mistocr 0.3.1__py3-none-any.whl → 0.3.2__py3-none-any.whl

mistocr/__init__.py

@@ -1 +1 @@
-__version__ = "0.3.1"
+__version__ = "0.3.2"

mistocr/_modidx.py

@@ -21,7 +21,7 @@ d = { 'settings': { 'branch': 'main',
                               'mistocr.core.subset_pdf': ('core.html#subset_pdf', 'mistocr/core.py'),
                               'mistocr.core.upload_pdf': ('core.html#upload_pdf', 'mistocr/core.py'),
                               'mistocr.core.wait_for_job': ('core.html#wait_for_job', 'mistocr/core.py')},
-            'mistocr.pipeline': {},
+            'mistocr.pipeline': {'mistocr.pipeline.pdf_to_md': ('pipeline.html#pdf_to_md', 'mistocr/pipeline.py')},
             'mistocr.refine': { 'mistocr.refine.HeadingCorrection': ('refine.html#headingcorrection', 'mistocr/refine.py'),
                                 'mistocr.refine.HeadingCorrections': ('refine.html#headingcorrections', 'mistocr/refine.py'),
                                 'mistocr.refine.ImgDescription': ('refine.html#imgdescription', 'mistocr/refine.py'),

mistocr/pipeline.py

@@ -3,7 +3,7 @@
 # AUTOGENERATED! DO NOT EDIT! File to edit: ../nbs/02_pipeline.ipynb.
 
 # %% auto 0
-__all__ = []
+__all__ = ['pdf_to_md']
 
 # %% ../nbs/02_pipeline.ipynb 3
 from fastcore.all import *
@@ -13,3 +13,29 @@ from pathlib import Path
 from asyncio import Semaphore, gather, sleep
 import tempfile
 import os, json, shutil
+
+# %% ../nbs/02_pipeline.ipynb 4
+@delegates(add_img_descs)
+async def pdf_to_md(
+    pdf_path:str,                   # Path to input PDF file
+    dst:str,                        # Destination directory for output markdown
+    ocr_dst:str=None,               # Optional OCR output directory
+    model:str='claude-sonnet-4-5',  # Model to use for heading fixes and image descriptions
+    add_img_desc:bool=True,         # Whether to add image descriptions
+    progress:bool=True,             # Whether to show progress messages
+    **kwargs
+    ):
+    "Convert PDF to markdown with OCR, fixed heading hierarchy, and optional image descriptions"
+    cleanup = ocr_dst is None
+    if cleanup: ocr_dst = tempfile.mkdtemp()
+    n_steps = 3 if add_img_desc else 2
+    if progress: print(f"Step 1/{n_steps}: Running OCR on {pdf_path}...")
+    ocr_dir = ocr_pdf(pdf_path, ocr_dst)[0]
+    if progress: print(f"Step 2/{n_steps}: Fixing heading hierarchy...")
+    fix_hdgs(ocr_dir, model=model)
+    if add_img_desc:
+        if progress: print(f"Step 3/{n_steps}: Adding image descriptions...")
+        await add_img_descs(ocr_dir, dst=dst, model=model, progress=progress, **kwargs)
+    elif dst != str(ocr_dir): shutil.copytree(ocr_dir, dst, dirs_exist_ok=True)
+    if cleanup: shutil.rmtree(ocr_dst)
+    if progress: print("Done!")

{mistocr-0.3.1.dist-info → mistocr-0.3.2.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: mistocr
-Version: 0.3.1
+Version: 0.3.2
 Summary: Batch OCR for PDFs with heading restoration and visual content integration
 Home-page: https://github.com/franckalbinet/mistocr
 Author: Solveit

mistocr-0.3.2.dist-info/RECORD

@@ -0,0 +1,11 @@
+mistocr/__init__.py,sha256=vNiWJ14r_cw5t_7UDqDQIVZvladKFGyHH2avsLpN7Vg,22
+mistocr/_modidx.py,sha256=2rHVTcz3A3BfDCmGqUNsJFW3_n3Ch1vxSorrPVyLvMI,4256
+mistocr/core.py,sha256=EQYQgpnX2skgSX123u3dYaJHc1oDk5Nhgt5uBdXnCKs,8386
+mistocr/pipeline.py,sha256=hVXpxRYtshaiUm9qXgfSLlyHCAxHZ6nAfPzoGXGmJMQ,1769
+mistocr/refine.py,sha256=hQg4ZYRTkPFEiCNNBU2ykgXFfVLCNAp2IuwOwHBVQ2k,12983
+mistocr-0.3.2.dist-info/licenses/LICENSE,sha256=xV8xoN4VOL0uw9X8RSs2IMuD_Ss_a9yAbtGNeBWZwnw,11337
+mistocr-0.3.2.dist-info/METADATA,sha256=igTgaDeBu00u_xJYtIcGlQswQCj2gIrdBi6NLiN5NNU,8438
+mistocr-0.3.2.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
+mistocr-0.3.2.dist-info/entry_points.txt,sha256=HjAiHozobM-alm_6bTF-ehRr2DD3KYE9PgRngelONOY,36
+mistocr-0.3.2.dist-info/top_level.txt,sha256=LelTYnSpSXLy1Hb1m2YP3gt8luwP-I8KV0NjP_ucdSs,8
+mistocr-0.3.2.dist-info/RECORD,,

mistocr-0.3.1.dist-info/RECORD

@@ -1,11 +0,0 @@
-mistocr/__init__.py,sha256=r4xAFihOf72W9TD-lpMi6ntWSTKTP2SlzKP1ytkjRbI,22
-mistocr/_modidx.py,sha256=4apaXSL_JCEHVA8B-tH3w_23jIZBzBflE3vnQ9TQDDo,4176
-mistocr/core.py,sha256=EQYQgpnX2skgSX123u3dYaJHc1oDk5Nhgt5uBdXnCKs,8386
-mistocr/pipeline.py,sha256=TDk8qMGFLpQReGelNyQ9mHnh2cIb-5P2YgpDHC0iqdI,438
-mistocr/refine.py,sha256=hQg4ZYRTkPFEiCNNBU2ykgXFfVLCNAp2IuwOwHBVQ2k,12983
-mistocr-0.3.1.dist-info/licenses/LICENSE,sha256=xV8xoN4VOL0uw9X8RSs2IMuD_Ss_a9yAbtGNeBWZwnw,11337
-mistocr-0.3.1.dist-info/METADATA,sha256=O6X89wifAKyOWhee-7SN5SuDkuZZ0W5FIOEEGZ_8J3M,8438
-mistocr-0.3.1.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
-mistocr-0.3.1.dist-info/entry_points.txt,sha256=HjAiHozobM-alm_6bTF-ehRr2DD3KYE9PgRngelONOY,36
-mistocr-0.3.1.dist-info/top_level.txt,sha256=LelTYnSpSXLy1Hb1m2YP3gt8luwP-I8KV0NjP_ucdSs,8
-mistocr-0.3.1.dist-info/RECORD,,