HowdenParser 5.2.2__tar.gz → 5.2.3__tar.gz

{howdenparser-5.2.2 → howdenparser-5.2.3}/HowdenParser/parsers/llama_parser.py

@@ -1,7 +1,9 @@
 import os
+import json
 import logging
 from pathlib import Path
 from ..parser import BaseParser
+from typing import Any
 
 class LlamaParser(BaseParser, name="llamaparser"):
     def __init__(self, result_type: str, model: str, parse_mode: str, provider_and_model: str,
@@ -44,8 +46,12 @@ class LlamaParser(BaseParser, name="llamaparser"):
             raise EnvironmentError("Missing LLAMA-PARSER-API-TOKEN in .env file.")
 
         rt = self.result_type
-        if rt.lower() in ("md", "markdown"):
+
+        rt_lower = str(rt).lower()
+        if rt_lower in ("md", "markdown"):
             rt = ResultType.MD
+        elif rt_lower in ("json", "layout_json", "jsonl"):
+            rt = ResultType.JSON
 
         # Construct the parser INSIDE the worker thread
         self._parser = LlamaParse(
@@ -61,7 +67,7 @@ class LlamaParser(BaseParser, name="llamaparser"):
 
         logging.info("LlamaParser initialized inside worker thread")
 
-    def parse(self, file_path: Path, include_pagenumbers: bool = False) -> str:
+    def parse(self, file_path: Path, include_pagenumbers: bool = False) -> Any:
         # Ensure parser is created inside the thread event loop
         self._lazy_init()
 
@@ -79,6 +85,9 @@ class LlamaParser(BaseParser, name="llamaparser"):
                 for idx, page in enumerate(documents.pages, start=1)
             )
 
+        if self.result_type.lower() in ("json", "layout_json", "jsonl"):
+            return documents.model_dump() if hasattr(documents, "model_dump") else documents
+
         return "\n".join(
             f"<PAGE_NUMBER {idx}>{page.text}</PAGE_NUMBER {idx}>"
             for idx, page in enumerate(documents.pages, start=1)
@@ -88,4 +97,4 @@ class LlamaParser(BaseParser, name="llamaparser"):
         return self.parse(file_path)
 
     def write_json_hyperparameter(self, folder_file_path: Path) -> None:
-        self.write_parameters(self._input_params, folder_file_path)
+        self.write_parameters(self._input_params, folder_file_path)

{howdenparser-5.2.2 → howdenparser-5.2.3}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: HowdenParser
-Version: 5.2.2
+Version: 5.2.3
 Summary: A simple configuration manager with Pydantic and JSON export.
 License: MIT
 Keywords: config,configuration,pydantic,json

{howdenparser-5.2.2 → howdenparser-5.2.3}/pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "HowdenParser"
-version = "5.2.2"
+version = "5.2.3"
 description = "A simple configuration manager with Pydantic and JSON export."
 readme = "README.md"
 requires-python = ">=3.12,<3.14"