HowdenParser 0.1.2__tar.gz

howdenparser-0.1.2/HowdenParser/__init__.py

@@ -0,0 +1,3 @@
+from .parser import ParserFactory
+
+__all__ = ["ParserFactory"]

howdenparser-0.1.2/HowdenParser/parser.py

@@ -0,0 +1,137 @@
+from abc import ABC, abstractmethod
+from dotenv import load_dotenv
+import os
+from pathlib import Path
+
+load_dotenv()
+
+class BaseParser(ABC):
+    @abstractmethod
+    def parse(self, text: str) -> dict:
+        pass
+
+
+class MistralOCRParser(BaseParser):
+    def __init__(self, result_type: str, **kwargs) -> None:
+        from mistralai import Mistral
+        name = "MISTRAL-OCR-API-TOKEN"
+        api_key = os.getenv(name)
+        if not api_key:
+            raise EnvironmentError(f"Missing {name} in .env file.")
+
+        self.result_type = "md" if result_type.lower() in ("md", "markdown") else "text"
+        self.client = Mistral(api_key=api_key)
+
+    def parse(self, file_path: Path) -> str:
+        def upload_pdf(filename):
+            uploaded_pdf = self.client.files.upload(
+            file={
+            "file_name": filename,
+                "content": open(filename, "rb"),
+            },
+            purpose="ocr"
+            )
+            signed_url = self.client.files.get_signed_url(file_id=uploaded_pdf.id)
+            return signed_url.url
+
+        ocr_response = self.client.ocr.process(
+            model="mistral-ocr-latest",
+            document={
+            "type": "document_url",
+            "document_url": upload_pdf(file_path),
+            },
+            include_image_base64=True,
+            )
+
+        return "\n".join(doc.markdown for doc in ocr_response.pages)
+
+
+class LangChainParser(BaseParser):
+    def __init__(self, model_name: str):
+        from langchain.llms import OpenAI
+        self.model_name = model_name
+        self.model = OpenAI(model_name=model_name)
+
+    def parse(self, text: str) -> dict:
+        response = self.model(text)
+        return {"source": "LangChain", "output": response}
+
+
+class LlamaParser(BaseParser):
+    def __init__(self, result_type: str, mode: bool, **kwargs) -> None:
+        from llama_parse import LlamaParse, ResultType
+        if result_type.lower() == "md" or result_type.lower() == "markdown":
+            self.result_type = ResultType.MD
+        name = "LLAMA-PARSER-API-TOKEN"
+        self.api_key = os.getenv(name)
+        if not self.api_key:
+            raise EnvironmentError(f"Missing {name} in .env file.")
+
+        self.parser = LlamaParse(
+            api_key=self.api_key,
+            result_type=self.result_type,
+            premium_mode=mode
+        )
+
+    def parse(self, file_path: Path) -> str:
+        documents = self.parser.load_data(str(file_path))
+        return "\n".join(doc.text for doc in documents)
+
+class HuggingFaceParser(BaseParser):
+    def __init__(self, result_type: str, mode: bool, **kwargs) -> None:
+        from transformers import pipeline
+
+        if result_type.lower() in ("md", "markdown"):
+            self.result_type = "markdown"
+        else:
+            self.result_type = "text"
+
+        name = "HF-API-TOKEN"
+        self.api_key = os.getenv(name)
+        if not self.api_key:
+            raise EnvironmentError(f"Missing {name} in .env file.")
+
+        # OCR + text extraction pipeline
+        # Example model: microsoft/layoutlmv3-base-finetuned-docvqa
+        self.parser = pipeline(
+            task="document-question-answering",
+            model="impira/layoutlm-document-qa",
+            use_auth_token=self.api_key
+        )
+
+    def parse(self, file_path: Path) -> str:
+        import fitz  # PyMuPDF for PDF → images
+
+        pdf_doc = fitz.open(file_path)
+        output_parts = []
+
+        for page in pdf_doc:
+            pix = page.get_pixmap(dpi=200)
+            img_bytes = pix.tobytes("png")
+            response = self.parser(img_bytes, question="Extract all text")
+            if response and "answer" in response[0]:
+                output_parts.append(response[0]["answer"])
+
+        if self.result_type == "markdown":
+            # Here you could add rules to format into markdown
+            return "\n\n".join(output_parts)
+        else:
+            return " ".join(output_parts)
+
+# === Step 3: Dynamic factory using string input ===
+class ParserFactory:
+    @staticmethod
+    def get_parser(parser_type: str, **kwargs) -> BaseParser:
+        provider = parser_type.partition(":")[0].lower()
+        model = parser_type.partition(":")[2]
+        if provider == "langchain":
+            return LangChainParser(model_name=model)
+        elif provider == "llamaparser":
+            return LlamaParser(**kwargs)
+        elif provider == "huggingface":
+            return HuggingFaceParser(model_name=model, **kwargs)
+        elif provider == "mistralocr":
+            return MistralOCRParser(**kwargs)
+        else:
+            raise ValueError(f"Unknown parser type: {parser_type}")
+

