HowdenParser 5.2.5__tar.gz → 6.0.0__tar.gz

{howdenparser-5.2.5 → howdenparser-6.0.0}/HowdenParser/parser.py

@@ -96,13 +96,11 @@ class Parser(BaseParser):
             cls,
             *,
             provider_and_model: Literal["llamaparser:"],
+            tier: str,
+            version: str,
             result_type: str,
-            model: str,
-            parse_mode: str,
             preserve_layout_alignment_across_pages: bool,
-            merge_tables_across_pages_in_markdown: bool,
-            hide_footers: bool,
-            hide_headers: bool,
+            merge_continued_tables: bool,
     ) -> "LlamaParser": ...
 
     @overload

{howdenparser-5.2.5 → howdenparser-6.0.0}/HowdenParser/parsers/llama_parser.py

@@ -1,15 +1,18 @@
 import os
-import json
 import logging
 from pathlib import Path
 from ..parser import BaseParser
 from typing import Any
-from llama_parse import ResultType
 
 class LlamaParser(BaseParser, name="llamaparser"):
-    def __init__(self, result_type: str, model: str, parse_mode: str, provider_and_model: str,
-                 merge_tables_across_pages_in_markdown: bool, preserve_layout_alignment_across_pages: bool,
-                 hide_footers: bool, hide_headers: bool) -> None:
+    def __init__(self,
+                 result_type: str,
+                 tier: str, # One of: ["fast", "cost_effective", "agentic", "agentic_plus"]
+                 version: str, # one of: ["latest","2026-03-12","2026-03-11"...] (many not listed)
+                 provider_and_model: str,
+                 merge_continued_tables: bool,
+                 preserve_layout_alignment_across_pages: bool,
+                 ) -> None:
 
         super().__init__()
         # Capture only input parameters
@@ -23,50 +26,39 @@ class LlamaParser(BaseParser, name="llamaparser"):
 
         # Store configuration only
         self.result_type = result_type
-        self.model = model
-        self.parse_mode = parse_mode
+        self.tier = tier
+        self.version = version
         self.provider_and_model = provider_and_model
-        self.merge_tables_across_pages_in_markdown = merge_tables_across_pages_in_markdown
+        self.merge_continued_tables = merge_continued_tables
         self.preserve_layout_alignment_across_pages = preserve_layout_alignment_across_pages
-        self.hide_footers = hide_footers
-        self.hide_headers = hide_headers
         self.hashed = self.compute_hash(self._input_params)
         # DO NOT CREATE LlamaParse HERE
         # It creates asyncio objects bound to the wrong loop
         self._parser = None
+        self.rt = None
 
     def _lazy_init(self):
         """Initialize LlamaParse inside the worker thread event loop."""
         if self._parser is not None:
             return
 
-        from llama_parse import LlamaParse, ResultType
+        from llama_cloud import LlamaCloud
 
         api_key = os.getenv("LLAMA-PARSER-API-TOKEN")
         if not api_key:
             raise EnvironmentError("Missing LLAMA-PARSER-API-TOKEN in .env file.")
 
-        rt_lower = str(self.result_type).lower()
+        rt_lower = str(self.result_type).lower() # Possible values for LlamaCloud expand: ["markdown", "text", "metadata", "items"]
         if rt_lower in ("md","markdown"):
-            rt = ResultType.MD
+            self.rt = 'markdown'
         elif rt_lower in ("txt","text"):
-            rt = ResultType.TXT
+            self.rt = 'text'
         elif rt_lower in ("json"):
-            rt = ResultType.JSON
+            self.rt = 'items'
         else:
             raise NotImplementedError(f"Result type {self.result_type} is not supported.")
 
