mistocr 0.1.3__py3-none-any.whl → 0.1.5__py3-none-any.whl

mistocr/__init__.py

@@ -1 +1 @@
-__version__ = "0.1.3"
+__version__ = "0.1.5"

mistocr/_modidx.py

@@ -21,9 +21,11 @@ d = { 'settings': { 'branch': 'main',
                               '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.refine': { 'mistocr.refine.HeadingCorrections': ('refine.html#headingcorrections', 'mistocr/refine.py'),
+                                'mistocr.refine.add_pg_hdgs': ('refine.html#add_pg_hdgs', 'mistocr/refine.py'),
                                 'mistocr.refine.apply_hdg_fixes': ('refine.html#apply_hdg_fixes', 'mistocr/refine.py'),
                                 'mistocr.refine.fix_hdg_hierarchy': ('refine.html#fix_hdg_hierarchy', 'mistocr/refine.py'),
                                 'mistocr.refine.fix_md_hdgs': ('refine.html#fix_md_hdgs', 'mistocr/refine.py'),
                                 'mistocr.refine.fmt_hdgs_idx': ('refine.html#fmt_hdgs_idx', 'mistocr/refine.py'),
                                 'mistocr.refine.get_hdgs': ('refine.html#get_hdgs', 'mistocr/refine.py'),
-                                'mistocr.refine.mk_fixes_lut': ('refine.html#mk_fixes_lut', 'mistocr/refine.py')}}}
+                                'mistocr.refine.mk_fixes_lut': ('refine.html#mk_fixes_lut', 'mistocr/refine.py'),
+                                'mistocr.refine.read_pgs_pg': ('refine.html#read_pgs_pg', 'mistocr/refine.py')}}}

mistocr/refine.py

@@ -3,8 +3,8 @@
 # AUTOGENERATED! DO NOT EDIT! File to edit: ../nbs/01_refine.ipynb.
 
 # %% auto 0
-__all__ = ['prompt_fix_hdgs', 'get_hdgs', 'fmt_hdgs_idx', 'HeadingCorrections', 'fix_hdg_hierarchy', 'mk_fixes_lut',
-           'apply_hdg_fixes', 'fix_md_hdgs']
+__all__ = ['prompt_fix_hdgs', 'get_hdgs', 'add_pg_hdgs', 'read_pgs_pg', 'fmt_hdgs_idx', 'HeadingCorrections', 'fix_hdg_hierarchy',
+           'mk_fixes_lut', 'apply_hdg_fixes', 'fix_md_hdgs']
 
 # %% ../nbs/01_refine.ipynb 3
 from fastcore.all import *
@@ -15,7 +15,7 @@ from lisette.core import completion
 import os
 import json
 
-# %% ../nbs/01_refine.ipynb 7
+# %% ../nbs/01_refine.ipynb 8
 def get_hdgs(
     md:str # Markdown file string
     ):
@@ -26,7 +26,20 @@ def get_hdgs(
 
 
 
-# %% ../nbs/01_refine.ipynb 10
+# %% ../nbs/01_refine.ipynb 9
+def add_pg_hdgs(md, n):
+    "Add page number to all headings in markdown"
+    md = sub(r'```[\s\S]*?```', '', md)
+    def repl(m): return m.group(0) + f' ... page {n}'
+    return sub(r'^#{1,6} .+$', repl, md, flags=MULTILINE)
+
+# %% ../nbs/01_refine.ipynb 11
+def read_pgs_pg(path):
+    "Read all pages of a markdown file and add page numbers to all headings"
+    pgs = read_pgs(path, join=False)
+    return L([add_pg_hdgs(pg, n) for n, pg in enumerate(pgs, 1)]).concat()
+
+# %% ../nbs/01_refine.ipynb 15
 def fmt_hdgs_idx(
     hdgs: list[str] # List of markdown headings
     ) -> str: # Formatted string with index
@@ -34,19 +47,21 @@ def fmt_hdgs_idx(
     return '\n'.join(f"{i}. {h}" for i, h in enumerate(hdgs))
 
 
-# %% ../nbs/01_refine.ipynb 13
+# %% ../nbs/01_refine.ipynb 18
 class HeadingCorrections(BaseModel):
     corrections: dict[int, str]  # index → corrected heading
 
-# %% ../nbs/01_refine.ipynb 15
+# %% ../nbs/01_refine.ipynb 20
 prompt_fix_hdgs = """Fix markdown heading hierarchy errors while preserving the document's intended structure.
 
