HowdenParser 5.2.6__tar.gz → 6.0.1__tar.gz

{howdenparser-5.2.6 → howdenparser-6.0.1}/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-6.0.1/HowdenParser/parsers/llama_parser.py

@@ -0,0 +1,116 @@
+import os
+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,
+                 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
+        self._input_params = {
+            k: v
+            for k, v in locals().items()
+            if k not in ("self", "__class__", "__len__")
+        }
+
+        logging.info("Configuring LlamaParser (lazy init enabled)...")
+
+        # Store configuration only
+        self.result_type = result_type
+        self.tier = tier
+        self.version = version
+        self.provider_and_model = provider_and_model
+        self.merge_continued_tables = merge_continued_tables
+        self.preserve_layout_alignment_across_pages = preserve_layout_alignment_across_pages
+        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_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()  # Possible values for LlamaCloud expand: ["markdown", "markdown_full", "text", "metadata", "items"]
+        if rt_lower in ("md", "markdown"):
+            self.rt = 'markdown'
+        elif rt_lower in ("md_full", "markdown_full"):
+            self.rt = 'markdown_full'
+        elif rt_lower in ("txt", "text"):
+            self.rt = 'text'
+        elif rt_lower in ("json"):
+            self.rt = 'items'
+        else:
+            raise NotImplementedError(f"Result type {self.result_type} is not supported.")
+
+        self._parser = LlamaCloud(api_key=os.getenv("LLAMA-PARSER-API-TOKEN"))
+
+        logging.info("LlamaParser initialized inside worker thread")
+
+    def parse(self, file_path: Path, include_pagenumbers: bool = False) -> Any:
+        # Ensure parser is created inside the thread event loop
+        self._lazy_init()
+
+        def page_break(text: str, idx: int) -> str:
+            return f"<PAGE_NUMBER {idx}>{text}</PAGE_NUMBER {idx}>"
+
+        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 == 'markdown_full':
+            return documents.markdown_full
+        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.")
+
+    def __call__(self, file_path: Path) -> str:
+        return self.parse(file_path)
+
+    def write_json_hyperparameter(self, folder_file_path: Path) -> None:
+        self.write_parameters(self._input_params, folder_file_path)

{howdenparser-5.2.6 → howdenparser-6.0.1}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: HowdenParser
-Version: 5.2.6
+Version: 6.0.1
 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-5.2.6 → howdenparser-6.0.1}/pyproject.toml

@@ -1,43 +1,44 @@
-[project]
-name = "HowdenParser"
-version = "5.2.6"
-description = "A simple configuration manager with Pydantic and JSON export."
-readme = "README.md"
-requires-python = ">=3.12,<3.14"
-authors = [
-    { name = "JesperThoftIllemannJ", email = "jesper.jaeger@howdendanmark.dk" },
-]
-keywords = [
-    "config",
-    "configuration",
-    "pydantic",
-    "json",
-]
-homepage = "https://github.com/yourusername/config"
-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)",
-    "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)",
-]
-
-[project.license]
-text = "MIT"
-
-[build-system]
-requires = [
-    "poetry-core>=2.0.0,<3.0.0",
-]
-build-backend = "poetry.core.masonry.api"
-
-[dependency-groups]
-dev = [
-    "toml (>=0.10.2,<0.11.0)",
-    "tomli-w (>=1.2.0,<2.0.0)",
-    "howdenconfig (>=1.0.6,<2.0.0)",
-]
+[project]
+name = "HowdenParser"
+version = "6.0.1"
+description = "A simple configuration manager with Pydantic and JSON export."
+readme = "README.md"
+requires-python = ">=3.12,<3.14"
+authors = [
+    { name = "JesperThoftIllemannJ", email = "jesper.jaeger@howdendanmark.dk" },
+]
+keywords = [
+    "config",
+    "configuration",
+    "pydantic",
+    "json",
+]
+homepage = "https://github.com/yourusername/config"
+repository = "https://github.com/yourusername/config"
+documentation = "https://github.com/yourusername/config"
+dependencies = [
+    "mistralai (>=1.9.3,<2.0.0)",
+    "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]
+text = "MIT"
+
+[build-system]
+requires = [
+    "poetry-core>=2.0.0,<3.0.0",
+]
+build-backend = "poetry.core.masonry.api"
+
+[dependency-groups]
+dev = [
+    "toml (>=0.10.2,<0.11.0)",
+    "tomli-w (>=1.2.0,<2.0.0)",
+    "howdenconfig (>=1.0.6,<2.0.0)",
+]

howdenparser-5.2.6/HowdenParser/parsers/llama_parser.py

@@ -1,106 +0,0 @@
-import os
-import json
-import logging
-from pathlib import Path
-from ..parser import BaseParser
-from typing import Any, Optional
-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,
-                 precise_bounding_box: Optional[bool] = None,
-                 line_level_bounding_box: Optional[bool] = None,
-                 ) -> None:
-
-        super().__init__()
-        # Capture only input parameters
-        self._input_params = {
-            k: v
-            for k, v in locals().items()
-            if k not in ("self", "__class__", "__len__")
-        }
-
-        logging.info("Configuring LlamaParser (lazy init enabled)...")
-
-        # Store configuration only
-        self.result_type = result_type
-        self.model = model
-        self.parse_mode = parse_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
-        self.precise_bounding_box = precise_bounding_box
-        self.line_level_bounding_box = line_level_bounding_box
-        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
-
-    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
-
-        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()
-        if rt_lower in ("md","markdown"):
-            rt = ResultType.MD
-        elif rt_lower in ("txt","text"):
-            rt = ResultType.TXT
-        elif rt_lower in ("json"):
-            rt = ResultType.JSON
-        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,
-            precise_bounding_box=self.precise_bounding_box,
-            line_level_bounding_box=self.line_level_bounding_box,
-        )
-
-        logging.info("LlamaParser initialized inside worker thread")
-
-    def parse(self, file_path: Path, include_pagenumbers: bool = False) -> Any:
-        # Ensure parser is created inside the thread event loop
-        self._lazy_init()
-        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':
-            return documents.model_dump()
-        else:
-            raise NotImplementedError(f"Result type {self.result_type} is not supported.")
-
-    def __call__(self, file_path: Path) -> str:
-        return self.parse(file_path)
-
-    def write_json_hyperparameter(self, folder_file_path: Path) -> None:
-        self.write_parameters(self._input_params, folder_file_path)