fmtr.tools 1.0.33__tar.gz → 1.0.34__tar.gz
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.
Potentially problematic release.
This version of fmtr.tools might be problematic. Click here for more details.
- {fmtr.tools-1.0.33 → fmtr.tools-1.0.34}/PKG-INFO +1 -1
- {fmtr.tools-1.0.33 → fmtr.tools-1.0.34}/fmtr/tools/pdf_tools.py +3 -8
- fmtr.tools-1.0.34/fmtr/tools/version +1 -0
- {fmtr.tools-1.0.33 → fmtr.tools-1.0.34}/fmtr.tools.egg-info/PKG-INFO +1 -1
- fmtr.tools-1.0.33/fmtr/tools/version +0 -1
- {fmtr.tools-1.0.33 → fmtr.tools-1.0.34}/LICENSE +0 -0
- {fmtr.tools-1.0.33 → fmtr.tools-1.0.34}/README.md +0 -0
- {fmtr.tools-1.0.33 → fmtr.tools-1.0.34}/fmtr/tools/__init__.py +0 -0
- {fmtr.tools-1.0.33 → fmtr.tools-1.0.34}/fmtr/tools/ai_tools.py +0 -0
- {fmtr.tools-1.0.33 → fmtr.tools-1.0.34}/fmtr/tools/api_tools.py +0 -0
- {fmtr.tools-1.0.33 → fmtr.tools-1.0.34}/fmtr/tools/async_tools.py +0 -0
- {fmtr.tools-1.0.33 → fmtr.tools-1.0.34}/fmtr/tools/augmentation_tools.py +0 -0
- {fmtr.tools-1.0.33 → fmtr.tools-1.0.34}/fmtr/tools/caching_tools.py +0 -0
- {fmtr.tools-1.0.33 → fmtr.tools-1.0.34}/fmtr/tools/config.py +0 -0
- {fmtr.tools-1.0.33 → fmtr.tools-1.0.34}/fmtr/tools/config_tools.py +0 -0
- {fmtr.tools-1.0.33 → fmtr.tools-1.0.34}/fmtr/tools/console_script_tools.py +0 -0
- {fmtr.tools-1.0.33 → fmtr.tools-1.0.34}/fmtr/tools/data_modelling_tools.py +0 -0
- {fmtr.tools-1.0.33 → fmtr.tools-1.0.34}/fmtr/tools/dataclass_tools.py +0 -0
- {fmtr.tools-1.0.33 → fmtr.tools-1.0.34}/fmtr/tools/datatype_tools.py +0 -0
- {fmtr.tools-1.0.33 → fmtr.tools-1.0.34}/fmtr/tools/docker_tools.py +0 -0
- {fmtr.tools-1.0.33 → fmtr.tools-1.0.34}/fmtr/tools/environment_tools.py +0 -0
- {fmtr.tools-1.0.33 → fmtr.tools-1.0.34}/fmtr/tools/function_tools.py +0 -0
- {fmtr.tools-1.0.33 → fmtr.tools-1.0.34}/fmtr/tools/google_api_tools.py +0 -0
- {fmtr.tools-1.0.33 → fmtr.tools-1.0.34}/fmtr/tools/hash_tools.py +0 -0
- {fmtr.tools-1.0.33 → fmtr.tools-1.0.34}/fmtr/tools/hfh_tools.py +0 -0
- {fmtr.tools-1.0.33 → fmtr.tools-1.0.34}/fmtr/tools/html_tools.py +0 -0
- {fmtr.tools-1.0.33 → fmtr.tools-1.0.34}/fmtr/tools/import_tools.py +0 -0
- {fmtr.tools-1.0.33 → fmtr.tools-1.0.34}/fmtr/tools/inspection_tools.py +0 -0
- {fmtr.tools-1.0.33 → fmtr.tools-1.0.34}/fmtr/tools/interface_tools.py +0 -0
- {fmtr.tools-1.0.33 → fmtr.tools-1.0.34}/fmtr/tools/iterator_tools.py +0 -0
- {fmtr.tools-1.0.33 → fmtr.tools-1.0.34}/fmtr/tools/json_fix_tools.py +0 -0
- {fmtr.tools-1.0.33 → fmtr.tools-1.0.34}/fmtr/tools/json_tools.py +0 -0
- {fmtr.tools-1.0.33 → fmtr.tools-1.0.34}/fmtr/tools/logging_tools.py +0 -0
- {fmtr.tools-1.0.33 → fmtr.tools-1.0.34}/fmtr/tools/merging_tools.py +0 -0
- {fmtr.tools-1.0.33 → fmtr.tools-1.0.34}/fmtr/tools/metric_tools.py +0 -0
- {fmtr.tools-1.0.33 → fmtr.tools-1.0.34}/fmtr/tools/name_tools.py +0 -0
- {fmtr.tools-1.0.33 → fmtr.tools-1.0.34}/fmtr/tools/netrc_tools.py +0 -0
- {fmtr.tools-1.0.33 → fmtr.tools-1.0.34}/fmtr/tools/openai_tools.py +0 -0
- {fmtr.tools-1.0.33 → fmtr.tools-1.0.34}/fmtr/tools/parallel_tools.py +0 -0
- {fmtr.tools-1.0.33 → fmtr.tools-1.0.34}/fmtr/tools/path_tools.py +0 -0
- {fmtr.tools-1.0.33 → fmtr.tools-1.0.34}/fmtr/tools/platform_tools.py +0 -0