-        # Construct the parser INSIDE the worker thread
-        self._parser = LlamaParse(
-            api_key=api_key,
-            result_type=rt,
-            model=self.model,
-            parse_mode=self.parse_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,
-        )
+        self._parser = LlamaCloud(api_key=os.getenv("LLAMA-PARSER-API-TOKEN"))
 
         logging.info("LlamaParser initialized inside worker thread")
 
@@ -76,14 +68,31 @@ class LlamaParser(BaseParser, name="llamaparser"):
         def page_break(text: str, idx: int) -> str:
             return f"<PAGE_NUMBER {idx}>{text}</PAGE_NUMBER {idx}>"
 
-        documents = self._parser.parse(file_path)
-        if self.result_type in ['md','markdown']:
-            if include_pagenumbers: return "\n".join(page_break(page.md, idx) for idx, page in enumerate(documents.pages, start=1))
-            else: return "\n".join(page.md for page in documents.pages)
-        elif self.result_type in ['txt','text']:
-            if include_pagenumbers: return "\n".join(page_break(page.text, idx) for idx, page in enumerate(documents.pages, start=1))
-            else: return "\n".join(page.text for page in documents.pages)
-        elif self.result_type.lower() == 'json':
+        documents = self._parser.parsing.parse(
+            upload_file=file_path,
+            tier=self.tier,
+            version=self.version,
+            output_options={
+                "markdown": {
+                    "tables": {
+                        "output_tables_as_markdown": True,
+                        "merge_continued_tables": self.merge_continued_tables,
+                    }
+                },
+                "spatial_text": {
+                    "preserve_layout_alignment_across_pages": self.preserve_layout_alignment_across_pages,
+                }
+            },
+            expand=[self.rt],
+        )
+
+        if self.rt == 'markdown':
+            if include_pagenumbers: return "\n".join(page_break(page.markdown, idx) for idx, page in enumerate(documents.markdown.pages, start=1))
+            else: return "\n".join(page.markdown for page in documents.markdown.pages)
+        elif self.rt == 'text':
+            if include_pagenumbers: return "\n".join(page_break(page.text, idx) for idx, page in enumerate(documents.text.pages, start=1))
+            else: return "\n".join(page.text for page in documents.text.pages)
+        elif self.rt == 'items':
             return documents.model_dump()
         else:
             raise NotImplementedError(f"Result type {self.result_type} is not supported.")

{howdenparser-5.2.5 → howdenparser-6.0.0}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: HowdenParser
-Version: 5.2.5
+Version: 6.0.0
 Summary: A simple configuration manager with Pydantic and JSON export.
 License: MIT
 Keywords: config,configuration,pydantic,json
@@ -11,8 +11,9 @@ Classifier: License :: OSI Approved :: MIT License
 Classifier: Programming Language :: Python :: 3
 Classifier: Programming Language :: Python :: 3.12
 Classifier: Programming Language :: Python :: 3.13
+Requires-Dist: dotenv (>=0.9.9,<0.10.0)
 Requires-Dist: langchain (>=1.1.3,<2.0.0)
-Requires-Dist: llama-parse (>=0.6.58,<0.7.0)
+Requires-Dist: llama-cloud (>=1.0)
 Requires-Dist: mistralai (>=1.9.3,<2.0.0)
 Requires-Dist: pdf2image (>=1.17.0,<2.0.0)
 Requires-Dist: pypdf2 (>=3.0.1,<4.0.0)

{howdenparser-5.2.5 → howdenparser-6.0.0}/pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "HowdenParser"
-version = "5.2.5"
+version = "6.0.0"
 description = "A simple configuration manager with Pydantic and JSON export."
 readme = "README.md"
 requires-python = ">=3.12,<3.14"
@@ -18,12 +18,13 @@ repository = "https://github.com/yourusername/config"
 documentation = "https://github.com/yourusername/config"
 dependencies = [
     "mistralai (>=1.9.3,<2.0.0)",
-    "llama-parse (>=0.6.58,<0.7.0)",
+    "llama-cloud (>=1.0)",
     "pypdf2 (>=3.0.1,<4.0.0)",
     "pdf2image (>=1.17.0,<2.0.0)",
     "langchain (>=1.1.3,<2.0.0)",
     "pytest (>=9.0.2,<10.0.0)",
     "tomli-w (>=1.2.0,<2.0.0)",
+    "dotenv (>=0.9.9,<0.10.0)",
 ]
 
 [project.license]