fmtr.tools 1.3.61__tar.gz → 1.3.62__tar.gz

{fmtr_tools-1.3.61 → fmtr_tools-1.3.62}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: fmtr.tools
-Version: 1.3.61
+Version: 1.3.62
 Summary: Collection of high-level tools to simplify everyday development tasks, with a focus on AI/ML
 Home-page: https://github.com/fmtr/fmtr.tools
 Author: Frontmatter
@@ -154,68 +154,68 @@ Provides-Extra: db-document
 Requires-Dist: beanie[odm]; extra == "db-document"
 Requires-Dist: motor; extra == "db-document"
 Provides-Extra: all
+Requires-Dist: setuptools; extra == "all"
+Requires-Dist: google-auth; extra == "all"
+Requires-Dist: google-auth-oauthlib; extra == "all"
+Requires-Dist: cachetools; extra == "all"
+Requires-Dist: uvicorn[standard]; extra == "all"
+Requires-Dist: pydantic; extra == "all"
+Requires-Dist: openai; extra == "all"
+Requires-Dist: ollama; extra == "all"
+Requires-Dist: json_repair; extra == "all"
+Requires-Dist: sentence_transformers; extra == "all"
 Requires-Dist: tabulate; extra == "all"
-Requires-Dist: pydantic-extra-types; extra == "all"
-Requires-Dist: google-auth-httplib2; extra == "all"
-Requires-Dist: filetype; extra == "all"
-Requires-Dist: odfpy; extra == "all"
-Requires-Dist: logfire[fastapi]; extra == "all"
-Requires-Dist: torchaudio; extra == "all"
-Requires-Dist: python-on-whales; extra == "all"
-Requires-Dist: torchvision; extra == "all"
-Requires-Dist: pydantic-settings; extra == "all"
-Requires-Dist: logfire[httpx]; extra == "all"
-Requires-Dist: dnspython[doh]; extra == "all"
-Requires-Dist: deepmerge; extra == "all"
+Requires-Dist: sre_yield; extra == "all"
+Requires-Dist: google-api-python-client; extra == "all"
 Requires-Dist: pyyaml; extra == "all"
-Requires-Dist: pydantic-ai[logfire,openai]; extra == "all"
-Requires-Dist: pydevd-pycharm~=251.25410.159; extra == "all"
+Requires-Dist: regex; extra == "all"
+Requires-Dist: httpx_retries; extra == "all"
+Requires-Dist: flet-video; extra == "all"
 Requires-Dist: pandas; extra == "all"
-Requires-Dist: beanie[odm]; extra == "all"
-Requires-Dist: sentence_transformers; extra == "all"
-Requires-Dist: playwright; extra == "all"
-Requires-Dist: dask[bag]; extra == "all"
+Requires-Dist: openpyxl; extra == "all"
+Requires-Dist: google-auth-httplib2; extra == "all"
+Requires-Dist: semver; extra == "all"
 Requires-Dist: yamlscript; extra == "all"
-Requires-Dist: Unidecode; extra == "all"
-Requires-Dist: pytest-cov; extra == "all"
-Requires-Dist: httpx; extra == "all"
+Requires-Dist: dask[bag]; extra == "all"
+Requires-Dist: peft; extra == "all"
 Requires-Dist: html2text; extra == "all"
 Requires-Dist: transformers[sentencepiece]; extra == "all"
-Requires-Dist: regex; extra == "all"
-Requires-Dist: tokenizers; extra == "all"
-Requires-Dist: pydantic; extra == "all"
-Requires-Dist: pycountry; extra == "all"
+Requires-Dist: tinynetrc; extra == "all"
+Requires-Dist: flet[all]; extra == "all"
+Requires-Dist: appdirs; extra == "all"
+Requires-Dist: deepmerge; extra == "all"
+Requires-Dist: httpx; extra == "all"
 Requires-Dist: huggingface_hub; extra == "all"
-Requires-Dist: contexttimer; extra == "all"
-Requires-Dist: bokeh; extra == "all"
-Requires-Dist: semver; extra == "all"
-Requires-Dist: flet-webview; extra == "all"
-Requires-Dist: logfire; extra == "all"
-Requires-Dist: sre_yield; extra == "all"
+Requires-Dist: pymupdf; extra == "all"
+Requires-Dist: pytest-cov; extra == "all"
+Requires-Dist: beanie[odm]; extra == "all"
+Requires-Dist: logfire[httpx]; extra == "all"
 Requires-Dist: deepdiff; extra == "all"
-Requires-Dist: json_repair; extra == "all"
-Requires-Dist: cachetools; extra == "all"
-Requires-Dist: distributed; extra == "all"
-Requires-Dist: pymupdf4llm; extra == "all"
-Requires-Dist: httpx_retries; extra == "all"
+Requires-Dist: bokeh; extra == "all"
+Requires-Dist: pycountry; extra == "all"
+Requires-Dist: dnspython[doh]; extra == "all"
 Requires-Dist: faker; extra == "all"
-Requires-Dist: flet[all]; extra == "all"
+Requires-Dist: pydantic-ai[logfire,openai]; extra == "all"
+Requires-Dist: odfpy; extra == "all"
+Requires-Dist: pydantic-settings; extra == "all"
+Requires-Dist: pydantic-extra-types; extra == "all"
+Requires-Dist: torchaudio; extra == "all"
 Requires-Dist: motor; extra == "all"
-Requires-Dist: uvicorn[standard]; extra == "all"
-Requires-Dist: diskcache; extra == "all"
-Requires-Dist: appdirs; extra == "all"
-Requires-Dist: openai; extra == "all"
-Requires-Dist: google-auth-oauthlib; extra == "all"
-Requires-Dist: setuptools; extra == "all"
-Requires-Dist: flet-video; extra == "all"
-Requires-Dist: google-api-python-client; extra == "all"
-Requires-Dist: ollama; extra == "all"
-Requires-Dist: openpyxl; extra == "all"
-Requires-Dist: google-auth; extra == "all"
-Requires-Dist: tinynetrc; extra == "all"
-Requires-Dist: pymupdf; extra == "all"
-Requires-Dist: peft; extra == "all"
+Requires-Dist: torchvision; extra == "all"
+Requires-Dist: filetype; extra == "all"
+Requires-Dist: python-on-whales; extra == "all"
+Requires-Dist: pymupdf4llm; extra == "all"
+Requires-Dist: logfire; extra == "all"
+Requires-Dist: pydevd-pycharm~=251.25410.159; extra == "all"
+Requires-Dist: logfire[fastapi]; extra == "all"
+Requires-Dist: contexttimer; extra == "all"
 Requires-Dist: fastapi; extra == "all"
+Requires-Dist: flet-webview; extra == "all"
+Requires-Dist: playwright; extra == "all"
+Requires-Dist: Unidecode; extra == "all"
+Requires-Dist: distributed; extra == "all"
+Requires-Dist: diskcache; extra == "all"
+Requires-Dist: tokenizers; extra == "all"
 Dynamic: author
 Dynamic: author-email
 Dynamic: description