-INPUT FORMAT: Each heading is prefixed with its index number (e.g., "0. # Title")
+INPUT FORMAT: Each heading is prefixed with its index number (e.g., "0. # Title ... page 1")
 
 RULES - Apply these fixes in order:
 
-1. **Single H1 rule**: Documents must have exactly ONE # heading (the title/main heading)
-   - All other headings should be ## or deeper
+1. **Single H1 rule**: Documents must have exactly ONE # heading (typically the document title at the top)
+   - If index 0 is already #, then all subsequent headings (index 1+) must be ## or deeper
+   - If no H1 exists, the first major heading should be #, and all others ## or deeper
+   - NO exceptions: appendices, references, and all sections are ## or deeper after the title
 
 2. **Infer depth from numbering patterns**: If headings contain section numbers, deeper nesting means deeper heading level
    - Parent section (e.g., "1", "2", "A") should be shallower than child (e.g., "1.1", "2.a", "A.1")
@@ -60,13 +75,15 @@ RULES - Apply these fixes in order:
 4. **Decreasing levels is OK**: Moving back up the hierarchy (### to ##) is valid for new sections
 
 OUTPUT: Return a Python dictionary mapping index to corrected heading (without the index prefix).
+IMPORTANT: Preserve the " ... page N" suffix in all corrected headings.
 Only include entries that need changes.
 
 Headings to analyze:
 {headings_list}
 """
 
-# %% ../nbs/01_refine.ipynb 16
+
+# %% ../nbs/01_refine.ipynb 21
 def fix_hdg_hierarchy(
     hdgs: list[str], # List of markdown headings
     prompt: str=prompt_fix_hdgs, # Prompt to use
@@ -82,7 +99,7 @@ def fix_hdg_hierarchy(
         )
     return json.loads(r.choices[0].message.content)['corrections']
 
-# %% ../nbs/01_refine.ipynb 19
+# %% ../nbs/01_refine.ipynb 24
 def mk_fixes_lut(
     hdgs: list[str], # List of markdown headings
     model: str='claude-sonnet-4-5', # Model to use
@@ -92,26 +109,25 @@ def mk_fixes_lut(
     fixes = fix_hdg_hierarchy(hdgs, model=model, api_key=api_key)
     return {hdgs[int(k)]:v for k,v in fixes.items()}
 
-# %% ../nbs/01_refine.ipynb 22
+# %% ../nbs/01_refine.ipynb 27
 def apply_hdg_fixes(
     p:str, # Page to fix
     lut_fixes: dict[str, str], # Lookup table of fixes
-    pg: int=None, # Optionnaly specify the page number to append to original heading
     ) -> str: # Page with fixes applied
     "Apply the fixes to the page"
-    for old in get_hdgs(p): p = p.replace(old, lut_fixes.get(old, old) + (f' .... page {pg}' if pg else ''))
+    for old in get_hdgs(p): p = p.replace(old, lut_fixes.get(old, old))
     return p
 
