feat!: Massive quality improvements to v2 parser and new sanitize_cells API (#73)

Signed-off-by: Peter Staar <[email protected]>
Signed-off-by: Christoph Auer <[email protected]>
Co-authored-by: Christoph Auer <[email protected]>

Author: PeterStaar-IBM

.github/scripts/build_rhel.sh

@@ -40,7 +40,7 @@ sudo -E XDG_RUNTIME_DIR= podman build --progress=plain \
             --no-deps --no-build-isolation -w /dist/ \
             /src/docling_parse*.tar.gz \
         && pip3.11 install /dist/docling_parse*.whl \
-        && python3.11 -c 'from docling_parse.docling_parse import pdf_parser_v1, pdf_parser_v2'
+        && python3.11 -c 'from docling_parse.pdf_parsers import pdf_parser_v1, pdf_parser_v2'
     
     COPY ./tests /src/tests
 

.pre-commit-config.yaml

@@ -4,18 +4,42 @@ repos:
     hooks:
       - id: system
         name: Black
-        entry: poetry run black docling_parse
+        entry: poetry run black docling_parse tests
         pass_filenames: false
         language: system
         files: '\.py$'
   - repo: local
     hooks:
       - id: system
         name: isort
-        entry: poetry run isort docling_parse
+        entry: poetry run isort docling_parse tests
         pass_filenames: false
         language: system
         files: '\.py$'
+  - repo: local
+    hooks:
+      - id: autoflake
+        name: autoflake
+        entry: poetry run autoflake docling_parse
+        pass_filenames: false
+        language: system
+        files: '\.py$'
+  - repo: local
+    hooks:
+      - id: mypy
+        name: MyPy
+        entry: poetry run mypy docling_parse tests
+        pass_filenames: false
+        language: system
+        files: '\.py$'
+#  - repo: local
+#    hooks:
+#      - id: pytest
+#        name: Pytest
+#        entry: poetry run pytest tests/
+#        pass_filenames: false
+#        language: system
+#        files: '\.py$'	
   - repo: local
     hooks:
       - id: system

CMakeLists.txt

@@ -192,18 +192,24 @@ target_link_libraries(parse_v2 ${DEPENDENCIES} ${LIB_LINK})
 # https://pybind11.readthedocs.io/en/stable/compiling.html#configuration-variables
 find_package(pybind11 CONFIG REQUIRED)
 
-pybind11_add_module(docling_parse "${TOPLEVEL_PREFIX_PATH}/app/pybind_parse.cpp")
+#pybind11_add_module(docling_parse "${TOPLEVEL_PREFIX_PATH}/app/pybind_parse.cpp")
+pybind11_add_module(pdf_parsers "${TOPLEVEL_PREFIX_PATH}/app/pybind_parse.cpp")
 
-add_dependencies(docling_parse parse_v1 parse_v2)
+#add_dependencies(docling_parse parse_v1 parse_v2)
+add_dependencies(pdf_parsers parse_v1 parse_v2)
 
-target_include_directories(docling_parse INTERFACE ${DEPENDENCIES})
+#target_include_directories(docling_parse INTERFACE ${DEPENDENCIES})
+target_include_directories(pdf_parsers INTERFACE ${DEPENDENCIES})
 
-target_compile_definitions(docling_parse PRIVATE VERSION_INFO=${CMAKE_PROJECT_VERSION})
+#target_compile_definitions(docling_parse PRIVATE VERSION_INFO=${CMAKE_PROJECT_VERSION})
+target_compile_definitions(pdf_parsers PRIVATE VERSION_INFO=${CMAKE_PROJECT_VERSION})
 
-target_link_libraries(docling_parse PRIVATE parse_v1 parse_v2)
+#target_link_libraries(docling_parse PRIVATE parse_v1 parse_v2)
+target_link_libraries(pdf_parsers PRIVATE parse_v1 parse_v2)
 
 # *****************
 # ***  Install  ***
 # *****************
 
-install(TARGETS docling_parse DESTINATION "${TOPLEVEL_PREFIX_PATH}/docling_parse")
+#install(TARGETS docling_parse DESTINATION "${TOPLEVEL_PREFIX_PATH}/docling_parse")
+install(TARGETS pdf_parsers DESTINATION "${TOPLEVEL_PREFIX_PATH}/docling_parse")

README.md

@@ -48,9 +48,7 @@ Convert a PDF (look in the [visualise.py](docling_parse/visualise.py) for a more
 from docling_parse.docling_parse import pdf_parser_v2
 
 # Do this only once to load fonts (avoid initialising it many times)
-parser = pdf_parser_v2()
-
-# parser.set_loglevel(1) # 1=error, 2=warning, 3=success, 4=info
+parser = pdf_parser_v2("error") # info, warning, error, fatal
 
 doc_file = "my-doc.pdf" # filename
 doc_key = f"key={pdf_doc}" # unique document key (eg hash, UUID, etc)
@@ -167,7 +165,7 @@ If you dont have an input file, then a template input file will be printed on th
 To build the package, simply run (make sure [poetry](https://python-poetry.org/) is [installed](https://python-poetry.org/docs/#installing-with-the-official-installer)),
 
 ```
-poetry build
+poetry install
 ```
 
 To test the package, run: