HowdenParser 3.0.0__tar.gz → 3.1.1__tar.gz

{howdenparser-3.0.0 → howdenparser-3.1.1}/HowdenParser/parser.py

@@ -14,6 +14,10 @@ logger = logging.getLogger(__name__)
 class BaseParser(ABC):
     _registry: dict[str, type["BaseParser"]] = {}
 
+    def __init__(self):
+        # Every instance gets a .name attribute
+        self.name = getattr(self.__class__, "_parser_name", self.__class__.__name__)
+
     def __init_subclass__(cls, name: str | None = None, **kwargs):
         """Automatically register subclasses under a key."""
         super().__init_subclass__(**kwargs)
@@ -57,6 +61,10 @@ class Parser(BaseParser):
             provider_and_model: Literal["llamaparser:"],
             result_type: str,
             mode: bool,
+            preserve_layout_alignment_across_pages: bool,
+            merge_tables_across_pages_in_markdown: bool,
+            hide_footers: bool,
+            hide_headers: bool,
     ) -> "LlamaParser": ...
 
     @overload
@@ -154,7 +162,7 @@ class Parser(BaseParser):
         if missing:
             example = ""
             if provider == "llamaparser":
-                example = "Example: Parser('llamaparser:', result_type='md', mode=True)"
+                example = "Example: Parser('llamaparser:', result_type='md', mode=True, extract_tables=True)"
             elif provider == "mistralocr":
                 example = "Example: Parser('mistralocr:ocr-large')"
             elif provider == "huggingface":
@@ -227,11 +235,21 @@ class LangChainParser(BaseParser, name="langchain"):
 
 
 class LlamaParser(BaseParser, name="llamaparser"):
-    def __init__(self, result_type: str, mode: bool, provider_and_model) -> None:
+    def __init__(self, result_type: str,
+                 mode: bool,
+                 provider_and_model: str,
+                 merge_tables_across_pages_in_markdown: bool,
+                 preserve_layout_alignment_across_pages: bool,
+                 hide_footers: bool,
+                 hide_headers: bool) -> None:
         logging.info("Initializing LlamaParser...")
         self.result_type = result_type
         self.mode = mode
         self.provider_and_model = provider_and_model
+        self.merge_tables_across_pages_in_markdown = merge_tables_across_pages_in_markdown
+        self.preserve_layout_alignment_across_pages = preserve_layout_alignment_across_pages
+        self.hide_footers=hide_footers
+        self.hide_headers=hide_headers
 
         from llama_parse import LlamaParse, ResultType
         if result_type.lower() in ("md", "markdown"):
@@ -239,7 +257,13 @@ class LlamaParser(BaseParser, name="llamaparser"):
         api_key = os.getenv("LLAMA-PARSER-API-TOKEN")
         if not api_key:
             raise EnvironmentError("Missing LLAMA-PARSER-API-TOKEN in .env file.")
-        self._parser = LlamaParse(api_key=api_key, result_type=result_type, premium_mode=self.mode)
+        self._parser = LlamaParse(api_key=api_key,
+                                  result_type=result_type,
+                                  premium_mode=self.mode,
+                                  merge_tables_across_pages_in_markdown=self.merge_tables_across_pages_in_markdown,
+                                  preserve_layout_alignment_across_pages=self.preserve_layout_alignment_across_pages,
+                                  hide_footers=self.hide_footers,
+                                  hide_headers=self.hide_headers)
 
     def parse(self, file_path: Path) -> str:
         documents = self._parser.load_data(str(file_path))

{howdenparser-3.0.0 → howdenparser-3.1.1}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.3
 Name: HowdenParser
-Version: 3.0.0
+Version: 3.1.1
 Summary: A simple configuration manager with Pydantic and JSON export.
 License: MIT
 Keywords: config,configuration,pydantic,json

{howdenparser-3.0.0 → howdenparser-3.1.1}/pyproject.toml

@@ -14,7 +14,7 @@ build-backend = "poetry.core.masonry.api"
 
 [tool.poetry]
 name = "HowdenParser"
-version = "3.0.0"
+version = "3.1.1"
 description = "A simple configuration manager with Pydantic and JSON export."
 authors = [ "JesperThoftIllemannJ <jesper.jaeger@howdendanmark.dk>",]
 readme = "README.md"