howdenparser-0.1.2/PKG-INFO

@@ -0,0 +1,28 @@
+Metadata-Version: 2.3
+Name: HowdenParser
+Version: 0.1.2
+Summary: A simple configuration manager with Pydantic and JSON export.
+License: MIT
+Keywords: config,configuration,pydantic,json
+Author: JesperThoftIllemannJ
+Author-email: jesper.jaeger@howdendanmark.dk
+Classifier: License :: OSI Approved :: MIT License
+Classifier: Programming Language :: Python :: 2
+Classifier: Programming Language :: Python :: 2.7
+Classifier: Programming Language :: Python :: 3
+Classifier: Programming Language :: Python :: 3.4
+Classifier: Programming Language :: Python :: 3.5
+Classifier: Programming Language :: Python :: 3.6
+Classifier: Programming Language :: Python :: 3.7
+Classifier: Programming Language :: Python :: 3.8
+Classifier: Programming Language :: Python :: 3.9
+Classifier: Programming Language :: Python :: 3.10
+Classifier: Programming Language :: Python :: 3.11
+Classifier: Programming Language :: Python :: 3.12
+Classifier: Programming Language :: Python :: 3.13
+Project-URL: Documentation, https://github.com/yourusername/config
+Project-URL: Homepage, https://github.com/yourusername/config
+Project-URL: Repository, https://github.com/yourusername/config
+Description-Content-Type: text/markdown
+
+.\build.ps1

howdenparser-0.1.2/README.md

@@ -0,0 +1 @@
+.\build.ps1

howdenparser-0.1.2/parameter/__init__.py

@@ -0,0 +1,3 @@
+from .huggingface import Parameter as Huggingface
+from .llamaparser import Parameter as Llamaparser
+from .mistralocr import Parameter as  Mistralocr

howdenparser-0.1.2/parameter/huggingface.py

@@ -0,0 +1,6 @@
+from typing import Literal
+from pydantic import BaseModel
+
+class Parameter(BaseModel):
+    model: str = "huggingface:HURIDOCS/pdf-segmentation"
+

howdenparser-0.1.2/parameter/llamaparser.py

@@ -0,0 +1,8 @@
+
+from typing import Literal
+from pydantic import BaseModel
+
+class Parameter(BaseModel):
+    model: str = "llamaparser:"
+    result_type: Literal["md"] = "md"
+    mode: bool = False

howdenparser-0.1.2/parameter/mistralocr.py

@@ -0,0 +1,7 @@
+from typing import Literal
+from pydantic import BaseModel
+
+class Parameter(BaseModel):
+    model: str = "mistralocr:"
+    result_type: Literal["md"] = "md"
+    mode: bool = False

howdenparser-0.1.2/pyproject.toml

@@ -0,0 +1,21 @@
+[build-system]
+requires = [ "poetry-core>=2.0.0,<3.0.0",]
+build-backend = "poetry.core.masonry.api"
+
+[tool.poetry]
+name = "HowdenParser"
+version = "0.1.2"
+description = "A simple configuration manager with Pydantic and JSON export."
+authors = [ "JesperThoftIllemannJ <jesper.jaeger@howdendanmark.dk>",]
+readme = "README.md"
+license = "MIT"
+keywords = [ "config", "configuration", "pydantic", "json",]
+homepage = "https://github.com/yourusername/config"
+repository = "https://github.com/yourusername/config"
+documentation = "https://github.com/yourusername/config"
+[[tool.poetry.packages]]
+include = "HowdenParser"
+
+[[tool.poetry.packages]]
+include = "parameter"
+