-# %% ../nbs/01_refine.ipynb 25
+# %% ../nbs/01_refine.ipynb 30
 def fix_md_hdgs(
     src:str, # Source directory with markdown pages
     model:str='claude-sonnet-4-5', # Model
     dst:str=None, # Destination directory (None=overwrite)
-    pg_nums:bool=True # Add page numbers
 ):
     "Fix heading hierarchy in markdown document"
     src_path,dst_path = Path(src),Path(dst) if dst else Path(src)
     if dst_path != src_path: dst_path.mkdir(parents=True, exist_ok=True)
-    lut = mk_fixes_lut(get_hdgs(read_pgs(src_path)), model)
-    for i,p in enumerate(read_pgs(src_path, join=False), 1):
-        (dst_path/f'page_{i}.md').write_text(apply_hdg_fixes(p, lut, pg=i if pg_nums else None))
+    pgs_with_pg = read_pgs_pg(src_path)
+    lut = mk_fixes_lut(L([get_hdgs(pg) for pg in pgs_with_pg]).concat(), model)
+    for i,p in enumerate(pgs_with_pg, 1):
+        (dst_path/f'page_{i}.md').write_text(apply_hdg_fixes(p, lut))

{mistocr-0.1.3.dist-info → mistocr-0.1.5.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: mistocr
-Version: 0.1.3
+Version: 0.1.5
 Summary: Simple batch OCR for PDFs using Mistral's state-of-the-art vision model
 Home-page: https://github.com/franckalbinet/mistocr
 Author: Solveit

mistocr-0.1.5.dist-info/RECORD

@@ -0,0 +1,10 @@
+mistocr/__init__.py,sha256=rPSfWgIeq2YWVPyESOAwCBt8vftsTpIkuLAGDEzyRQc,22
+mistocr/_modidx.py,sha256=sZ3ISGF-2f7VEOD9MVgqMVs5SifUNe-1YP0wy8Ey0cU,2884
+mistocr/core.py,sha256=wtaYZ_Fz0dXIb1cYLAtymwR9Z7_KBI4ULy-UpM3YTX4,7154
+mistocr/refine.py,sha256=314r4MBZRIvUmu6B_dvvq9P4d4a_japKBpsg4wnU9oU,5253
+mistocr-0.1.5.dist-info/licenses/LICENSE,sha256=xV8xoN4VOL0uw9X8RSs2IMuD_Ss_a9yAbtGNeBWZwnw,11337
+mistocr-0.1.5.dist-info/METADATA,sha256=kcrDK0kJadP5Sze0tVzRo-pLWWTJQiEnvwlLpWJZz2o,4848
+mistocr-0.1.5.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
+mistocr-0.1.5.dist-info/entry_points.txt,sha256=HjAiHozobM-alm_6bTF-ehRr2DD3KYE9PgRngelONOY,36
+mistocr-0.1.5.dist-info/top_level.txt,sha256=LelTYnSpSXLy1Hb1m2YP3gt8luwP-I8KV0NjP_ucdSs,8
+mistocr-0.1.5.dist-info/RECORD,,

mistocr-0.1.3.dist-info/RECORD

@@ -1,10 +0,0 @@
-mistocr/__init__.py,sha256=XEqb2aiIn8fzGE68Mph4ck1FtQqsR_am0wRWvrYPffQ,22
-mistocr/_modidx.py,sha256=R9zVMv4dKz2sLStoB5wBoKRqjza216_z8xPXszoplU4,2660
-mistocr/core.py,sha256=wtaYZ_Fz0dXIb1cYLAtymwR9Z7_KBI4ULy-UpM3YTX4,7154
-mistocr/refine.py,sha256=572SDG8vhGjNMiET5eZhgVemNpUIHNFqi0ZSSl4eKCM,4545
-mistocr-0.1.3.dist-info/licenses/LICENSE,sha256=xV8xoN4VOL0uw9X8RSs2IMuD_Ss_a9yAbtGNeBWZwnw,11337
-mistocr-0.1.3.dist-info/METADATA,sha256=jHRc6nm_uk7V-03y6Bd268hUWmkkOFNdt4s5cH3YPu0,4848
-mistocr-0.1.3.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
-mistocr-0.1.3.dist-info/entry_points.txt,sha256=HjAiHozobM-alm_6bTF-ehRr2DD3KYE9PgRngelONOY,36
-mistocr-0.1.3.dist-info/top_level.txt,sha256=LelTYnSpSXLy1Hb1m2YP3gt8luwP-I8KV0NjP_ucdSs,8
-mistocr-0.1.3.dist-info/RECORD,,