{fmtr_tools-1.3.61 → fmtr_tools-1.3.62}/fmtr/tools/pdf_tools.py

@@ -1,6 +1,7 @@
+from typing import List, Tuple, Dict, Any, Self
+
 import pymupdf as pm
 import pymupdf4llm
-from typing import List, Tuple, Dict, Any, Self
 
 from fmtr.tools import data_modelling_tools
 
@@ -179,10 +180,10 @@ class Document(pm.Document):
         """
         return pymupdf4llm.to_markdown(self, **kwargs)
 
-    def to_text(self):
+    def to_text_pages(self) -> List[str]:
         """
 
-        Simple text output.
+        Simple text output per-page.
 
         """
         lines = []
@@ -190,9 +191,32 @@ class Document(pm.Document):
             text = page.get_text()
             lines.append(text)
 
-        text = '\n'.join(lines)
+        return lines
+
+    def to_text(self) -> str:
+        """
+
+        Simple text output.
+
+        """
+
+        text = '\n'.join(self.to_text_pages())
         return text
 
+    def split(self) -> List[Self]:
+        """
+
+        Split pages into individual documents.
+
+        """
+
+        documents = []
+        for i, page in enumerate(self, start=1):
+            document = self.__class__()
+            document.insert_pdf(self, from_page=i, to_page=i)
+            documents.append(document)
+
+        return documents
 
