mistocr 0.2.11__py3-none-any.whl → 0.3.0__py3-none-any.whl

mistocr/__init__.py

@@ -1 +1 @@
-__version__ = "0.2.11"
+__version__ = "0.3.0"

mistocr/_modidx.py

@@ -18,10 +18,12 @@ d = { 'settings': { 'branch': 'main',
                               'mistocr.core.save_page': ('core.html#save_page', 'mistocr/core.py'),
                               'mistocr.core.save_pages': ('core.html#save_pages', 'mistocr/core.py'),
                               'mistocr.core.submit_batch': ('core.html#submit_batch', 'mistocr/core.py'),
+                              '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.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.refine.add_descs_to_pg': ('refine.html#add_descs_to_pg', 'mistocr/refine.py'),
                                 'mistocr.refine.add_descs_to_pgs': ('refine.html#add_descs_to_pgs', 'mistocr/refine.py'),

mistocr/core.py

@@ -4,7 +4,8 @@
 
 # %% auto 0
 __all__ = ['ocr_model', 'ocr_endpoint', 'get_api_key', 'upload_pdf', 'create_batch_entry', 'prep_pdf_batch', 'submit_batch',
-           'wait_for_job', 'download_results', 'save_images', 'save_page', 'save_pages', 'ocr_pdf', 'read_pgs']
+           'wait_for_job', 'download_results', 'save_images', 'save_page', 'save_pages', 'ocr_pdf', 'read_pgs',
+           'subset_pdf']
 
 # %% ../nbs/00_core.ipynb 3
 from fastcore.all import *
@@ -13,6 +14,7 @@ from io import BytesIO
 from pathlib import Path
 from PIL import Image
 from mistralai import Mistral
+import PyPDF2
 
 # %% ../nbs/00_core.ipynb 6
 def get_api_key(
@@ -181,7 +183,7 @@ def ocr_pdf(
     results = _run_batch(entries, c, poll_interval)
     return L([save_pages(r['response']['body'], dst, r['custom_id']) for r in results])
 
-# %% ../nbs/00_core.ipynb 47
+# %% ../nbs/00_core.ipynb 50
 def read_pgs(
     path:str, # OCR output directory, 
     join:bool=True # Join pages into single string
@@ -191,3 +193,24 @@ def read_pgs(
     pgs = sorted(path.glob('page_*.md'), key=lambda p: int(p.stem.split('_')[1]))
     contents = L([p.read_text() for p in pgs])
     return '\n\n'.join(contents) if join else contents
+
+# %% ../nbs/00_core.ipynb 57
+def subset_pdf(
+    path:str, # Path to PDF file
+    start:int=1, # Start page (1-based)
+    end:int=None, # End page (1-based, inclusive)
+    dst:str='.' # Output directory
+    ) -> Path: # Path to subset PDF
+    "Extract page range from PDF and save with range suffix"
+    path = Path(path)
+    writer = PyPDF2.PdfWriter()
+    with open(path, 'rb') as f:
+        reader = PyPDF2.PdfReader(f)
+        n = len(reader.pages)
+        end = end or n
+        s, e = max(0, start-1), min(n, end) - 1
+        for i in range(s, e+1): writer.add_page(reader.pages[i])
+    suffix = f"_p{s+1}-{e+1}" if s>0 or e<n-1 else ""
+    out = Path(dst) / f"{path.stem}{suffix}.pdf"
+    with open(out, 'wb') as f: writer.write(f)
+    return out

mistocr/refine.py

@@ -4,9 +4,9 @@
 
 # %% auto 0
 __all__ = ['prompt_fix_hdgs', 'describe_img_prompt', 'get_hdgs', 'add_pg_hdgs', 'read_pgs_pg', 'fmt_hdgs_idx',
-           'HeadingCorrection', 'fix_hdg_hierarchy', 'mk_fixes_lut', 'apply_hdg_fixes', 'fix_hdgs', 'ImgDescription',
-           'describe_img', 'limit', 'parse_r', 'describe_imgs', 'save_img_descs', 'add_descs_to_pg', 'add_descs_to_pgs',
-           'add_img_descs']
+           'HeadingCorrection', 'HeadingCorrections', 'fix_hdg_hierarchy', 'mk_fixes_lut', 'apply_hdg_fixes',
+           'fix_hdgs', 'ImgDescription', 'describe_img', 'limit', 'parse_r', 'describe_imgs', 'save_img_descs',
+           'add_descs_to_pg', 'add_descs_to_pgs', 'add_img_descs']
 
 # %% ../nbs/01_refine.ipynb 3
 from fastcore.all import *
@@ -64,6 +64,11 @@ class HeadingCorrection(BaseModel):
     index: int
     corrected: str
 
+# %% ../nbs/01_refine.ipynb 19
+class HeadingCorrections(BaseModel):
+    "Collection of heading corrections returned by the LLM"
+    corrections: list[HeadingCorrection]
+
 # %% ../nbs/01_refine.ipynb 21
 prompt_fix_hdgs = """Fix markdown heading hierarchy errors while preserving the document's intended structure.
 

{mistocr-0.2.11.dist-info → mistocr-0.3.0.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: mistocr
-Version: 0.2.11
+Version: 0.3.0
 Summary: Batch OCR for PDFs with heading restoration and visual content integration
 Home-page: https://github.com/franckalbinet/mistocr
 Author: Solveit
@@ -23,6 +23,7 @@ Requires-Dist: mistralai
 Requires-Dist: pillow
 Requires-Dist: dotenv
 Requires-Dist: lisette
+Requires-Dist: PyPDF2
 Provides-Extra: dev
 Dynamic: author
 Dynamic: author-email

mistocr-0.3.0.dist-info/RECORD

@@ -0,0 +1,11 @@
+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,,

mistocr-0.2.11.dist-info/RECORD

@@ -1,11 +0,0 @@
-mistocr/__init__.py,sha256=_MLx4ac1juJPWEEiC9kMQISX3x3jFBr507jM2P_hxMg,23
-mistocr/_modidx.py,sha256=WTS9JpZdbrp2LghjhOV-CK0JYChHE4PzttgKfh7pTy4,4028
-mistocr/core.py,sha256=ohh2ru05gUKbIQCRHPMz_hw4ui39FtpoV3_S3n4bl_c,7592
-mistocr/pipeline.py,sha256=k5rtr9ccxTgwQAzw533iMIL9qwiCofugWv0fhotLzaI,1639
-mistocr/refine.py,sha256=H_IAF02k6CwBQdDJm9txknzUcTlz245zXitaHELX-P4,12791
-mistocr-0.2.11.dist-info/licenses/LICENSE,sha256=xV8xoN4VOL0uw9X8RSs2IMuD_Ss_a9yAbtGNeBWZwnw,11337
-mistocr-0.2.11.dist-info/METADATA,sha256=CZ9TO24sY0OT9B19KGPIuy3MF9T7B1G9qHaVBoEMRIk,8417
-mistocr-0.2.11.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
-mistocr-0.2.11.dist-info/entry_points.txt,sha256=HjAiHozobM-alm_6bTF-ehRr2DD3KYE9PgRngelONOY,36
-mistocr-0.2.11.dist-info/top_level.txt,sha256=LelTYnSpSXLy1Hb1m2YP3gt8luwP-I8KV0NjP_ucdSs,8
-mistocr-0.2.11.dist-info/RECORD,,