 from ..tags.utils import get_cusip_using_regex, get_isin_using_regex, get_figi_using_regex,get_all_tickers, get_full_names,get_full_names_dictionary_lookup, analyze_lm_sentiment_fragment
+from ..utils.pdf import has_extractable_text
 
 class DataWithTags(dict):
     def __init__(self, data, document):
@@ -113,29 +114,9 @@ class TextAnalysisBase:
                 # Original behavior - single text source
                 self._text_sources = [{'id': None, 'text': str(self.document.text)}]
             else:  # mode == 'data'
-                # New behavior - multiple text fragments
-                self._text_sources = []
-                self._extract_text_fragments(self.document.data, '')
+                self._text_sources = [{'id':data_tuple[0],'text':data_tuple[2]} for data_tuple in self.document.data_tuples if data_tuple[1] in ['text','title','textsmall']]
         return self._text_sources
     
-    def _extract_text_fragments(self, data, parent_id=''):
-        """Extract all text fragments with their document IDs from parsed data"""
-        if isinstance(data, dict):
-            for key, value in data.items():
-                if key in ["text", "title"] and isinstance(value, str):
-                    # Use the current dictionary's parent key as the fragment ID
-                    self._text_sources.append({
-                        'id': parent_id,
-                        'text': value
-                    })
-                elif isinstance(value, (dict, list)):
-                    # Pass the current key as the parent_id for the next level
-                    self._extract_text_fragments(value, key)
-        elif isinstance(data, list):
-            for i, item in enumerate(data):
-                if isinstance(item, (dict, list)):
-                    self._extract_text_fragments(item, parent_id)
-    
     def _format_results(self, results, fragment_id):
         """Format results based on mode"""
         if self.mode == 'text':
@@ -286,12 +267,20 @@ class Document:
 
         # this will be filled by parsed
         self._data = None
+        self._data_tuples = None
         self._tables = None
         self._text = None
         self._markdown = None
 
         # booleans
         self._data_bool = self.extension in ('.htm', '.html','.txt')
+
+        # may slow things down?
+        if self.extension == '.pdf':
+            if has_extractable_text(pdf_bytes=self.content):
+                self._data_bool = True
+
+        self._data_tuples_bool = self._data_bool
         self._text_bool = self._data_bool
         self._markdown_bool = self._data_bool
         self._visualize_bool = self._data_bool
@@ -417,22 +406,30 @@ class Document:
 
     @property
     def data(self):
-        if self._data is None:
-            self.parse()
+        if self._data_bool:
+            if self._data is None:
+                self.parse()
 
-        if self._data is None:
-            self._data = {}
-        
-        if not isinstance(self._data, DataWithTags):
-            self._data = DataWithTags(self._data, self)
+            if self._data is None:
+                self._data = {}
+            
+            if not isinstance(self._data, DataWithTags):
+                self._data = DataWithTags(self._data, self)
             
         return self._data
     
+    @property
+    def data_tuples(self):
+        if self._data_bool:
+            if self._data_tuples is None:
+                self._data_tuples = unnest_dict(self.data)
+        return self._data_tuples
+    
     @property
     def text(self):
         if self._text_bool:
             if self._text is None:
-                text = flatten_dict(self.data,'text')
+                text = flatten_dict(tuples_list=self.data_tuples,format='text')
                 self._text = TextWithTags(text, self)
         return self._text
     
@@ -440,7 +437,7 @@ class Document:
     def markdown(self):
         if self._markdown_bool:
             if self._markdown is None:
-                self._markdown = flatten_dict(self.data,'markdown')
+                self._markdown = flatten_dict(tuples_list=self.data_tuples,format='markdown')
 
         return self._markdown
 
@@ -556,18 +553,16 @@ class Document:
             webbrowser.open('file://' + temp_path)
         else:
             print(f"Cannot open files with extension {self.extension}")
-
     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=title,title_regex=title_regex,title_class=title_class)
+        if self._data_bool:
+            if not self.data:
+                self.parse()
 
-        if format == 'text':
-            result = [item[1] for item in result]
-            result = [unnest_dict(item) for item in result]
-
-        return result
+            result = get_title(self.data,title=title,title_regex=title_regex,title_class=title_class)
+            if format == 'dict':
+                return [item[1] for item in result]
+            else:
+                return [flatten_dict(item[1],format) for item in result]
 
    
    # TODO CHANGE THIS

datamule-2.2.8/datamule/utils/pdf.py

@@ -0,0 +1,25 @@
+def has_extractable_text(pdf_bytes, search_range=50000):
+    """
+    Check if PDF contains extractable text within first N bytes
+    Returns True if found in range, False otherwise
+    
+    Args:
+        pdf_bytes: PDF content as bytes
+        search_range: Number of bytes to search from start (default 50KB)
+    """
+    # Text indicators to search for
+    indicators = [
+        b'BT',  # Begin text - most common
+        b'Tj',  # Show text
+        b'TJ',  # Show text with positioning
+        b'Tf',  # Set font
+    ]
+    
+    # Search only within the specified range
+    search_data = pdf_bytes[:search_range]
+    
+    for indicator in indicators:
+        if indicator in search_data:
+            return True
+    
+    return False

{datamule-2.2.6 → datamule-2.2.8}/datamule.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: datamule
-Version: 2.2.6
+Version: 2.2.8
 Summary: Work with SEC submissions at scale.
 Home-page: https://github.com/john-friedman/datamule-python
 Author: John Friedman

{datamule-2.2.6 → datamule-2.2.8}/datamule.egg-info/SOURCES.txt

@@ -62,4 +62,5 @@ datamule/tags/regex.py
 datamule/tags/utils.py
 datamule/utils/__init__.py
 datamule/utils/construct_submissions_data.py
-datamule/utils/format_accession.py
+datamule/utils/format_accession.py
+datamule/utils/pdf.py

{datamule-2.2.6 → datamule-2.2.8}/setup.py

@@ -32,7 +32,7 @@ if not os.path.exists(file_path):
 setup(
     name="datamule",
     author="John Friedman",
-    version="2.2.6",
+    version="2.2.8",
     description="Work with SEC submissions at scale.",
     packages=find_packages(include=['datamule', 'datamule.*']),
     url="https://github.com/john-friedman/datamule-python",