 if __name__ == '__main__':
     from fmtr.tools.path_tools import Path

fmtr_tools-1.3.62/fmtr/tools/version

@@ -0,0 +1 @@
+1.3.62

{fmtr_tools-1.3.61 → fmtr_tools-1.3.62}/fmtr.tools.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: fmtr.tools
-Version: 1.3.61
+Version: 1.3.62
 Summary: Collection of high-level tools to simplify everyday development tasks, with a focus on AI/ML
 Home-page: https://github.com/fmtr/fmtr.tools
 Author: Frontmatter
@@ -154,68 +154,68 @@ Provides-Extra: db-document
 Requires-Dist: beanie[odm]; extra == "db-document"
 Requires-Dist: motor; extra == "db-document"
 Provides-Extra: all
+Requires-Dist: setuptools; extra == "all"
+Requires-Dist: google-auth; extra == "all"
+Requires-Dist: google-auth-oauthlib; extra == "all"
+Requires-Dist: cachetools; extra == "all"
+Requires-Dist: uvicorn[standard]; extra == "all"
+Requires-Dist: pydantic; extra == "all"
+Requires-Dist: openai; extra == "all"
+Requires-Dist: ollama; extra == "all"
+Requires-Dist: json_repair; extra == "all"
+Requires-Dist: sentence_transformers; extra == "all"
 Requires-Dist: tabulate; extra == "all"
-Requires-Dist: pydantic-extra-types; extra == "all"
-Requires-Dist: google-auth-httplib2; extra == "all"
-Requires-Dist: filetype; extra == "all"
-Requires-Dist: odfpy; extra == "all"
-Requires-Dist: logfire[fastapi]; extra == "all"
-Requires-Dist: torchaudio; extra == "all"
-Requires-Dist: python-on-whales; extra == "all"
-Requires-Dist: torchvision; extra == "all"
-Requires-Dist: pydantic-settings; extra == "all"
-Requires-Dist: logfire[httpx]; extra == "all"
-Requires-Dist: dnspython[doh]; extra == "all"
-Requires-Dist: deepmerge; extra == "all"
+Requires-Dist: sre_yield; extra == "all"
+Requires-Dist: google-api-python-client; extra == "all"
 Requires-Dist: pyyaml; extra == "all"
-Requires-Dist: pydantic-ai[logfire,openai]; extra == "all"
-Requires-Dist: pydevd-pycharm~=251.25410.159; extra == "all"
+Requires-Dist: regex; extra == "all"
+Requires-Dist: httpx_retries; extra == "all"
+Requires-Dist: flet-video; extra == "all"
 Requires-Dist: pandas; extra == "all"
-Requires-Dist: beanie[odm]; extra == "all"
-Requires-Dist: sentence_transformers; extra == "all"
-Requires-Dist: playwright; extra == "all"
-Requires-Dist: dask[bag]; extra == "all"
+Requires-Dist: openpyxl; extra == "all"
+Requires-Dist: google-auth-httplib2; extra == "all"
+Requires-Dist: semver; extra == "all"
 Requires-Dist: yamlscript; extra == "all"
-Requires-Dist: Unidecode; extra == "all"
-Requires-Dist: pytest-cov; extra == "all"
-Requires-Dist: httpx; extra == "all"
+Requires-Dist: dask[bag]; extra == "all"
+Requires-Dist: peft; extra == "all"
 Requires-Dist: html2text; extra == "all"
 Requires-Dist: transformers[sentencepiece]; extra == "all"
-Requires-Dist: regex; extra == "all"
-Requires-Dist: tokenizers; extra == "all"
-Requires-Dist: pydantic; extra == "all"
-Requires-Dist: pycountry; extra == "all"
+Requires-Dist: tinynetrc; extra == "all"
+Requires-Dist: flet[all]; extra == "all"
+Requires-Dist: appdirs; extra == "all"
+Requires-Dist: deepmerge; extra == "all"
+Requires-Dist: httpx; extra == "all"
 Requires-Dist: huggingface_hub; extra == "all"
-Requires-Dist: contexttimer; extra == "all"
-Requires-Dist: bokeh; extra == "all"
-Requires-Dist: semver; extra == "all"
-Requires-Dist: flet-webview; extra == "all"
-Requires-Dist: logfire; extra == "all"
-Requires-Dist: sre_yield; extra == "all"
+Requires-Dist: pymupdf; extra == "all"
+Requires-Dist: pytest-cov; extra == "all"
+Requires-Dist: beanie[odm]; extra == "all"
+Requires-Dist: logfire[httpx]; extra == "all"
 Requires-Dist: deepdiff; extra == "all"
-Requires-Dist: json_repair; extra == "all"
-Requires-Dist: cachetools; extra == "all"
-Requires-Dist: distributed; extra == "all"
-Requires-Dist: pymupdf4llm; extra == "all"
-Requires-Dist: httpx_retries; extra == "all"
+Requires-Dist: bokeh; extra == "all"
+Requires-Dist: pycountry; extra == "all"
+Requires-Dist: dnspython[doh]; extra == "all"
 Requires-Dist: faker; extra == "all"
-Requires-Dist: flet[all]; extra == "all"
+Requires-Dist: pydantic-ai[logfire,openai]; extra == "all"
+Requires-Dist: odfpy; extra == "all"
+Requires-Dist: pydantic-settings; extra == "all"
+Requires-Dist: pydantic-extra-types; extra == "all"
+Requires-Dist: torchaudio; extra == "all"
 Requires-Dist: motor; extra == "all"
-Requires-Dist: uvicorn[standard]; extra == "all"
-Requires-Dist: diskcache; extra == "all"
-Requires-Dist: appdirs; extra == "all"
-Requires-Dist: openai; extra == "all"
-Requires-Dist: google-auth-oauthlib; extra == "all"
-Requires-Dist: setuptools; extra == "all"
-Requires-Dist: flet-video; extra == "all"
-Requires-Dist: google-api-python-client; extra == "all"
-Requires-Dist: ollama; extra == "all"
-Requires-Dist: openpyxl; extra == "all"
-Requires-Dist: google-auth; extra == "all"
-Requires-Dist: tinynetrc; extra == "all"
-Requires-Dist: pymupdf; extra == "all"
-Requires-Dist: peft; extra == "all"
+Requires-Dist: torchvision; extra == "all"
+Requires-Dist: filetype; extra == "all"
+Requires-Dist: python-on-whales; extra == "all"
+Requires-Dist: pymupdf4llm; extra == "all"
+Requires-Dist: logfire; extra == "all"
+Requires-Dist: pydevd-pycharm~=251.25410.159; extra == "all"
+Requires-Dist: logfire[fastapi]; extra == "all"
+Requires-Dist: contexttimer; extra == "all"
 Requires-Dist: fastapi; extra == "all"
+Requires-Dist: flet-webview; extra == "all"
+Requires-Dist: playwright; extra == "all"
+Requires-Dist: Unidecode; extra == "all"
+Requires-Dist: distributed; extra == "all"
+Requires-Dist: diskcache; extra == "all"
+Requires-Dist: tokenizers; extra == "all"
 Dynamic: author
 Dynamic: author-email
 Dynamic: description

