mistocr 0.3.0__py3-none-any.whl → 0.3.2__py3-none-any.whl

mistocr/__init__.py

@@ -1 +1 @@
-__version__ = "0.3.0"
+__version__ = "0.3.2"

mistocr/pipeline.py

@@ -11,27 +11,31 @@ from .core import read_pgs, ocr_pdf
 from .refine import add_img_descs, fix_hdgs
 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_output:str=None, # Optional OCR output directory (defaults to pdf_path stem)
-    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):
+    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_dirs = ocr_pdf(pdf_path, ocr_output or 'ocr_temp')
-    ocr_dir = ocr_dirs[0]
+    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 and Path(dst) != ocr_dir: shutil.copytree(ocr_dir, dst, dirs_exist_ok=True)
+    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-0.3.0.dist-info → mistocr-0.3.2.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: mistocr
-Version: 0.3.0
+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-0.3.0.dist-info/RECORD

@@ -1,11 +0,0 @@
-mistocr/__init__.py,sha256=VrXpHDu3erkzwl_WXrqINBm9xWkcyUy53IQOj042dOs,22
-mistocr/_modidx.py,sha256=2rHVTcz3A3BfDCmGqUNsJFW3_n3Ch1vxSorrPVyLvMI,4256
-mistocr/core.py,sha256=EQYQgpnX2skgSX123u3dYaJHc1oDk5Nhgt5uBdXnCKs,8386
-mistocr/pipeline.py,sha256=k5rtr9ccxTgwQAzw533iMIL9qwiCofugWv0fhotLzaI,1639
-mistocr/refine.py,sha256=hQg4ZYRTkPFEiCNNBU2ykgXFfVLCNAp2IuwOwHBVQ2k,12983
-mistocr-0.3.0.dist-info/licenses/LICENSE,sha256=xV8xoN4VOL0uw9X8RSs2IMuD_Ss_a9yAbtGNeBWZwnw,11337
-mistocr-0.3.0.dist-info/METADATA,sha256=XTduA6AA_UcHYrjeJo4oOl9tQYnIfqfiP3oNJvy_nTI,8438
-mistocr-0.3.0.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
-mistocr-0.3.0.dist-info/entry_points.txt,sha256=HjAiHozobM-alm_6bTF-ehRr2DD3KYE9PgRngelONOY,36
-mistocr-0.3.0.dist-info/top_level.txt,sha256=LelTYnSpSXLy1Hb1m2YP3gt8luwP-I8KV0NjP_ucdSs,8
-mistocr-0.3.0.dist-info/RECORD,,