- {fmtr.tools-1.0.33 → fmtr.tools-1.0.34}/fmtr/tools/process_tools.py +0 -0
- {fmtr.tools-1.0.33 → fmtr.tools-1.0.34}/fmtr/tools/profiling_tools.py +0 -0
- {fmtr.tools-1.0.33 → fmtr.tools-1.0.34}/fmtr/tools/random_tools.py +0 -0
- {fmtr.tools-1.0.33 → fmtr.tools-1.0.34}/fmtr/tools/semantic_tools.py +0 -0
- {fmtr.tools-1.0.33 → fmtr.tools-1.0.34}/fmtr/tools/spaces_tools.py +0 -0
- {fmtr.tools-1.0.33 → fmtr.tools-1.0.34}/fmtr/tools/string_tools.py +0 -0
- {fmtr.tools-1.0.33 → fmtr.tools-1.0.34}/fmtr/tools/tests/__init__.py +0 -0
- {fmtr.tools-1.0.33 → fmtr.tools-1.0.34}/fmtr/tools/tests/conftest.py +0 -0
- {fmtr.tools-1.0.33 → fmtr.tools-1.0.34}/fmtr/tools/tests/helpers.py +0 -0
- {fmtr.tools-1.0.33 → fmtr.tools-1.0.34}/fmtr/tools/tests/test_datatype.py +0 -0
- {fmtr.tools-1.0.33 → fmtr.tools-1.0.34}/fmtr/tools/tests/test_environment.py +0 -0
- {fmtr.tools-1.0.33 → fmtr.tools-1.0.34}/fmtr/tools/tests/test_json.py +0 -0
- {fmtr.tools-1.0.33 → fmtr.tools-1.0.34}/fmtr/tools/tests/test_path.py +0 -0
- {fmtr.tools-1.0.33 → fmtr.tools-1.0.34}/fmtr/tools/tests/test_yaml.py +0 -0
- {fmtr.tools-1.0.33 → fmtr.tools-1.0.34}/fmtr/tools/tokenization_tools.py +0 -0
- {fmtr.tools-1.0.33 → fmtr.tools-1.0.34}/fmtr/tools/tools.py +0 -0
- {fmtr.tools-1.0.33 → fmtr.tools-1.0.34}/fmtr/tools/unicode_tools.py +0 -0
- {fmtr.tools-1.0.33 → fmtr.tools-1.0.34}/fmtr/tools/version_tools.py +0 -0
- {fmtr.tools-1.0.33 → fmtr.tools-1.0.34}/fmtr/tools/yaml_tools.py +0 -0
- {fmtr.tools-1.0.33 → fmtr.tools-1.0.34}/fmtr.tools.egg-info/SOURCES.txt +0 -0
- {fmtr.tools-1.0.33 → fmtr.tools-1.0.34}/fmtr.tools.egg-info/dependency_links.txt +0 -0
- {fmtr.tools-1.0.33 → fmtr.tools-1.0.34}/fmtr.tools.egg-info/entry_points.txt +0 -0
- {fmtr.tools-1.0.33 → fmtr.tools-1.0.34}/fmtr.tools.egg-info/requires.txt +0 -0
- {fmtr.tools-1.0.33 → fmtr.tools-1.0.34}/fmtr.tools.egg-info/top_level.txt +0 -0
- {fmtr.tools-1.0.33 → fmtr.tools-1.0.34}/setup.cfg +0 -0
- {fmtr.tools-1.0.33 → fmtr.tools-1.0.34}/setup.py +0 -0
|
@@ -97,7 +97,7 @@ class Block(data_modelling_tools.Base):
|
|
|
97
97
|
Simple text representation
|
|
98
98
|
|
|
99
99
|
"""
|
|
100
|
-
return ' '.join([
|
|
100
|
+
return ' '.join([line.text for line in self.lines])
|
|
101
101
|
|
|
102
102
|
@classmethod
|
|
103
103
|
def from_dict(cls, data: Dict) -> Self:
|
|
@@ -138,7 +138,7 @@ class Document(pm.Document):
|
|
|
138
138
|
blocks = []
|
|
139
139
|
|
|
140
140
|
for page in self:
|
|
141
|
-
for block in page.get_text("dict")["blocks"]:
|
|
141
|
+
for block in page.get_text("dict", flags=pm.TEXTFLAGS_TEXT | pm.TEXT_ACCURATE_BBOXES)["blocks"]:
|
|
142
142
|
obj = Block.from_dict(block)
|
|
143
143
|
blocks.append(obj)
|
|
144
144
|
|
|
@@ -162,11 +162,6 @@ if __name__ == '__main__':
|
|
|
162
162
|
assert PATH_PDF.exists()
|
|
163
163
|
|
|
164
164
|
doc = Document(PATH_PDF)
|
|
165
|
-
doc.data
|
|
166
|
-
|
|
167
|
-
for page in doc:
|
|
168
|
-
print(page.get_text('dict'))
|
|
169
|
-
print(page.get_text('html'))
|
|
170
|
-
|
|
165
|
+
data = doc.data
|
|
171
166
|
md = doc.to_markdown()
|
|
172
167
|
md
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
1.0.34
|
|
@@ -1 +0,0 @@
|
|
|
1
|
-
1.0.33
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|