fmtr_tools-1.3.61/fmtr/tools/version

@@ -1 +0,0 @@
-1.3.61

{fmtr_tools-1.3.61 → fmtr_tools-1.3.62}/fmtr.tools.egg-info/requires.txt

@@ -18,68 +18,68 @@ pydantic-ai[logfire,openai]
 ollama
 
 [all]
+setuptools
+google-auth
+google-auth-oauthlib
+cachetools
+uvicorn[standard]
+pydantic
+openai
+ollama
+json_repair
+sentence_transformers
 tabulate
-pydantic-extra-types
-google-auth-httplib2
-filetype
-odfpy
-logfire[fastapi]
-torchaudio
-python-on-whales
-torchvision
-pydantic-settings
-logfire[httpx]
-dnspython[doh]
-deepmerge
+sre_yield
+google-api-python-client
 pyyaml
-pydantic-ai[logfire,openai]
-pydevd-pycharm~=251.25410.159
+regex
+httpx_retries
+flet-video
 pandas
-beanie[odm]
-sentence_transformers
-playwright
-dask[bag]
+openpyxl
+google-auth-httplib2
+semver
 yamlscript
-Unidecode
-pytest-cov
-httpx
+dask[bag]
+peft
 html2text
 transformers[sentencepiece]
-regex
-tokenizers
-pydantic
-pycountry
+tinynetrc
+flet[all]
+appdirs
+deepmerge
+httpx
 huggingface_hub
-contexttimer
-bokeh
-semver
-flet-webview
-logfire
-sre_yield
+pymupdf
+pytest-cov
+beanie[odm]
+logfire[httpx]
 deepdiff
-json_repair
-cachetools
-distributed
-pymupdf4llm
-httpx_retries
+bokeh
+pycountry
+dnspython[doh]
 faker
-flet[all]
+pydantic-ai[logfire,openai]
+odfpy
+pydantic-settings
+pydantic-extra-types
+torchaudio
 motor
-uvicorn[standard]
-diskcache
-appdirs
-openai
-google-auth-oauthlib
-setuptools
-flet-video
-google-api-python-client
-ollama
-openpyxl
-google-auth
-tinynetrc
-pymupdf
-peft
+torchvision
+filetype
+python-on-whales
+pymupdf4llm
+logfire
+pydevd-pycharm~=251.25410.159
+logfire[fastapi]
+contexttimer
 fastapi
+flet-webview
+playwright
+Unidecode
+distributed
+diskcache
+tokenizers
 
 [api]
 fastapi